Citation
Forecasting advertising communication effects using media exposure distribution models

Material Information

Title:
Forecasting advertising communication effects using media exposure distribution models test market results in South Korea
Creator:
Park, Hyunsoo, 1966-
Publication Date:
Language:
English
Physical Description:
xiv, 211 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Advertising ( jstor )
Advertising campaigns ( jstor )
Advertising expenditures ( jstor )
Advertising media ( jstor )
Advertising research ( jstor )
Advertising to sales ratios ( jstor )
Brands ( jstor )
Consumer advertising ( jstor )
Modeling ( jstor )
Target audiences ( jstor )
Dissertations, Academic -- Mass Communication -- UF ( lcsh )
Mass Communication thesis, Ph.D ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1998.
Bibliography:
Includes bibliographical references (leaves 203-210).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Hyunsoo Park.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
029246829 ( ALEPH )
39556936 ( OCLC )

Downloads

This item has the following downloads:


Full Text










FORECASTING ADVERTISING COMMUNICATION EFFECTS
USING MEDIA EXPOSURE DISTRIBUTION MODELS:
TEST MARKET RESULTS IN SOUTH KOREA














By

HYUNSOO PARK


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1998






























Copyright 1998

by

HYUNSOO PARK













ACKNOWLEDGEMENTS


First of all, heartfelt thanks goes to my adviser, Dr. Kent M. Lancaster. I really
appreciate his kind guidance in developing the research design and the normative
framework, and in writing this dissertation. Without his scholarly advice, guidance, and
great help during my whole Ph.D. study, this dissertation would not have been possible. I
also would like to thank my committee professors, Dr. Joseph Pisani, Dr. Leonard Tipton,
Dr. Marilyn Roberts, and Dr. Richard Scheaffer for excellent suggestions, encouragement,
and help.
Acknowledgments are due to several professionals and professors in Korea.
Special gratitude goes to the CEO of SangAm agency, Dr. Tong-Yeol Sung, and Ms.
Nan-Hee Choi, Vice-chief of the marketing department, for their considerable efforts in
providing valuable data. Mr. Hyun-Tae Kim, Dr. Jung-Sik Jo, and Dr. Yang-Ho Choi also
provided valuable help.
Appreciation is also extended to June-Hyung Ji and my brother-in-law, Yeon-Soo
Kang, for sending me precious Korean papers and data for this study. Special thanks go
to my brothers and friends, Yoon-Soo Park, Jeong-Soo Park, Kwang-Ho Lee, Dong-
Hyun Han, Yoon-Sung Hwang, Il-Joon Yoo, Seung-Kwan Ryu, and Ho-Cheol Kim
whose encouragement and friendship made this dissertation possible.
I am really grateful to my lovely wife, Yeon-Mi, whose self-sacrificing support and
help during my Ph.D. study have given me the inspiration, confidence, and balance
necessary to complete this endeavor.
Finally, I would like to dedicate this dissertation and my Ph.D. degree to my
parents, Jong-gug Park and Bum-Rok Lee. "Thank you for life, love, and help every
moment of my life. I love you."










i4 oi 1 4 9I RE 7-44 l-*g4 Zi-
-S-a-a- "r l a-$ rB O P 4. 01 -*)0 11013) = -7
924- 471xir @AMr 8 E 44. Il 2r< 'gol golt 1
0oi- dr + MA-l, 01 7s0l 71 Ot -E 1A % f-4-


*& s-g- i-444 44.1 444g -9- l -4q-0 11% 4TSi4
Dr. Kent M. Lancaster d9 13-- A4* } -E5'-14. 0A} 4Y91
sEf 463- j4 94 rQg4 7fi4X 0 f3R j41s-4q4. O}1-Bi V
-- W 44 2 -2 xl~4 Z Mel of.l Q- t4 Dr. Joseph
Pisani, Dr. Leonard Tipton, Dr. Marilyn Roberts, Dr. Richard
Scheaffer -1+-- 7*A4 -g qq4.


"A ^S-4. 44 n# SH4 1444 F-1 S.^*0-S. ^^ F
(2)7714Q rARl014 I 44I, 4A +A4i4 61 4 t)a 7


-44-?, A \OVl SP A elm +14 4utBl a 4 ^)), 3W


i3)401, -2 S01 University of Florida .4 EIdW-4 *4A4SW4 174
718-4 WP4 -a4.
5s1-al *-14-k4% -YO-I SIx as l4AA4 A471* 1l A tr :
94 AA 7j a oI 'i, + -efl, 0) -, J g-, V ,, g
st -^s4 41941 +4g rgo sgo] al-4.
11R94 4971 M49?! -Aa 0 -1 t i-44^1 etA
4t^}% Of4 4aD14'11, p-g44 491r 40 _1 *ia-, a-^4R 4A-=
*-4A -V 44. 'LLI, 71' ^a4 "'t J 4A 1 4 sil qL 0l 9 sA-0
4M9413"-4 94% el 4^l r t ^q o 4^P 414 42ltl 3--t
^-tJ;-017 44-1 !L 44.











TABLE OF CONTENTS


page


ACKNOWLEDGEMENT N III

LIST O F T A B LE S ........................ ..... ...... ............................ ........ ....... ix

LIST OF FIG URES......... ..... ...... .. .. ............................. .. ........ .. .. xi

ABSTRACT................................ ............ .. ............. .. ................. xii

CHAPTER I IN TRODU CTION ........................... ..... .............. .... ......... .. 1

A advertising C am paign E effects ....................................... ................................... ......... 1
Com plexity of the Problem ........................................ ....... .................. ...... ...............2
M edia Exposure Distribution M odels............................. ............3
Objectives of this Study. .................. .... ...... .. .. ... ...................... 5
R atio n ale .................................................................. .... ......... 6
O organization of the D issertation................... ..................... .............. ........................ 8

CHAPTER II. MEASURING AND FORECASTING ADVERTISING EFFECTS ......9

Primary Reason for Evaluating and Forecasting Advertising Campaigns......................... 10
Problems in Advertising Campaign Measurement................................................ 11
Advertising Effects on Sales............... .... ............... .................. .......... ................. 13
Difficulties in Measuring Advertising Effects on Sales.................................... 14
Advertising Communication Effects............ ................ ..................... ............... 16
Awareness .................................................... .... ......... ....... 17
Previous Research on Aw areness ...................................... .......... 19
Recall ..................................................................... .. ........ 20
Liking and Attitude Toward Advertisement........................... ................. 21
Preference ........................... ...... ........... .. .... ..... ............... .. 23
Purchase Intention and Actual Purchase ...................... ............................24
Achieving Desired Communication Effects........... ........ ....... ... ........... 25
Effective Exposure to A dvertising................ .............................................. ............... 28
Advertising Exposure......................... ... .... ....... .. 28
R each .......................... .. ..... ............................. ....... ......... 2 9
Frequency and Frequency Distribution.............................................. ............ 30
Effective Frequency and Reach.......................................................................... 32
Problems with Using the Concept of Effective Frequency............................... 34
How to Set Effective Frequency Levels........................................... ...... .. ....37








CHAPTER III. BACKGROUND ON THE SOUTH KOREAN ADVERTISING
INDUSTRY...................... .. .. ........ ................... 42

Introduction......................... ............ ............................ .................... 42
South Korean Economic Development ....................................... ..... .......... 42
The Growth of the Advertising Industry in South Korea.................................... 43
A advertising M edia in South K orea .................................................... ........... 45
Recent Changes in the South Korean Advertising Industry ................................. 47
Studies on Measuring and Forecasting Advertising Effects in South Korea......... 48
Media Planning and Evaluation Procedures of South Korean Advertising Practitioners ..51
Purpose of the Study ......................... ......................... .................................. 52
M eth o d s .............. ..... .......... ................................. ............ 5 2
R results ................................... ... ................... ................... ........ 53
Definition of Reach and Effective Reach................................................ 56
Com m unication Effects U sed...................... ..................... ................... 57
Message Weighting and Reasons of Not Weighting ............................... 59
N eeded Im provem ents ...................................... ....... ....... ........ 60
Discussion and Implications for Further Research............................ ........... 61

CHAPTER IV. M ETHODOLOGY ............................... ... ................... ....... 64

C ase S tu d y ............ ... .. .. ..... ...... .. .. .. ... .. .. .. .. .. ... .. .. .... .. .. .. .. 6 6
A SIA N A A irline.................... .... ... ..... ...... .. ... .... 66
Target Audience and Advertising Campaign ................ ............................ 69
C o m p e tito r ......................................................................................................... 7 0
M edia D ata ................ ........... ...... ........... ........... ..... ..... .................. 72
T V R atin g s ......................................... ................................... 7 2
M magazine R atings ....................................... ..... 73
T racking D ata ........................................ ...... ... ..... .. 74
S am pling ............................................ .. .. 74
M easurem ent ......................... ... ... ... .. ............................. 75
Media Evaluation Model Used in This Study ............................................... 76
Beta Binomial Exposure Distribution Model.................................... 76
M odel Selection................................. .. ..... .. ....... .. .. 79
M odel M modification ..... .................................................. .............. ... 82
Error Estimation for the Media Evaluation Model ............................. 83
Suggestions for Improving Campaign Efficiency and Budget Levels ................. 87
Sophisticated versus N alive Comparison ............................................................. 88
Summary of the Normative Framework................... ..................................... 89
R research H ypotheses ................................................................................... .. ............. 90
Testing Procedures ...................... ......................................... 93
Statistical Procedures ............ ...... ........... ............. .......... 97
Calibrating Forecasting Procedures .......................................................... .. .. 99








CH A PTER V RESULTS............................................... ................................ 104

Intro du ctio n ............................ .. .. ... .. .. .. .. ... .. .. .. .. .. .. .. .. .. ..... 104
T est M market R esults................. .. ............... ..... ....................... ........... 106
M edia Evaluation M odel Prediction..................................... ...................... 106
T racking Study R results ............................... ...................... ....................... 109
Testing Hypotheses............................ ....................................... ..... .. 111
Awareness of the Campaign Message ......................................... .... 111
Preference ..................... ......................... .. .. .... .. .. ... .. 111
W willingness to U se ................................ ......................... .............. 112
Calibrating Procedures ................................................ 113
Critical Analysis of the ASIANA Advertising Campaign.................... ...... ....... 120
Suggestions for Future ASIANA Campaigns.................. ...............125
Selecting Appropriate Vehicles................. .............................. ....... .. 125
Appropriate Budget for Monthly Schedule................................. 127
Sophisticated versus Naive Comparison ................. ......... .......... 132
Pooled Approach............................... ............. 132
Vehicle versus Message ................ .......... ....... .............. 133

CHAPTER VI. SUMMARY, CONCLUSIONS, AND IMPLICATIONS ................. 135

Sum m ary and C conclusions ............................................................. ....... 135
Im plications ................................... ........................ ............ .. 139
Limitations and Suggestions for Future Study .......................................................... 140

APPENDICES

A. SURVEY QUESTIONNAIRE IN BOTH ENGLISH AND KOREAN...... 143
EN G LISH V ER SIO N ................................. ..................... ........ 144
KOREAN VERSION .......................................... .......... 151

B. TWO TYPES OF MAGAZINE ADVERTISEMENTS ........................... 158

C. TV PROGRAM RATINGS, ADVERTISING MESSAGE RATINGS,
MAGAZINE RATINGS, AND COSTS FOR THE ASIANA
ADVERTISING CAMPAIGN IN BOTH ENGLISH AND
KOREAN... .............................. ..... ........ 161
ENGLISH VERSION ................................................. ................ 162
KOREAN VERSION ....... ...................................... ................. 166

D. GENERAL FORM OF THE BBMD MODEL................... ............. .. 170

E. ADPLUS RESULTS FOR NETWORK TELEVISION IN FEBRUARY.... 176

F. ADPLUS RESULTS FOR NETWORK TELEVISION IN MARCH.......... 179








G. ADPLUS RESULTS FOR NETWORK TELEVISION IN APRIL............. 182

H. ADPLUS RESULTS FOR MAGAZINE ADVERTISEMENTS................. 185

I. ADPLUS RESULTS FOR NETWORK TELEVISION AND
MAGAZINES COMBINED IN FEBRUARY ............................. 187

J. ADPLUS RESULTS FOR NETWORK TELEVISION AND
MAGAZINES COMBINED IN MARCH................. .................... 191

K. ADPLUS RESULTS FOR NETWORK TELEVISION AND
MAGAZINES COMBINED IN APRIL............................ ........... 195

L. ADPLUS RESULTS WITH FLOWCHART......................................... 198

R EFE R E N C E S ............................. ...... ... .... .. .... .. .. .. .... .. ... ... 203

BIO G RA PH ICA L SK ETCH .............................................................. ........ ....... 211














LIST OF TABLES


Table page

1. Common measures for advertising communication effects.................................. 17
2. Relationship between common communication measures and needed
exposure frequency in a normal marketing, copy, and media situation ............. 27
3. Example of frequency distribution................... ....... ....... ....... ..... ......... 32
4 Marketing, copy and media factors that affect effective frequency........................ 38
5 Suggested additional points and modified effective frequency levels
for brand X ....... ... ... ....... .............. ...................................... .. 41
6. The progress of the South Korean advertising industry................... .......... .. 45
7. South Korean advertising media, total expenses, and costs...................................... 46
8. Sung's independent and dependent variables and regression equations................... 50
9. Systems used to develop advertising media plans ..........................................54
10. Factors used in evaluating media plans.............................. ....................... 55
11. Representative definitions of effective reach in comparison with two
previous Am erican studies .................................... .......... ................. 58
12. Variables used to evaluate advertising communication effects............................ 59
13. Normative framework necessary to predict advertising communication effects ........ 65
14 D om estic and international m market shares ...............................................................68
15. Advertising insertions by media category and expenses of
"Her name is ASIANA" campaign from February 17 to May 16, 1996............. 70
16. Competitor GRPs and advertising expenses.... ........................... .......... 71
17. Sample characteristics ...................................... .......................... .. 75
18. R2, and R, estimation equations for network television............ ......................82

19. Summary of the normative framework applied to the ASIANA case study............... 89








20. Suggested points and modified effective frequency levels
for ASIANA Airline campaign effects ................ ................... ......... 91
21. Adjustment options, expected effects, and decisions in this study......................... 102
22. Consumer tracking study results and model predictions.................................... 110
23. Adjustment options and changes in predictions. .......................... ................. 116
24. Consumer tracking study results and model predictions with calibration ............... 118
25. CPM-MSG figures for the most efficient 11 television programs (vehicles) ........... 122
26. CPM-MSG figures for all magazine advertisements.............. ......................... 124
27. Relatively high CPM-MSG vehicle list in the ASIANA campaign.......................... 126
28. Relatively low CPM-MSG vehicles added for schedule modification ...................... 126
29. Suggested monthly budget levels and estimated message reach 3+........................ 131

















LIST OF FIGURES


Figure page

1. The effects of top-of-mind awareness of the advertisement............... ............ 20
2. The shape of a typical m message reach curve.................................................... ...... 30
3. Sung's model to estimate advertising effects ............... ................ ............... 50
4 A D p lu s re su lts ................................................... ................................... .......... 8 1
5. M edia evaluation m odel prediction............................................. ..................... 108
6. M edia evaluation model prediction with calibration............................................... 117
7. Estimated frequency distributions for the three-month ASIANA campaign............ 119
8. Budget allocation among television programs............................. .................... 120
9. Budget allocation among m agazines....................... ......................................... 121
10. Reach percentage changes with budget increases for February vehicles,
television and magazines combined .......................................... .............. 128
11. Reach percentage changes with budget increases for March vehicles,
television and magazines combined ................................. ................. 128
12. Reach percentage changes with budget increases for April vehicles,

television and m magazines com bined ..................................... ...................... 129
13. Estimated reach (3+) for each monthly advertising schedule .................................. 130
14. Estimated message reach (3+), naive versus sophisticated approaches.................. 133
15. Estimated frequency distributions with vehicle and message data............................ 134
16. Ideal procedures for predicting and evaluating advertising effects using media
exposure distribution models ............ ... ........ ... .. ....... 136














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy

FORECASTING ADVERTISING COMMUNICATION EFFECTS
USING MEDIA EXPOSURE DISTRIBUTION MODELS:
TEST MARKET RESULTS IN SOUTH KOREA

By

Hyunsoo Park

May, 1998

Chairman: Dr. Kent M. Lancaster
Major Department: Mass Communication

The primary objective of this study is to demonstrate how to predict advertising

communication effects using a variation of the beta binomial exposure distribution (BBD)

model. A computer based model which embodies important aspects of normative theories

of media planning is used to forecast the communication effects of a South Korean

advertising campaign.

Due to the lack of systematic study of advertising effects in South Korea, the

South Korean advertising industry needs objective and scientific methods to develop a

model that is applicable to the South Korean advertising environment. Therefore, this

study applied a media exposure distribution model that is used in the United States and

worldwide to the South Korean market to probe how accurately such models and

procedures can predict the communication effects of a South Korean advertising








campaign. A consumer tracking study is used to determine the accuracy of forecasting

advertising campaign effects.

Prior to the application of the media evaluation model and procedures, survey

research was conducted among South Korean media practitioners. The goal was to show

how the media practitioners view the current state of reach/frequency estimation methods

and which communication effects are used to predict or measure potential advertising

impact. Based on the survey results, previous studies as well as marketing, copy, and

media factors, an effective exposure frequency level is estimated to predict target audience

awareness of a South Korean advertising campaign message.

In addition to advertising message awareness, this study also examines target

audience preference and willingness to purchase for the brand used as a case study.

Minimum frequencies of consumer exposure to the advertising campaign necessary to

achieve preference and willingness to purchase also are estimated with the same normative

framework. Then, the minimum effective frequency levels are analyzed in comparison

with actual consumer tracking data to determine whether more accurate predictions would

have been possible.

The study results showed that the normative framework and media exposure

distribution model used in this study can be accurately used in predicting target audience

awareness of the campaign message. The prediction for target audience awareness of the

campaign message was close to the tracking study results. With additional calibrating

options, the predicted awareness of the campaign message could be exactly matched with

the tracking study results.








This study also suggests that there may be a hierarchy in consumer processing of

advertising messages. The needed target audience exposure to achieve preference was

significantly higher than the needed exposure for awareness. The needed message

exposure for willingness to use was also significantly higher than the target audience

exposure for awareness and preference. Though many more factors should be considered

for accurately predicting or evaluating advertising effects in terms of preference and

willingness to use, this study still provided useful results for the estimation of advertising

effects on preference and willingness to use.

The advertising campaign used as a case study also was critically analyzed to

suggest appropriate monthly budgets and vehicle selections for future campaigns. The

sophisticated approach used in this study was compared with typical naive approaches to

show how the potential gaps can be misleading.

In this study, the normative frameworks, media exposure distribution models,

evaluation methods, including many influential factors, and calibrating procedures provide

useful guidance for forecasting advertising communication effects.














CHAPTER I
INTRODUCTION




Advertising Campaign Effects


Advertising, an important promotional tools of modern marketing management,

has been used worldwide to present information to potential buyers and reinforce

consumer preference for a company's products or services (Lilien, Kotler, & Moorthy,

1992). Every year billions of dollars are invested by companies around the world to test,

produce, and implement advertising campaigns. According to Aaker, Batra, and Myers

(1992), estimated advertising expenditures will be $780 billion worldwide and $320 billion

in the United States alone by the year 2000. Because of these enormous expenditures,

advertisers want the advertising planners and managers to identify exactly what are the

results obtained from the advertising investment and to provide evidence of the return on

investment. Therefore, industry practitioners and social scientists throughout the world

continue to predict and evaluate the effects of advertising campaigns not only prior to

campaign execution but also after advertising is aired or printed.








Complexity of the Problem


The effects of advertising campaigns are evaluated according to campaign

objectives. No matter what objectives an advertising campaign has, such as an increase in

sales, market share, brand awareness, advertising message awareness, recall, preference,

or something else, there seem to be limited means of accurately measuring advertising

campaign effects. The main reason is that advertising effects typically interact with other

elements in the marketing mix, such as an advertiser's image, price, competitor

promotional activity, distribution, product quality, and other promotion activities.

Although technology is making the prediction and evaluation of advertising effects easier,

skepticism about the possibility of accurately predicting or evaluating advertising effects

has continued over the years (Lancaster & Helander, 1987).

Another problem in predicting or evaluating advertising effects is what is to be

measured? Two frequently used approaches are advertising communication effects and

sales. Since one of the main purposes of advertising is to communicate persuasive

messages, measuring communication effects, such as brand and advertising message

awareness, recall, attitude toward brand, preference, and willingness to purchase, is a

natural way to evaluate advertising campaign effects (Schultz, Martin, & Brown, 1984;

McDonald, 1995; Woodside, 1996).

Measuring actual behavior of consumers, such as sales or purchases, is another

frequently used approach to measure advertising campaign effects. Since the ultimate goal

of advertising may be the increase of sales, the amount of sales has been used to measure

advertising effects (Leone, 1983; Assumus, Farley, & Lehmann, 1984; Holstius, 1990;








Schreiber, & Appel, 1991; Lodish et al., 1995). However, since sales are the result of

various marketing efforts, it is very difficult to measure the absolute effects of an

advertising campaign in terms of sales.

Advertising communication effects can be used more efficiently as a method of

evaluating the effects of advertising campaigns. Brand awareness, advertising message

awareness, recall, preference, and willingness to purchase are the most frequently used

measures in terms of communication effects. For example, advertising campaign message

awareness can be measured with much less interaction with other marketing factors than

the measure of sales. However, the communication measures should be explained in

relation to the final objectives, such as increase in sales or market share. Many previous

studies in the United States suggest that there is a positive relationship between brand or

advertising message awareness and preference, sales, and market share (Burke &

Schoeffleer, 1980; Wilson, 1981; Zufryden, 1996; Woodside, 1996).

What is best among various communication effects is up to the planners to decide

in relation to campaign objectives. However, no matter what kind of approach is used to

measure advertising campaign effects, it is still difficult to identify actual advertising

effects while completely controlling for all other important controllable or uncontrollable

dimensions of the marketing mix, such as pricing behavior, distribution intensity, product

quality, competitor target market selection, and other promotion activities.



Media Exposure Distribution Models


Because of the enormous advertising expenditures and the complexity of the

problem, studies of advertising effects have been active in the advanced countries like the








United States in an effort to overcome the problems explained above (Leckenby & Kishi,

1982a; Lancaster & Helander, 1987; Rust & Klompmaker, 1981; Rice, 1985; DeKimpe &

Hanssens, 1995a; Hansen, 1995). Models have been developed to predict and measure

advertising effects.

One of the most frequently used models in the United States is the beta binomial

exposure distribution (BBD) model (Leckenby & Kishi, 1982), which was developed by

Metheringham in 1964. The general goal of the BBD model is to estimate the probability

and distribution of consumer exposures to advertising media schedules. The model

incorporates both beta and binomial distributions of schedule exposures. The beta

distribution is used to estimate the probability of consumer exposures to any number of

advertising vehicles or messages included in a schedule, such as one, two, three, or up to

any number of exposures to advertising vehicles or messages. These probabilities are then

used with the binomial distribution which estimates the proportion of consumer exposures

to advertising vehicles or messages, for example, once, twice, or three times (Lancaster &

Katz, 1989).

Computer-based models also have been developed based on the beta binomial

exposure distribution (BBD). One such microcomputer program is ADplus, which is used

to evaluate advertising schedules based on a variation of the beta binomial exposure

distribution model. It uses procedures that vary by media category to account for

differences in available audience and cost data (Lancaster, 1993).

Three test market studies in the United States (Lancaster & Helander, 1987;

Lancaster & Katz, 1989) demonstrate the ability of the predecessor of ADplus to predict








the effects of advertising campaigns, including communication effects such as advertising

exposure and brand awareness.



Objectives of this Study


While the study of advertising effects has been active in advanced countries, it has

not been widespread in some other countries such as South Korea. Even though the

South Korean advertising industry has in recent years grown into the world's 13th largest

market (US $6.1 billion in 1995), its research remains at an immature level. Only a few

studies have been conducted about measuring advertising effectiveness in South Korea

(Sung, 1989; Korean Gallup, 1992).

Because of the lack of systematic study of advertising effects in South Korea, the

development of an advertising effects model applicable to the advertising environment of

South Korea is relatively poor. Therefore, a model of measurable advertising effects that

meets the advertising environment of South Korea is needed. Such a model should be

developed by means of an objective and scientific method for forecasting and evaluating

advertising effects. Much more research should be conducted in South Korea concerning

the evaluation of desired communication effects.

Therefore, one of the primary objectives of this study is to demonstrate how to

predict the effects of an advertising campaign using a variation of the beta binomial media

exposure distribution model in the South Korean market. The computer-based beta

binomial exposure distribution model ADplus will be used to forecast the effects of an

advertising campaign. These predicted advertising effects will be compared with actual

consumer survey results in the South Korean market. Consequently, the computer-based








BBD model developed in the United States will be used to examine how accurately the

model can predict the effects of a South Korean advertising campaign.

Little information is available about how South Korean media practitioners view

current media evaluation terms and which communication effects are frequently used to

predict or measure potential advertising impact. Therefore, a survey was conducted prior

to the application of the media evaluation model and procedures. Based on the survey

results and previous American and Korean studies, optimal advertising exposure

frequencies will be estimated to predict desired communication effects, such as consumer

awareness of an advertising campaign message, preference, and willingness to purchase.

Then, a consumer tracking study will be used to determine how accurately the model

predicts the potential advertising impact.

This study also will examine methods used to measure advertising effects as well

as factors that may influence advertising communication effects, such as advertising

message quality, market share, and competitor promotional activities. For the brand used

in this study, these factors will be used to determine the level of effective frequency, which

is a predetermined level of consumer exposure repetition to an advertising campaign that

is needed to achieve desired communication effects. For example, the brand that has

strong competitor advertising, low market share, low brand loyalty, or low advertising

message quality may need a relatively high effective frequency level.



Rationale


Advertising planners have experienced difficulties in predicting and measuring the

effects of advertising campaigns over the years. One of the major challenges in evaluating








advertising effects is the inability of researchers to estimate the accurate functions between

advertising and consumer responses. Predicting advertising effects is certainly challenging

due to the difficulty in controlling other important dimensions of the marketing mix

(Lancaster & Helander, 1987). Therefore, this study will attempt to overcome some of

these problems. Advertising expenditures will be analyzed by their essential message and

media characteristics, instead of examining the amount in relation to only sales or

communication variables. Actual advertising media schedules and messages gathered from

the South Korean market will be analyzed from the perspective of media planners in terms

of the likely effects on the intended target consumers.

Since the reliability of the beta binomial exposure distribution model has been

proven over decades in the United States (Leckenby & Boyd, 1984; Leckenby & Kishi,

1982; Kreshel, Lancaster, & Toomey, 1985; Rust 1986; Lancaster & Helander, 1987), it

will be worthwhile to apply the BBD model to a totally different market, such as South

Korea, to examine whether the model works as it does in the American market.

The total expenses for advertising in South Korea increased more than 70-fold

during the last two decades. However, in spite of the recent rapid growth of the South

Korean advertising industry, there have been few Korean studies on measuring and

predicting advertising effects. Studies should be conducted to develop a model that is

suitable for the Korean advertising environment. This study, which applies a model used

in the United States and around the world, such as Europe and South Africa, to the South

Korean market, will help South Korean advertising planners develop scientific models.

These can then be used to evaluate and forecast advertising results in terms of

communication effects.








Organization of the Dissertation


This dissertation is organized into six chapters. The introductory chapter provides

the conceptual framework of this study, including the objectives and rationale. The

second chapter presents a literature review related to the purpose of this study. The

review includes many problems of measuring advertising effects as well as previous studies

on predicting and measuring advertising communication effects. Chapter II also presents

the concept of effective exposure to advertising campaigns from the prospective of media

planners.

Chapter III presents South Korean economic development, the growth of the

South Korean advertising industry, the advertising media, recent changes in the advertising

industry, and previous studies measuring and forecasting advertising effects. Survey

results of 61 South Korean advertising agencies are also included in Chapter II. This

survey research demonstrates how South Korean media practitioners perceive media

evaluation terms and the current state of reach and frequency methods.

Chapter IV, Methodology, explains the brand used as a case study, the framework

used to analyze this case study, data collection, research hypotheses, testing procedures,

and calibrating options. Chapter V concerns the results of hypothesis testing and the

application of a commercial model to the South Korean market. Test market results are

provided with comparison to the model predictions of advertising communication effects.

Finally, a summary, conclusions, implications, limitations and suggestions for

future research are presented in Chapter VI.















CHAPTER II
MEASURING AND FORECASTING ADVERTISING EFFECTS






Over the past decades, the research of advertising effects has evolved with great

detail and diversity of theoretical, methodological, and technical approaches. These

research efforts suggest that a possible solution to the ultimate problems of measuring or

predicting advertising campaign effects may be within reach (Dutka, 1995). In this

chapter, the reasons and problems with forecasting and measuring advertising effects, and

publications conducted to help measure or forecast advertising effects are reviewed,

focusing on recent developments. The review of literature will help develop appropriate

methods to conduct this study.

Advertising target audiences should be exposed to the advertising campaign with

optimal frequency to achieve the desired advertising effects. Since there is little agreement

on precisely how many exposures to an advertising campaign message are sufficient to

achieve the advertising campaign goal, studies have been conducted to examine this issue

(Krugmen, 1972; Craig, & Ghosh, 1993; Dekimpe, & Hanssens, 1995;).

In this chapter, in addition to the previous studies on predicting or measuring

advertising effects, the factors that may influence the number of sufficient exposures are








also reviewed. This study will attempt to integrate the possible influential factors to

predict and measure advertising effects from the perspective of media planners.



Primary Reasons for Evaluating and Forecasting Advertising Campaigns


The decision process for advertising typically begins with some form of planning

and either an implicit or explicit strategy or objective. After setting or establishing

measurable objectives and a budget for the various stages and elements in the advertising

campaign, creative strategies and media plans are prepared. Consequently, the planners

need to know whether the advertising campaigns achieved the goals established. Because

of enormous expenditures on advertising campaigns, advertisers want advertising planners

and managers to identify exactly what are the results to be obtained from the advertising

investment and to provide evidence of the return on investment. The advertising planner's

role is to be able to answer the question of whether advertising is the best way to use the

company's resources. Some form of advertising research, such as advertising evaluation,

is required to answer the question.

Most advertising research can be divided into two forms. One is used to predict

what might occur in the real-world marketplace. On the other hand, evaluation research

also can be used as a basis for future actions. In this case, its basic purpose is to measure

what occurred as a result of the advertising campaign and what was returned on the

advertising investment.

According to Schultz and Barnes (1994), there are three primary reasons for

evaluating an advertising campaign. First, advertising campaign evaluation can be used to

determine whether or not the overall advertising campaign objectives are accomplished.








Individual objectives for the various elements of the campaign also can be evaluated and

then, based on these evaluations, the planners can determine why failure may have

occurred.

Second, the return on the campaign investment can be quantified and justified by

knowing what is accomplished. Management also can use the information to explain the

relationship with other potential uses for funds and determine the cost effectiveness of the

campaign.

Third, the measurement information also can be used to change, add, or correct

the campaign plans for future campaigns. Since advertising campaigns can always be

improved, evaluations of previous campaigns and predictions of current campaigns are a

great help to improve the elements and makeup of future campaigns.



Problems in Advertising Campaign Measurement


Although it is important to measure and forecast advertising campaign effects, it

has been difficult for advertising campaign planners to accurately evaluate and predict the

effects of advertising campaigns over the years.

One of the major challenges in measuring advertising campaign effects is what

should be measured. Advertising communication effects, such as consumer awareness of

a brand, or actual behavior of consumers, such as sales or purchases, can be used to

measure and evaluate advertising campaign effects. According to the campaign objectives

established, advertising planners should decide what measures are best. However, no

matter what objectives were set for the campaign, such as sales, market share,








communication effects, or some other, there are problems inherent in the measurement

process.

Schultz and Bares (1994) summarized the problems in measuring advertising

campaign effects. First, there is a problem in differentiating between what is advertising

and what is not. In most cases, advertising is one part of the whole marketing effort.

Certainly advertising and all other marketing efforts, such as sales promotion and public

relations, do not work in isolation. For example, if a customer sees advertising sales

messages carried on the manufacturer's coupon and remembers it, but not the actual

advertising media message itself, this calls into question whether the advertising campaign

should be credited for the consumers' awareness of the advertising sales message. It is

very difficult to identify actual advertising effects while controlling for all other important

controllable or uncontrollable dimensions of the marketing mix, such as pricing behavior,

distribution intensity, product quality, competitor target market selection, and other

promotion activities. Usually advertising effects can not be separated from other

marketing efforts.

Second, since most campaigns are advertised over several months or even a year, it

is hard to identify the exact effects of one campaign. In most cases, the advertising

campaign effects are built over time and there are lagged effects of advertising campaigns.

Third, the goals of most advertising campaigns are subjective to interpretation.

For example, when the advertising objectives are communication effects, such as

awareness, knowledge, preference, or recall, the measurement of advertising effects often

can not be as precise as those for direct marketing in which consumer response is

measured more accurately.








The fourth problem is attributed to human memory. No one can remember all the

advertisements he or she has seen or been exposed to. Due to the lack of information

about how human memory and information storage works, it is certainly difficult to say

exactly what should be measured to evaluate advertising campaigns.

As specified above, skepticism at the possibility of accurately measuring and

forecasting the effects of advertising campaigns has grown over the years. However, in

spite of all these problems, the results of advertising campaigns must be measured and

predicted because of the advertisers' and advertising planners' needs and requirements for

accountability. Two frequently used approaches are sales and communication effects.



Advertising Effects on Sales


The purpose of advertising is to provide information that changes consumers'

mental and behavioral responses in a manner favored by the advertiser. However, it is

difficult to measure or predict advertising effects accurately because, typically, there is

more than one response resulting from an advertisement. Therefore, advertisers should

decide what is the best consumer response describing the effectiveness of the

advertisement (McDonald, 1995).

Generally, advertising can be measured and predicted at two levels, either

communication effects or sales effects. Sometimes, advertising effects are measured on

the behavior level, which includes not only purchase behavior but also information seeking

behavior before making a purchase. Direct advertising, which is general for industrial

marketing, makes it easy to measure advertising behavior effects. For example, customers

can receive detailed information about a certain product or service, if only they send a








coupon, which is available on a certain part of the advertisement. However, academically

and practically, advertising effects are usually evaluated in terms of communication effects

or sales effects.

Since the ultimate goal of most advertising campaigns is either to increase sales or

market share or to defend achieved positions, many researchers have studied the

advertising-sales relationship (Assumus, Farley, & Lehmann, 1984; Holstius, 1990;

Schreiber, & Appel, 1991; DeKimpe, & Hanssens, 1995a). However, in spite of all the

efforts that have been made to track advertising's effect on sales, it still seems to be an

elusive goal for researchers. Multiple effects exist in advertising and the other

components of the marketing mix. The promotional activities of competitors also easily

confound the measurement of advertising effects on sales (Holstius, 1990).



Difficulties in Measuring Advertising Effects on Sales


One of the major strengths of the evaluation of advertising effects on sales is that

advertising is evaluated in terms of its ultimate objective. However, studies show that

there are some limitations in measuring and predicting advertising effects on sales

(Schreiber, & Appel, 1991; Hoistius, 1990; Lodish et al., 1995). First, although it is true

that the ultimate long-term objective of advertising is to increase sales, it is also

incontestable that advertising frequently has other important short-term objectives, such as

an increase in awareness, knowledge, or recall. It may be true that ultimately all of these

objectives should lead to an increase in sales. However, it is still practically important to

recognize whether these intermediate advertising objectives are achieved or not. The sales

effect measurement approach can not provide the information about the intermediate








objectives of advertising. It should not be underestimated that the effectiveness of each

advertising program element can be evaluated solely in terms of its contribution to sales in

the complexity of the typical advertising efforts. In other words, the measured sales

effects only provide whether or not a particular program element meets the advertising

objectives. However, unlike the intermediate factors, the measured sales effects are not

able to explain why the results occurred and what corrective action is needed.

Second, sales are influenced by so many different factors that it is difficult to

isolate the sales effects provided by one particular element of an advertising program from

those provided by other factors. It is also difficult to measure sales response because of

the lag time between the advertising and the resulting sales. Lodish et al. (1995) indicated

that increased advertising weights increased sales of established brands in only 33 percent

of cases.

Therefore, desired advertising effects are mostly stated in terms of communication

effects in both South Korean and American advertising industries. According to Kreshel,

Lancaster, and Toomey's study (1985), most (79 %) of American advertising practitioners

use communication effects when evaluating advertising media plans. The survey

conducted by the author (1996) also revealed that the majority (90.3%) of the South

Korean advertising practitioners use advertising communication effects to measure or

predict the advertising impact. Only 4.9 percent of media practitioners answered that they

do not use communication effects to predict or measure advertising campaign results.








Advertising Communication Effects


Advertising effects are more often measured on cognitive and attitudinal levels

than sales so that they may be better understood. This cognitive and attitudinal method, of

course, provides the advertiser with more detailed information about the performance of

its communication activities (Holstius, 1990). Alternative measures such as awareness of

a brand or advertising messages delivered, attitude toward the brand advertised,

preference, or purchase intention can be used to evaluate the effects of advertising

campaigns.

Sometimes the purpose of advertising is to create awareness and favorable

attitudes that will ultimately change predispositions to companies and their products.

Most techniques designed to measure this kind of effectiveness are based on theories

about the ways in which consumers process information. These techniques help

advertising planners to determine whether the advertising communication was received

and what the impact was on consumers. For example, if an advertisement successfully

increased peoples' awareness of a product or service, then it would be evaluated to be

effective. In this case, the change in consumer awareness of a product or service is used

to measure advertising effectiveness, and it is presumed that the increase in consumer

awareness ultimately increases sales of the product or service (Cohen, 1988).

The common measures for the measurement of advertising communication effects

are shown in Table 1. Frequently used measures for advertising communication effects

and previous studies are reviewed in the next section.








Table 1. Common measures for advertising communication effects


Stages of Information Common Measures
Process


Cognition Awareness/recognition
Comprehension
Recall

Affect Attitudes toward product
Liking
Preference
Conviction

Conation Intention to try product
Intention to purchase
Actual product purchase





Awareness


Providing either awareness of the brand being advertised or the advertising

message itself is the first level of advertising's communication effects and the first

objective of an advertising campaign. Especially when the product is new or unknown,

the usual objective of the advertising campaign is to increase the consumers' awareness of

the brand being advertised. Therefore, awareness of either the brand or the sales message

is the simplest and the most widely used measure of advertising communication effects

(Woodside, 1996). In a typical measurement of advertising communication effects,

awareness is obtained both for the brand and its advertising message. For an existing and

well-known brand, the usual advertising goal is to develop awareness within the target








market about a specific benefit of the brand or the consumer problem that the brand can

solve. At this stage, the advertiser does not merely attempt to determine whether the

consumers are aware of the brand. If consumers are aware of an advertising message that

communicates specific benefits of the brand, advertising planners may determine that the

advertising achieved its basic objective.

Advertising communication effects also can be measured in terms of advertising

message awareness, instead of brand awareness. For an established brand that is well

known prior to the beginning of an advertising campaign, the awareness of the new

advertising campaign message can be measured at any time after the campaign begins.

The measurement of awareness is generally accomplished on both an unaided and

aided basis. In dealing with the relative utility of measurements between unaided and

aided awareness, Donius (1986) provides a helpful perspective that focuses on the

function of market fragmentation.

If a brand is in a fractionated, highly competitive market, or if it is new, aided
awareness may be the most useful, sensitive measure for that brand. If, on the
other hand, the brand is a household word, dominates its category and/or is in a
highly advertised category with few competitors, spontaneous awareness is
probably going to prove more sensitive and useful.

According to research conducted by Lowe (1984) in the United Kingdom, the

interrelationships among three measures of awareness, such as top-of-mind awareness,

spontaneous awareness, and aided awareness, were examined with over 16 waves of a

tracking study for two leading oil brands. The results suggest that the three measurements

of awareness are, to a considerable extent, measuring the same thing because the three

measurements tend to move together and at much the same rates. The correlation

coefficients of three awareness measurements were over 0.95.








Previous Research on Awareness


To use consumer awareness as a measurement tool of advertising communication

effects, evidence should be presented as to whether consumer awareness of brand or

advertising campaign messages is related to consumers' preference and purchase of the

brand. In a recent study of advertising for box office performance of new film releases

(Zufryden, 1996), the most significant predictors of consumer intention to see a film is

awareness. This study sequentially establishes a connection between advertising

expenditure and consumer awareness, intention to watch the movie and expected ticket

sales at the box office.

According to Gallup research (Mehta & Purvis, 1994) conducted to learn

consumers' top-of-mind awareness levels of competing brand advertising, top-of-mind

awareness is also related to preference. The empirical results showed a strong and

positive relationship between unaided top-of-mind brand awareness and brand preference

as Woodside (1996) stated.

Determining the size of a brand's share of consumers who identify that brand's
advertising in the top-of-mind awareness position may be one useful indicator of
effectiveness of the brand's advertising relative to the advertising of competing
brands.

In Woodside's study, the tests of association indicated that top-of-mind awareness

of an advertising message is related to "brand-mind-position" for three brands tested and

also related to preference for four of the brands investigated. These findings suggest that

a positive relationship may occur between top-of-mind awareness of advertising message

and consumer brand preferences. This study also suggests that advertising exposure leads







to top-of-mind awareness that may affect brand purchase behavior through brand

preference. These schemes are depicted in Figure 1.



1. -1 .
F I- r -



I




i r-






Figure 1. The effects of top-of-mind awareness of the advertisement

Source: adaptedfrom Woodside (1996).


Recall

Advertising planners can use recall to determine whether or not consumers are

exposed to the advertising campaign. Those who are exposed to an advertising campaign

might repeat or play back certain portions of ideas they have seen or heard.

Although recall of sales messages may have some effect on the consumers' future

purchase behavior, the relationship between recall and future consumer purchase seems to

be somewhat tenuous. According to a recent study by Lodish et al. (1995), the








examination of 389 real-world split cable television advertising experiments shows that

recall is not strongly related to sales.

Similar to awareness tests, two types of recall, such as aided recall and unaided

recall, are used in evaluating advertising campaigns. However, since the respondents are

asked only about the product category and are expected to spontaneously remember the

advertising, the unaided recall is believed to be the more useful measure.

When the marketing or advertising manager's objective is to determine the extent

to which consumers have learned or remembered advertising content, "day-after-recall

(DAR)" has been used as a measure of advertising effectiveness (Higie & Sewall, 1991).

DAR is perceived as an important factor to increase a brand's probability of becoming a

member of a consumer's evoked set (Stewart, 1986 and 1989). Therefore, in spite of the

arguments against the use of DAR as a measurement tool of advertising effectiveness,

DAR is frequently employed by advertising managers (Walker & Von Gonten, 1989).



Liking and Attitude Toward Advertisement


Liking assumes that consumers are aware of the brand and have some knowledge

of the brand either from the advertising or from actual product use. Usually, it is difficult

to separate advertising effects on consumer liking from experience, which makes

measurement more difficult. In the liking measure, it is assumed that there is an

advertising effect on the consumer's mental condition or some kind of attitude change.

The consumers move from the awareness or knowledge stage and form a positive

opinion about the product or service. However, liking does not necessarily mean that the

consumer will purchase the product. It simply means that there exist positive feelings or








impressions about the product or service. For example, liking can be measured by having

a consumer name several acceptable brands of products in a category. The assumption is

that if the products are acceptable, they are liked. Although liking is an important step to

evaluate the results of an advertising campaign, the liking measure still does not assume an

actual purchase action. Consumers may like many products but only purchase one or two

(Schultz & Barnes, 1994).

On the other hand, an individual's evaluation of an advertisement, as measured by

his/her attitude toward the advertisement, is thought to be an important mediator of the

advertising effects on brand attitude. Recently, there has been considerable research

conducted to understand the antecedents of the relationship between attitude toward the

advertisement and brand attitudes (Homer, 1990).

Research shows that attitude toward an advertisement is affected by a brand or

non-brand processing set, such as advertising exposure level (Burke & Edell, 1986; Cox &

Cox, 1988), message involvement (Park & Young, 1986; Muehling & Laczniak, 1988),

advertising message quality/content (Hastak & Olson, 1989), and effective responses

generated during advertising exposure (Burke & Edell, 1989; Machleit & Wilson, 1988).

Attitude toward an advertisement has also been observed to be related to the recall

of an advertisement (Pieters & Klerk-Warmerdam, 1996), purchase intention (Mitchel &

Olson, 1981), and attitude toward the act of buying the brand (Mitchel, 1993).

Biehal et al. (1992) suggested that attitude toward the advertisement also had a

direct and positive effect on brand choice. This may be an important issue for advertisers,

who should be concerned that attitude toward advertisements affect not only attitude

toward the brand and purchase intention but also brand choice.








Preference


Preference is defined as one brand being selected over the others among a number

of brands within a certain product category. In advertising campaign evaluation, it is

assumed that the advertised brand is preferred and likely purchased by consumers among

available alternatives, because advertising messages create a level of acceptance for a

specific brand.

Therefore, brand preference is another frequently used measure of advertising

effectiveness and has been of interest to marketers. If the major goal of advertising is to

create product preference, whether or not any actual sales results, the advertising

campaign is usually considered successful if the advertising only creates warm feelings

about the brand. Once a brand is preferred to other brands, it is usually believed that the

advertising or sales messages can do more to trigger the purchase decision. However, the

consumers may or may not purchase the product because there are still many market

factors, such as price, availability, and sales promotion that may influence consumer

buying decisions (Clark, Brock, & Stewart, 1993).

A commercial's ability to affect brand preference usually is measured as the

difference between before and after exposure to the advertisement (Fletcher & Bowers,

1988). However, only a few studies have examined the reliability of preference. In one

study, Stewart and Furse (1986) estimated fairly good r-squared (0.86) for the use of

preference, using test-retest methods.

In another study, Higie and Sewall (1991) suggested that although measuring

preference is more closely related to actual purchase, preference is less reliable than day-








after-recall or other communication effects. They also indicated that it is difficult to

measure significant differences in preference based on a single exposure to an

advertisement.



Purchase Intention and Actual Purchase


If an advertising message is successfully communicated with consumers and other

marketing variables are favorable, the final step of consumer reactions is purchase

intention or actual purchase of the brand. Like other measures explained earlier, it is

assumed that the target audience is exposed and reacts to the advertising campaign

messages. The consumers' awareness of the brand advertised may help develop a liking,

move through preference, and drive to the final step, such as purchasing the brand.

Possible measurement of advertising effects in terms of purchase intention is to identify

advertising as the force that creates a consideration of future brand switching or purchase

(Dutka, 1995).

According to Poiesz and Henry (1994), studies for advertising communication

effects are limited to the investigation of psychological dependent variables, such as

awareness, recall, or attitude toward a brand that are only assumed to precede choice

effects. Although the current availability of product scanner data and other point-of-

purchase systems can provide behavioral data on individual households, purchase

behaviors may be the most difficult concepts to measure because of the many variables

involved, like advertising effects on sales (Poiesz & Henry, 1994).

Nedungadi, Mitchell, and Berger (1993) also argued that although advertising

may have a variety of effects on consumers, an advertising campaign that increases








awareness or improves attitudes is only partially successful if the campaign does not

ultimately influence brand choice. Most evaluation questions on this level deal with what

was purchased or what will be purchased in the future. If change, such as brand

switching, did occur, or if there is an indication of future brand switching, research can be

attempted to relate the behavior change to the advertising exposure.



Achieving Desired Communication Effects


The common measures and previous studies of advertising communication effects

have been explained in this chapter. The desired communication effects can be consumer

awareness of a brand or advertising message, recall, attitude change, preference, or

purchase intention. No matter what communication goals an advertising campaign has,

target consumer's cognitive and attitudinal process is initiated by advertising exposure.

Therefore, one of the fundamental concerns of advertising planning and practice is to

develop and evaluate advertising media schedules in terms of advertising exposure.

The desired communication effects typically can be accomplished when target

audiences are exposed to the advertising campaign with optimal frequencies. For

example, if an advertising campaign goal is to achieve 50 percent awareness of the

advertising campaign message, the advertising planner should estimate how much

repetition can help accomplish the goal. Although no study was attempted to examine

how many times a consumer should be exposed to an advertising campaign to achieve

each level of the communication hierarchy described above, it is assumed that the optimal

frequency may vary with the desired communication objectives established. For example,

conative objectives such as purchase intention may need more consumer exposure than








cognitive goals such as awareness, because the target audience should first be aware of the

existence of a brand with about 3 to 5 exposures to the advertising message and then,

more advertising exposures after awareness may develop purchase intention through

preference and positive attitude toward the brand. It is assumed that consumers need

more information about the brand to reach purchase intention than awareness. The

additional exposures to advertising message will provide more information about a certain

product or service with the target audiences.

Affective and conative measures also may be influenced more by marketing, copy,

and media factors, such as product quality, advertiser image, and message creativity, than

cognitive measures. In spite of enough exposures to advertising messages, consumers

may not reach purchase intention, for instance, if a brand has poor quality and a weak

advertiser image. Table 2 shows the possible relationship between common

communication measures and needed exposures to achieve the goals in a normal

marketing, copy, and media situation. In other words, when ignoring all other influential

conditions such as marketing, copy, and media factors, optimal frequency of consumer

exposure to advertising to achieve cognitive goals may be lower than affective and

conative objectives. For example, if a brand needs consumer exposure to the

advertisement at least three times to gain 50 percent of awareness of the brand, consumers

might need to be exposed six times to achieve preference, and ten times for purchase

intention. In this case, the minimum numbers of consumer exposures to the

advertisement to achieve awareness (a+), preference (b+), and purchase intention (c+) are

3+, 6+, and 10+ in a normal marketing situation. However, until further study identifies a

scientific method to determine the needed exposure frequency for each level of consumer








information processing, it is up to advertising planners to decide how many exposures are

needed to achieve each communication goal established.


Table 2. Relationship between common communication measures and needed exposure
frequency in a normal marketing, copy, and media situation.


Common Measures Minimum Exposure
Frequency Neededfor
Average Consumer

Cognitive Awareness/Recognition
Recall a+

Affective Attitude toward brand
Liking b+
Preference (a < b)

Conative Purchase Intention c+
Actual purchase (a < b < c)


The concepts shown in Table 2 will be integrated with marketing, copy, and media

factors later in this chapter to estimate optimal frequency level for a selected case study.

Since this study is to estimate the optimal frequency of advertising exposure from the

prospective of media planners, the terms used to explain effective exposure to advertising,

such as reach, frequency, effective reach and effective frequency are reviewed to illustrate

what should be achieved to generate the desired advertising communication effects.








Effective Exposure to Advertising



Advertising Exposure


The meaning of exposure is "open eyes or listening ears" facing the advertising

vehicle or message (Sissors & Bumba, 1996). The most common standard to evaluate the

performance of an advertising media vehicle is its rating. Rating can be defined as the

percentage of people exposed to a media vehicle or message and expressed as a

proportion of a selected population base. The rating has been widely used to measure the

performance of media vehicles in terms of advertising vehicle delivery.

However, since audiences exposed to a media vehicle are not automatically

exposed to the actual advertising message, the rating can not indicate what percentage of

the target audience that is exposed to an advertising message within the vehicle.

According to Sissors and Surmanek (1986),

Vehicle exposure represents an opportunity-to-see (OTS). The number of
opportunities-to-see that a media vehicle develops does not guarantee any
exposure to advertising.

Since it is possible to be exposed to the vehicle without exposure to the advertising

message, generally, the vehicle audience is bigger than the advertising audience. In spite

of the difference between vehicle and message exposure, due to the lack of advertising

exposure data and the technical difficulties to measure accurate advertising exposure,

many media practitioners have frequently used vehicle exposure instead of advertising

exposure (Kreshel, Lancaster, & Toomey, 1985).

However, it is certain that the cognitive process explained earlier is initiated by

exposure to an advertising message, not exposure to the vehicle. Therefore, it is the








advertising exposure that should be measured and used for the estimation of media

evaluation factors such as reach, frequency, effective reach, and effective frequency.

In this study, advertising exposure represents exposure to advertising message, not

exposure to the vehicle. The estimation of media evaluation factors is explained based on

the exposure to advertising message in order to measure or predict advertising campaign

effects.



Reach


Recent developments suggest that an important element of the advertising media

decision-making process should be the evaluation of reach and frequency in terms of

advertising exposures or communication effects (Kreshel, Lancaster, & Toomey, 1985).

Reach (sometimes called "coverage") is a measurement of the proportion of the

population that is exposed to an advertising campaign at least once within a given period

of time. Reach is generally explained as a percentage, such as 20, 30, or 40 percent. For

example, if the target market of brand X is 4 million and 2 million of them are exposed to

the brand X's advertisement at least once within a given period of time, the reach would

be 50 percent. However, since reach is accumulated, the duplication should be subtracted.

Suppose only two magazine advertisements were used for brand X's advertising

campaign. If the two advertisements were read by 30 and 20 percent of target audiences,

and 10 percent of them were exposed to both advertisements, the reach of this schedule

would be 40 percent [ (30 + 20) 10 = 40) ].

Reach is usually built in a fairly consistent pattern over time. Using television as

an example, the first time an advertisement is telecast, it accumulates the largest number of








viewers. The second time it is telecast, many of the viewers are repeat viewers. For the

third time, even fewer new viewers may see the advertisement (Sissors & Bumba, 1996).

The shape of a typical reach curve is convex as shown in Figure 2.



Households
Reach %

50
40
30
20
10

Week 1 2 3 4 5



Figure 2. The shape of a typical message reach curve

Source: adaptedfrom Sissors and Bumba (1996).



Frequency and Frequency Distribution


Frequency is a companion statistic to reach and indicates the average number of

times that target audience members are exposed to an advertising message. Whereas

reach is a measure of message dispersion, indicating how widely the advertising message

may be received in a target market, frequency is a measure of repetition, indicating how

many times a target market is exposed, on average, to the advertising message (Sissors &

Bumba, 1996).








For the prediction or measurement of the desired advertising communication

effects, such as awareness, recall, preference, or purchase, these two concepts, reach and

frequency, have been widely used (McDonald, 1995). An interesting example was

provided by a research agency, Millward Brown (Brown, 1993), claiming that the increase

in advertising message awareness can be produced by every 100 TVRs (Television Rating

Points).

The average frequency is calculated by dividing the gross rating points (GRPs) by

reach. For example, if the gross rating points of twenty TV advertisements is 80 and the

percent of the target audience exposed to the advertising schedule at least once (reach) is

40 percent, the average frequency would be 2 (80 + 40 = 2).

A frequency distribution is the percentage of the target audience exposed to each

frequency level of an advertising schedule. Since this study uses advertising message

exposure, instead of vehicle exposure, frequency exposure can show the distribution

patterns of message reach at each frequency level. For example, if an evaluation of an

advertising campaign predicted the frequency distribution shown in Table 3, it is assumed

that 70 percent of the target market is exposed to the advertising message at least once

and 30 percent of the target market is exposed to the advertising message more than two

times with the plan.

To achieve desired communication effects, target audiences need to be exposed to

the advertising message a certain number of times, such as three or four times. Since the

required number of exposures may vary with each advertising campaign or plan, frequency

distributions demonstrate more pertinent information than do individual reach and








frequency when comparing alternative plans. Frequency distributions also provide

planners with a method of determining the pattern of repetition that the plan may provide.



Table 3. Example of frequency distribution


Frequency of Message Exposure Advertising Plan


1 24%
2 16%
3 9%
4-5 6%
6-8 5%
9-12 5%
+13 5%

Average Frequency 4
Total Message Reach 70
Message Reach at 3+ Level 30

Source: adaptedfrom Sissors and Bumba (1996).



Effective Frequency and Reach


In recent years, one of the most significant changes in advertising media planning

and evaluation is the development of the concepts of effective reach and frequency

(Sissors & Bumba, 1996). With these concepts, media planners can understand a number

of important factors that help decide which plan is better among alternatives or how much

repetition is needed to accomplish the expected advertising communication effects.

Effective frequency is the number of advertising exposures needed to accomplish

desired communication effects on the target audience. Since effective reach is defined as

the percentage of the target market exposed to an advertising campaign schedule at the








level of the effective frequency, determining the effective frequency level is one of the

major steps in predicting or measuring advertising campaign effects. Advertising media

planners may decide the effective frequency level needed through test marketing and study

responses to previous advertising schedules. However, sometimes, there is no data by

which to determine objectively how much repetition is necessary to achieve the desired

communication effects. In this case, the planner's estimation may have to be based on

either experience, or specialized research.

Much of what is known about the effects of frequency can be traced to

psychologically trained researchers who explored the subject in a laboratory environment.

One of the major psychologically trained researchers is Krugmen. According to his "three

hit theory" (1972), the first exposure to an advertisement is only perceived by audiences to

raise questions about "What is it?" After the second exposure, the audiences may ask

"What of it?" Then, the second exposure makes audiences react to the advertisement and

begin to compare with alternative brands. The third exposure is a reminder of the other

two exposures and may be the optimal advertising frequency.

Many other studies (Craig & Ghosh, 1993 and Sterling, 1997) have hypothesized

that audiences begin to respond more effectively to advertising from the third exposure to

tenth exposure.

According to Sterling (1997), reach and frequency analysis does not tell us enough

about whether the advertising has been sufficiently exposed to the target audience. The

use of reach (the percentage of target audience members potentially exposed to an

advertising message at least once) is misleading because it often suggests that most any

media plan successfully reaches the vast majority of its intended target audience.








Consequently, Sterling suggests that effective frequency can solve such a problem because

effective frequency is the minimum number of times that target audience members should

be exposed to the advertising campaign in order for the advertising campaign to

accomplish its goal. He said that although the number varies from campaign to campaign,

common levels of message effective frequency are between 3+ and 6+. Threshold effects

tend to occur below three exposures. It means that there will be no or little advertising

effects below the threshold level. An S-shaped response function indicates that there is a

threshold effect.

On the other hand, Naples (1979) suggested that the optimal exposure frequency

should be at least 3+ within a purchase cycle and a "convex-shaped" response curve is

expected after three exposures. He also indicated that advertising copy, market share, and

total media expenses are significant factors to decide exposure frequency levels.

According to him, too many exposures can occur negative responses to the brand.

Sissors and Bumba (1996) also argued that many researchers of advertising

response curves have not observed the S-shape, which is the only shape that can explain

the threshold effect. In fact, "convex-shaped" curves have been frequently found. The

advertising effective frequency level may also vary in response to many other marketing,

copy, and media factors, such as consumer involvement with the advertising message or

product, extent of brand loyalty, product life cycle, purchase cycle, creative execution,

extent of clutter, and other specific market or brand conditions (Ostrow, 1982; Sissors &

Bumba, 1996).








Problems with Using the Concept of Effective Frequency


Although many studies of the concept of effective frequency (Elliott, 1985;

Kreshel, Lancaster, & Toomey, 1985; McDonald, 1995; Leckenby, & Kim, 1994) showed

that effective frequency is widely accepted among media practitioners in the United States,

there have been a number of questions about the use of the effective frequency concept.

The first problem is the differentiation within product categories. Schumann et al.

(1990) indicated that when subjects were exposed to the advertisements only under

conditions of high rather than low involvement, repeated advertising exposure led to more

positive attitudes toward the advertised brand. Recently, a growing number of media

planners suggest that there should be differences among various product category

frequency levels. Although there is some speculation that, for instance, low-involvement

products may need more frequency than high-involvement products, there is still no

practical measurement to differentiate the frequency levels needed for different kinds of

products (Sissors & Bumba, 1996).

Another problem with the concept of effective frequency is whether a "threshold

effect" exists. As Krugmen suggested, many media planners in the US still believe that

there is a threshold effect and that advertising begins to be effective with the third

impression. However, according to Simon and Arndt (1980), most results of studies

concerning the threshold effect showed that the advertising response curves are mostly

convex, not S-shaped. Furthermore, many direct marketing practitioners objected to the

threshold, arguing that there is no answer about the tremendous consumer responses to

their first impressions (Sissors & Bumba, 1996).








The relationship between advertising frequency and its wear-out is another

problem. Studies show that repeated exposure to an advertisement after a certain number

of initial exposures might actually produce negative feelings toward the brand (Calder and

Sternthal 1980 ; Pechmann, & Stewart 1988 ). On the other hand, according to Russell &

Verrill (1986), each additional frequency increases advertising communication effects,

such as recall, recognition, and purchase up to 20 exposures. Therefore, the main problem

that media planners have is when, and under what circumstances, the frequency effects

wear-out. Until further research provides the answer, planners seem to be responsible for

the judgement based on their subjective evaluation.

McDonald's study (1982) also recommended that product-purchase cycle should

be considered to decide the effective frequency level. However, since each customer in a

product category has a totally different purchase cycle, it may not be useful to help

planners decide an effective frequency level. For example, if a product has a purchase

cycle of two months, each customer may purchase the product on different days during

the 60 days, such as the first day of the first month, or the last day of the second month

(Sissors & Bumba, 1996).

There are more problems to be considered. The quality of advertising message

also may affect the effective frequency level because uninteresting copy may require much

more frequency. As explained earlier, the difference between vehicle and message

exposure also may confuse the decision of an effective frequency level. According to

Lancaster et al. (1986), due to the lack of information or technical difficulties, more than

half of advertising practitioners in the United States use vehicle exposure, in spite of their

understanding of the difference between vehicle and message exposure. However, since








the difference between vehicle and message is apparent (Joyce, 1984; Sissors &

Surmanek, 1986; Lancaster & Katz, 1988; Lee, 1996), for the media planners to use the

effective frequency concept, they should focus on how much message frequency is

necessary to achieve the desired advertising exposure. It also might be necessary for

media planners to understand the ratio of message to vehicle exposure.



How to Set Effective Frequency Levels


One of the difficult issues facing any advertising planner is that of deciding how

much effective frequency is enough for an advertising campaign. Since effective frequency

levels vary with marketing, copy, and media factors, those factors should be considered

when choosing an effective frequency level for each advertising campaign.

One suggestion about how to set a frequency level was specified by Ostrow in

1982. According to Ostrow,

The right level of frequency for an advertising campaign is the point at which
effective communication takes place. For example, getting consumers to
understand the advertising message; helping consumers become more positive
about a product or influencing the purchase decisions directly.

He itemized a number of conditions that should be considered based on marketing,

media, and creative strategy factors. For example, new brands need higher frequency

levels than established brands, because new brands need to be learned by the target

market. Planners can use Ostrow's items (see Table 4) by adding or subtracting points

from a base such as 3+ or 4+. The base levels can be chosen according to the

communication goals established. For example, the base may be 3+ for consumer

awareness of a brand while the base may be increased to 6+ to achieve consumer








preference. In Table 4, Sissors and Bumba (1996) suggested specific points with

Ostrow's categorized items. However, these items and points are only suggestions.


Table 4. Marketing, copy, and media factors that affect effective frequency


Marketing factors that affect effective frequency.

Established brands (2)-.1 +.1 +.2 New brands

High market share -.2 -.1 +.1 1 I Low market share

Dominant brand -.2 -. 1 +2 Smaller, less well-known brands
in market

High brand loyalty -.2 -.1 +.1 Low brand loyalty

Long purchase cycle -.2 -.1 +.1 +.2 Short purchase cycle

Products used daily -.2 -.1 +.1 72) Products used occasionally

+.1 ) Needed to beat competition


Copy factors that affect effective frequency


Simple copy

Copy more unique than
competition

Continuing campaign

Product sell copy

Single kind of message

To avoid wearout:
New messages

Large ad units


-.2 +.1 +.2

-.2 +.1 +.2


-.2 -.1 +.1

-.2 -.1 +.1 X

-.l +.1 +.2


-.1 +.1 +.2

0-.l +.1 +.2


Complex copy

Copy less unique than competition


New copy campaign

Image type copy

More different kinds of messages


Old messages

Small ad units








Media factors that affect effective frequency.

Lower ad clutter -.2 U1 +.1 +.2 High ad clutter

Compatible editorial -.2 -.1 +.1 +.2 Non-compatible environment
environment

Attentiveness high -.2 -.1 +.1 +.2 Attentiveness low

Continuous advertising -. +.1 +.2 Pulsed or flighted advertising

Few media used -.2 ( +.1 +.2 Many media used

Opportunities for media )-. +1 +.2 Fewer opportunities
repetition

Source: adaptedfrom Sissors & Bumba (1996) p251-52.
Circled points are applied to the example of brand X described below, while the
remainingfactors are assumed to be neutral.


For example, the effective frequency for a certain brand X can be estimated with the

factors in Table 4. If brand X is an established brand which has a strong market leading

competitor with low brand loyalty and occasionally used, the sum of the scales with

marketing factors will be (+ 0.7). For the copy factors, if brand X has a new advertising

campaign with simple image type copy that is more unique than the competition, single

kind of new message, and large advertising units, the sum of scales applied with copy

factors will be (- 0.4). Finally, media factors can be applied. If the advertising campaign

for brand X is continuous and also has low advertising clutter, high opportunities for media

repetition, and a few advertising media, the sum of scales applied to the example with

media factors will be (- 0.6). When the remaining factors are assumed to be neutral, the

sum of scales applied to brand X with three factors will be (- 0.3), [ (+ 0.7) + (- 0.4) +

(- 0.6) = (- 0.3) ].








Therefore, if the advertising campaign goal is to gain 50 percent consumer

awareness of brand X and the media planner decided the base is 3+, the estimated effective

frequency would be 2.7 or 3+. In this case, the base 3+ indicates that at least three

exposures are needed to accomplish the campaign goal when ignoring marketing, copy, and

media factors.

However, different items or scales may be applied for the estimation of effective

frequency levels as the situation dictates. If the brand X company has a relatively good

image (-. 1) and high advertising message quality (-. 1) but more active competitor

promotional performances (+.2) and relatively poor product or service quality (+.2), the

estimated sum of points will be (-. 1). Table 5 shows the additional points which are

suggested by the author and applied to brand X. It is assumed that the effective frequency

base for advertising message awareness is 3+, then the minimum bases for preference (b+)

and purchase intention (c+) can be estimated as follows:

b > 3 or at least 4+

c > 4 or at least 5+

In addition to Ostrow's items and Sissors and Bumba's point scales, the items suggested

by the author can be applied to the bases of awareness (3+), preference (b+), and purchase

intention (c+). Therefore, the modified effective frequency levels for awareness,

preference, and purchase intention estimation are shown in Table 5.








Table 5. Suggested additional points and modified effective frequency levels for brand X

SuggestedAdditional Points


Good advertiser image -.1 Advertising message quality (high) -.1
Product or service quality +.2 More competitor promotional activities +.2
(relatively poor)



Sum of Points from the Ostrow's items = (- 0.3)
Sum of Points from the additional items = (+ 0.2)

Total Sum of Points = (- 0.1)


Communication Goal Effective Modified Effective
Frequency Frequency Level
Base

Awareness of Brand X 3+ 3.0 0.1 = 2.9 or 3+

Preference b+ 4.0 0.1 = 3.9 or
(b > 3 or at least at least 4+
4+)

Purchase Intention c+ 5.0 0.1 = 4.9 or
(c > 4 or at least at least 5+
5+)



Since one of the main goals of this study is to apply a frequently used American

media evaluation model to the South Korean market, background information on the

South Korean advertising industry will be reviewed in the next chapter. Survey

research conducted by the author and presented in Chapter III also shows how South

Korean media practitioners perceive the current state of reach and frequency

estimation.














CHAPTER III
BACKGROUND ON THE SOUTH KOREAN ADVERTISING INDUSTRY






Introduction


In this study, a frequently used American media evaluation model will be applied to

a South Korean case study. Therefore, it is necessary to understand South Korean

economic development, the growth and recent changes in the advertising industry,

advertising media, and the present state of South Korean advertising. Also in this chapter,

South Korean media practitioners' perceptions of media evaluation terms and concepts,

such as effective frequency and the current state of reach/frequency methods are reviewed.



South Korean Economic Development


The Republic of Korea has become the object of growing international interest

because of its rapid rise to world industrial prominence. The South Korean economic

growth has been driven by the export market, which rose from 20.6 percent of GNP in

1972 to 45 percent by 1988. According to Choi (1994):

The composition of the South Korean exports has changed dramatically from
agricultural and simple labor-intensive goods to increasingly more sophisticated
value-added products, including steel, petrochemicals, automobiles, consumer
electronics, computers and large shipping vessels.








With the dramatic growth of exports, the South Korean per capital GNP has been

increased from $1,012 in 1977 to $4,994 in 1989 (AD DATA, 1995). The total

population of South Korea was 36 million in 1977 and 42 million in 1989. South Korea's

world trade success has not been simply the result of entrepreneurial talent and the

contribution of high quality, but of low-cost labor as well. South Korean economic

development has been planned, directed, and controlled by aggressive government

intervention and financed by government-controlled banks (Koo, 1993).



The Growth of the Advertising Industry in South Korea


In Korea, the first modem advertisement was seen in the "Han-Sung weekly

newspaper" in 1886; it advertised a German trading house which sold products imported

from Germany (Shin, 1986). However, successive political, economic, and social

disorders deterred the development of a capitalistic economic system. Accordingly,

advertising and industrial activity did not fully develop in relation to other specialized

fields and was generally held in lower esteem than other business. The stages of Korean

economic development and the circumstances of consumer life were so different from

those of advanced countries that consumer awareness of advertising was relatively low

(Yoon, 1980).

After the Korean War ended in the 1950s, the Korean advertising agency business

had to start over because the few Japanese agencies operating in Korea during its

occupation (1910-1945) left Korea after World War II (Shin, 1989). The introduction of

television broadcasting, which was initiated by the American corporation RCA, greatly








facilitated the growth of advertising in South Korea (Yoon, 1994). South Korean TV

broadcasting has quantitatively and qualitatively broadened since the introduction.

Along with the influence of the United States on South Korean TV broadcasting,

foreign professionals also initiated the establishment of advertising agencies in the 1960s.

A media representative company, Impact, started its operation as an advertising agency in

1962 (Shin, 1989). The birth of Manbosa Advertising in 1969 led the market entry of

'COKE' and 'Pepsi' and helped to shape agency recognition and the commission system

(Kim, 1996).

The first Korean Advertising agency, Hap-Dong, appeared in 1967 as a technical

"tie-up" with a Japanese advertising agency, McCann-Erickson Hakuhodo. However, it

was a failure. A successful Korean advertising agency did not appear until the Korean

advertising agency, Cheil Communications, started in 1973. With the success of Cheil

Communications, several advertising agencies such as Oricom and Yon-Hap Advertising

began their business in 1973. The progress of Cheil Communication's annual billings,

South Korean per capital GNP, and total expenses for advertising in South Korea

increased drastically as shown in Table 6.

The South Korean advertising industry has grown very fast, primarily as a result of

the democratic reforms of 1987, the development of a free press and the effects of the

1988 Summer Olympics held in Seoul. South Korea has a population of about 44 million,

a per capital GNP of US $10,994 (1996), and spent US $6,184 million on advertising in

1995 (AD DATA, 1995).








Table 6. The progress of the South Korean advertising industry

Year Cheil Billings Per Capita GNP Total Expenses For
(US $) (US $) Advertising
(Us $)

1973 1,100,000 361 28,875,000
1975 5,000,000 532 81,250,000
1983 61,000,000 2,014 706,625,000
1994 575,492,500 8,483 5,355,000,000
1995 750,125,000 10,096 6,184,071,250
(US $1 = 800 Won)



Advertising Media in South Korea


In South Korea, newspaper, television, radio, and magazines are the four major

advertising media. Newspaper advertising has grown very fast for the last 10 years and it

covers approximately 56.4 percent of current total advertising expenses in South Korea.

There are about 90 newspapers that can be used as advertising media. The cost for each

1/3 page advertisement with four major national newspapers is about $900 to $2,400 per

advertisement.

TV is the second most widely used medium. About one third of total advertising

expenses (34.3%) are spent on television advertising. Currently, there are three network

television channels that can be used as advertising media in South Korea. Like the United

States, South Korean television advertising is categorized as program advertising and spot

television advertising. Program advertisements are 15, 20, and 30 second in length. The

cost depends on the popularity of each television program and the time when it is on the








air. In South Korea, most program advertisements are 15 seconds and the cost range is

$8 to $46 per second. On the other hand, spot television advertisements are on the air

between two programs. Most spot television advertisements are 30 seconds and cost

about $60 to $240 per second.

There are approximately 450 magazines that can be used as advertising media in

South Korea (Park, 1995). Among them, women's magazines and news magazines for

men are the most widely used advertising media. The estimated cost for a one page

advertisement ranges from $588 to $2,500 in South Korea. The advertising media, total

expenses and costs in South Korea are summarized in Table 7.


Table 7. South Korean advertising media, total expenses, and costs

Newspapers TVNetworks Magazines Radio Stations


No. of Advertising 90 3 450 7
Media

Total expenses in
1995 (US $) 2,676,323,000 1,627,854,000 220,170,000 216,734,000


Cost Range $588 2,400 Program Ads. $540- Program Ads.
for one 1/3 $8 ~ 46 2,400 $3 ~ 14
page with per second for per second
major national one page.
newspapers Spot TV Ads. Spot Radio
$60 240 Ads.
per second $6 ~ 25
per second

(US 1$= 800 Won)








Recent Changes in the South Korean Advertising Industry


Despite its growing market size, the South Korean advertising industry prohibited

foreign investment until the 1980s. It was the United States that estimated the market

potential and began to address market liberalization. During the 1985 American South

Korean Economic Consultations, liberalization of the South Korean advertising industry

first emerged as a trade issue. In 1986, the American trade representatives became much

more aggressive and then, during the ninth American-South Korean Economic

Consultations in 1988, South Korea announced liberalization of its advertising industry

(Kim, 1990).

Along with the liberalization of the economy and the increased economic power,

South Korean mass media started to be expanded dramatically to satisfy diverse consumer

interests. The number of print media outlets was doubled and the number of weekly

newspapers was increased about five times between 1987 and 1990. Daily newspapers

also increased their editorial and advertising pages (1996, Kim). Broadcasting media

outlets also have increased in number. The new private broadcast network, SBS, opened

in 1991 and emerged as a new media channel.

Another potential medium for advertising space and time is cable television. The

South Korean CATV began service in March 1995 with 15 to 20 channels and plans to

increase the channels up to 40 within the next five years. Therefore, the coming CATV

era in the next five years in South Korea will provide much more opportunities for

advertising (Kim, 1996).








The opening of the South Korean advertising industry to foreign companies also

brought both the establishment of an audit bureau of circulation and the development of

television rating services that were awaited by the South Korean advertising industry. In

1989, although it has been experiencing difficulties, the Korean Audit Bureau of

Circulation (KABC) was finally established and started to organize circulation audits of

print media. In addition to the KABC, the Korean advertising industry also prompted the

need for a television rating system. Along with the establishment of the Korea Survey

Gallup Polls (KSG) in 1990, three additional media-rating companies began operating

such as Media Service Korea (MSK), Lee's PR, and Korean Marketing Research (KMR)

(Park, 1995).

The opening of the South Korean advertising industry also prompted advertising

agencies to turn to market research and audience studies. Since the leading South Korean

advertising agency, Cheil Communications, established the first marketing research

institute in 1991, other major large agencies, such as LGAd, Daehong, KeumKang, Korad,

Oricom, and Seoul DMB&B opened their research institutes (Kim, 1996).



Studies on Measuring and Forecasting Advertising Effects in South Korea


Although the studies on measuring and forecasting advertising effects have been

active in the advanced countries, such as the United States and Japan, since the early

1900s, there has been little systematic study of evaluating advertising effects in South

Korea. It is assumed that some South Korean advertising agencies have developed their

own advertising evaluation models that are applicable to the South Korean advertising








environment since the late 1980s. However, those models are proprietary and therefore

are not available for access from outside of the agencies (Kim, 1996).

In South Korea, the first study in measuring and forecasting advertising effects was

conducted by Sung in 1989. In his study, "The establishment of a measurable model for

pre-estimation and evaluation of advertising effects in television commercials," Sung tried

to develop a model of predicting and measuring advertising effects that is applicable to the

South Korean environment. Based on the Japanese JNN Data Bank (JDB) model, as

shown in Figure 3, he suggested that GRPs or exposure to advertisements lead to

awareness that may affect brand purchase through brand preference. Using regression

analysis, he examined the relationship between GRPs and advertising exposure, GRPs and

brand awareness, awareness and preference, preference and purchase, awareness and

purchase, and GRPs and purchase. The results of this study showed that all these six

relationships are statistically significant and different regression equations are suggested.

The independent and dependent variables used in his study and regression equations are

shown in Table 8.

However, his study identified problems for future studies. Sung assumed that all

the other intermediate factors, such as competitors' promotional activities and quality of

advertising message that may mediate the relationship between independent and dependent

variables, are the same for all brands studied. His model also may not be applicable to

established brands because he used only new products.

As explained earlier, although some South Korean advertising agencies have been

trying to develop their own models to evaluate and predict advertising effects, most of

them are not published. However, in 1991, LGAd, which is one of the top five advertising








agencies in South Korea, developed an advertising effectiveness measuring model (AEM)

and the model was introduced by Lee and Park (1995). In addition to brand awareness,

preference, and purchase, this model also considers advertising message awareness and

advertising preference to predict advertising communication effects (Lee & Park, 1995).







GRPs Brand Preference Purchase
Awareness


-_F I ___

Exposure to
Advertisements.



Figure 3. Sung's model to estimate advertising effects



Table 8. Sung's independent and dependent variables and regression equations

Independent Dependent Regression Equations
Variables Variables for Food Products



GRPs Exposure to Ads. Y = 100 61.207128exp(-0.000125065X)
GRPs Brand Awareness Y = 100- 55.473exp(-0.0000135X)
Brand Awareness Brand Preference Y = 2.37916exp(0.032628345X)
Brand Preference Brand Purchase Y = 1.185111X + 1.85653
Brand Awareness Brand Purchase Y = 2.35262279exp(0.035743144X)
GRPs Brand Purchase Y = 100 95.2616282exp(-0.000091683X)








Although the South Korean advertising industry has been growing very fast, the

evaluation of advertising effects is still in its early stage. Therefore, it is necessary that a

new measurable model of advertising effects that fits to advertising environment of South

Korea should be developed to predict and measure advertising effects. From the two

previous Korean studies, it is also assumed that awareness, preference, and purchase or

purchase intention are the most widely used measures to predict or evaluate advertising

communication effects in South Korea.



Media Planning and Evaluation Procedures of South Korean Advertising Practitioners


Chapter II introduced concepts and studies dealing with the aspects of media

planning, such as the kind of communication effects that can be used to predict or measure

advertising effects, how effective reach is typically measured to achieve communication

effects, and the minimum and maximum levels of frequency required for target market

members to be effectively reached.

Studies of the development of media plans, media directors' perceptions of

effective reach/frequency, and the consideration of weighting and timing factors have been

frequently conducted in the US (Leckenby & Kishi, 1982b; Leckenby & Boyd, 1984;

Kreshel, Lancaster, & Toomey, 1985; Lancaster, Kreshel, & Harris, 1986; Leckenby, &

Kim 1994). However, in spite of the fast growth of the South Korean advertising

industry, no such research has been conducted in South Korea. Therefore, prior to the

application of a frequently used American model to the South Korean market, it may be

important to examine how media directors of South Korean advertising agencies view the








current state of reach/frequency estimation methods and which communication effects are

used to predict or measure potential advertising impact.



Purpose of the Study


No information was available about how the South Korean advertising industry

perceives reach/frequency estimation methods and how these concepts are used by the

South Korean advertising agencies. Consequently, this study shows the South Korean

advertising industry's 1) vehicle/message rating sources, 2) perceptions of the difference

between vehicle and message, 3) definition of reach and effective reach, 4) lower and

upper limit of effective reach, 5) and the use of communication effects, among other

factors.

In this study, the survey results of South Korean media practitioners also will be

compared with previous American studies to inspect similarities and differences between

the two advertising industries. The American studies, which are partially available for

comparison, are Kreshel, Lancaster, and Toomey (1985), Leckenby and Kim (1993), and

Falcheck (1995).



Methods


Currently there are about 120 advertising agencies that can deal with both

broadcasting and print advertising in South Korea. In this study, a mail survey was

conducted of the media directors of these 120 advertising agencies.








The questionnaire used in this study was adapted from the one published in 1986

by Lancaster, Kreshel, and Harris. The questions were modified in accordance with the

South Korean situation, such as vehicle rating sources, and translated into Korean by the

author. Both Korean and English versions of the questionnaire are shown in Appendix A.

The questionnaire was mailed to the media director of each agency, if there was any, with

a personalized cover letter, return envelope, and a return postage stamp for international

mail. For the agencies which do not have media directors, a questionnaire was mailed to

senior executives or owners of the agencies.

A follow-up letter, questionnaire, and return envelope were mailed to all non-

respondents approximately seven weeks after the initial mailing. After two mailings, a

total of 61 responses were received. SPSS PC+ for Windows was used to analyze the

data.



Results


About half(61) of 120 South Korean advertising agencies responded to this study

and 43 of them answered that they develop advertising media plans.

For the titles of the respondents, 36 percent were media directors or vice directors,

3.3 percent were executives, and 32.8 percent were others, not directors or vice-directors

but individuals who work for either media planning or media purchases, and owners of the

agencies. Approximately 28 percent of respondents did not provide their titles.

For the first question regarding sources for television program ratings, 57.4

percent of the South Korean advertising agencies receive the data from Media Service

Korea (MSK). Other sources used in the South Korean advertising industry were Lee's








PR (23%), and Korean Marketing Research (KMR) (26.2%). Since some of the agencies

receive the rating data from more than one source, the sum of percentages for each source

exceeds 100 percent.

For the TV advertising message ratings, 53.5 percent of the South Korean

advertising agencies also use MSK data. Lee's PR data are used by 8.3 percent of

agencies and 24.6 percent of agencies do not receive advertising message ratings.

This study also shows that the majority (70.5%) of the South Korean advertising

agencies develop media plans for their advertising campaigns. Among the agencies that

develop media plans, the most frequently used systems are non-computerized programs

(39.5%) and in-house computer programs (32.6%) (see Table 9).


Table 9. Systems used to develop advertising media plans (N = 43)

System Used (%)


In-house computer program 32.6

Foreign computer program 4.7

Non-computerized program 39.5

Others or no answer 23.3


For the next item examining the factors which were generally used to evaluate

advertising media plans, most of the South Korean advertising agencies that develop

media plans answered that they use reach (79.1%) and GRPs (Gross Rating Points)

(72.1%). Other frequently used factors are average frequency (79.1%), CPM (67.4%),








and CPRP (72.1%). Two previous American studies (Kreshel, Lancaster, & Toomey,

1985; Leckenby & Kim, 1994) also revealed that reach, GRPs, average frequency, and

CPM are the most widely used factors to evaluate media plans. Table 10 illustrates the

factors used in evaluating media plans in the South Korean advertising industry in

comparison with two American studies.


Table 10. Factors used in evaluating media plans


Factors US 85 US 94 Korea 96
(N =91) (N= 51) (N=43)
(%) (%) (%)

Reach 90.4 81.0 79.1

Average Frequency 87.2 73.0 79.1

GRPs (Gross Rating Points) 89.4 73.0 72.1

CPRP (Cost-Per-Rating Point) N/A N/A 72.1

Effective Reach 86.2 68.3 67.4

CPM (Cost-Per-Thousand) 88.3 77.8 67.4

Gross Impressions N/A N/A 2.3

Frequency Distributions 75.5 74.6 N/A

Quintile Distribution 43.6 58.7 N/A
N/A indicates that the factors are not used or no answer.



The evident differences discovered between South Korea and the United States are

the use of CPRP and Frequency Distribution. In spite of little or no use by American

practitioners, the majority (72.1%) of South Korean media practitioners use CPRP to








evaluate media plans. On the contrary, frequency distributions are highly used by only

American media practitioners. In other words, the introduction of frequency distributions,

which can provide planners with more pertinent information and a method of determining

the pattern of repetition, can help the South Korean media directors to evaluate their

media plans more effectively.


Definition of Reach and Effective Reach

Although the majority (91.8%) of the South Korean advertising practitioners

answered that the distinction between advertising vehicle and message is important for the

evaluation of advertising campaigns, only 39.1 percent answered that the most

representative definition of 'reach' is message exposure. Nearly half of the South Korean

advertising practitioners (53.5%) defined 'reach' as vehicle exposure. In other words,

53.5 percent of the South Korean advertising agencies surveyed use vehicle audiences as

a proxy of advertising audiences, while nearly 40 percent of the agencies attempt to

estimate advertising message exposure.

The definition of effective reach applied in the media plan of a particular product

or service can vary across a number of factors, including the media categories and time

frames used. When multiple media categories are evaluated simultaneously, the same

definition of effective reach and time frame must be used for all media categories. This

study found that only half (46.5%) of the media plans developed by the South Korean

advertising agencies include evaluations of schedule reach, frequency, and GRPs for

combinations of two or more media categories.








The South Korean advertising practitioners were then asked to note the general

definition of effective reach. The majority of the South Korean advertising practitioners

recognize that the definition of effective reach requires a minimum frequency level. About

93 percent answered that a minimum frequency of two or more should be the lower limit

of effective reach and 74.4 percent answered it requires a frequency of three or more,

while 11.6 percent answered that a frequency of four or more is needed.

For the upper frequency limit, 41.9 percent of the respondents recognize no upper

frequency limit to effective reach, while 37.2 percent set some kind of upper limit, ranging

from 9 to 11. In spite of the different environmental situation, two American studies

(Kreshel, Lancaster, & Toomey, 1985; Falcheck, 1995) also show similar results to this

study. Table 11 shows the South Korean results in comparison with the previous

American studies.


Communication Effects Used

The majority (90.3%) of the South Korean advertising practitioners answered that

they use communication effects to evaluate their advertising media plans. The two most

frequently used variables are exposure to advertising and awareness/recognition. Among

those who use communication effects to evaluate media plans, 80 percent answered that

they use advertising exposure. The second most frequently used variable is awareness or

recognition (46.5%). Since there is no distinct difference between awareness and

recognition in the Korean language, these two terms are treated as the same expression.








Table 11. Representative definitions of effective reach in comparison with two previous
American studies


US 85 US 95 Korea 96
Message Impact Required us 85 Uo
Message Impact Required (N = 91) (N = 50) (N = 43)
(%J (%) (%)

Vehicle Exposure 48.4 70.0 53.5
Message Exposure 31.9 13.3 39.5
Advertising Impact 16.5 13.3 7.0
Others and no answer 3.3 3.3


Number of Exposures Required Regardless of Message Impact Level
Lower Limit US 85 US 95 Korea 96
(%) (%) (%)

1+ 6.6 13.3 4.7
2+ 5.5 6.7 18.6
3+ 61.5 50.0 62.8
4+ 17.6 16.7 11.6
Others and no answer 8.8 3.3 2.3

Upper Limit US 85 US 95 Korea 96
(%) (%) (%)

No Upper Limit 58.2 46.7 41.9
9 4.4 6.7 4.7
10 6.6 10.0 20.9
11 2.2 0.0 11.6
Others and no answer 28.6 13.3 21.0








Table 12 illustrates the variables used to evaluate advertising communication

effects in the South Korean advertising industry in comparison with a previous American

study (Kreshel, Lancaster, & Toomey, 1985). Though the American study was conducted

about ten years before this study, the two studies show similar results. The only

observable differences are that Korean agencies use attitude toward advertisement or

brand more frequently than American agencies, while American agencies use recall more

than Korean agencies.


Table 12. Variables used to evaluate advertising communication effects

US 85 Korea 96 US 85 Korea 96
Variable (N=94) (N=61) Variable (N=94) (N=61)
(%) (%) (%) (%)

Exposure to 52.1 79.1 Preference 18.1 32.6
advertisement

Awareness/ 47.9 46.5 Willingness to 12.8 32.6
Recognition 24.5 purchase

Attitude toward 11.7 44.2 Purchase 31.9 23.3
Ads.

Recall 58.5 37 2 Comprehension 16.0 20.9

Interest 16.0 34.9 Conviction 8.5 14.0

Attitude toward 18.1 32.6 Others 9.6 4.7
brand




Message Weighting and Reasons of Not Weighting

The media practitioners of the South Korean advertising agencies that develop

media plans and evaluate their plans (about 70 percent of all South Korean agencies) were








asked whether they use such information to weight vehicle audience data to estimate

advertising message exposures or other desired communication effects.

Among 43 agencies that develop media plans, 34.9 percent indicated that the data

are used to weight vehicle audiences. More than half (56.3%) of the agencies which

weight audience data use individual judgements to derive weights, while only 18.8%

derive weights from a formula or an established standard.

The respondents who do not weight vehicle audiences gave approximately six

different reasons for not doing so. The most common of these are that media planning

situation for each brand or service is unique and weighting is too difficult to use

accurately. Lack of information about weighting procedures and not enough time are two

other common answers.


Needed Improvements

Almost all respondents (96.7%) answered that their agency's current procedures

for evaluating advertising media plans are in need of improvement.

The majority (82%) of South Korean advertising practitioners indicated that

recognition of the importance of media planning should be improved. Nearly 80 percent

of them also answered that training of media specialists is the second most important thing

to improve current procedures to develop and evaluate advertising media plans (78.7%).

The South Korean advertising media practitioners showed split responses to the

introduction of foreign media planning procedures, such as American reach and frequency

computer programs, for better media planning. About 34.9 percent answered that the








foreign procedures would be helpful for better media planning, while 32.6 percent replied

that the foreign procedures would not be helpful due to the different media environment.



Discussion and Implication for Further Research


In the United States, models for estimating reach and frequency distributions have

made increasing use of the concept of effective reach and frequency among both

practitioners and academicians (Kreshel, Lancaster, & Toomey, 1985; Lancaster, Pelati, &

Cho, 1991; Leckenby, & Kim, 1994). The results of this study clearly demonstrate that

the majority of the South Korean advertising agencies are also aware of the media

evaluation terms such as effective frequency, and more than half of the agencies actually

use reach/frequency estimation methods.

In spite of the different environments, the South Korean media practitioners'

representative definition of effective reach is also close to the American practitioners as

shown in Table 11. The majority of both American and South Korean media practitioners

answered that the most representative definition of effective reach has 3+ as the lower

limit and no upper limit.

In addition to the definition of effective reach, the majority of both American and

South Korean media practitioners also use communication effects to evaluate advertising

campaign effects and media plans. Among the common measures for advertising

communication effects, advertising exposure is the most frequently used measure in both

countries. Awareness is the second most widely used measure in South Korea and third in

the United States.








With the beginning of CATV in South Korea since 1995, the South Korean

advertising industry started to recognize the importance of media planning and evaluation

of media schedules. According to this study, about 96.7 percent of the South Korean

advertising agencies claim that their current procedures for developing and evaluating

media plans should be improved. The South Korean media directors also indicate that

systematic media evaluation methods that fit the South Korean environment are needed in

the near future.

Although only 34.9 percent of the South Korean media practitioners responded

that the introduction of foreign procedures would be helpful for better media planning and

campaign evaluation, 32.6 percent replied that the procedures would not be helpful. The

application of a frequently used American reach and frequency computer program to the

South Korean market will help develop a model that fits the advertising environment of

South Korea.

This study indicates that reach/frequency estimation is a good method to predict or

evaluate advertising communication effects in the South Korean advertising industry

because of the wide acceptance and actual use of reach and frequency as shown in Table

10. Furthermore, since the South Korean media practitioners also define the general

effective frequency level similar to American media directors as shown in Table 11, it is

assumed that in South Korea the effective frequency levels for common communication

measures, such as awareness, preference, and purchase intention, can be decided using the

procedures explained in Chapter II. Based on the optimal exposures for common

communication measures, which are used as bases to estimate the advertising





63


communication effects, the effective frequency level can be decided for each advertising

campaign goal with consideration of marketing, copy, and media factors.

A test market study is a good method for examining these assumptions.

Therefore, Chapter IV, Methodology, will introduce the brand used as a case study, along

with the normative theories, methods, and procedures required to evaluate it.














CHAPTER IV
METHODOLOGY





The primary purpose of this study is to demonstrate how accurately media

exposure distribution models can predict advertising communication effects. In particular,

one goal is to apply such a model to the South Korean market. In this test market study,

advertising campaign communication effects are "predicted" by a variation of the beta

binomial exposure distribution model based on the information available before the

advertising execution. Then, the impact of the advertising campaign on its target audience

is measured after a three-month execution. The predicted communication effects will be

compared with the actual target consumer responses to the advertising campaign to probe

the capability of this normative framework.

Generally, as shown in Table 13, regardless of the country or brand being studied,

particular items need to be gathered or defined to complete this type of study. These

items are 1) target audience definition, 2) vehicle data, 3) message data, 4) advertising

expenditures, 5) a media evaluation model, 6) a narrow time frame, 7) tracking study

results, 8) and the media plan used.

Advertising planners who wish to accomplish this type of research should follow

these procedures. After defining the target audience, advertising message and vehicle

ratings and expenditures should be gathered prior to the period of the campaign in order








to predict the campaign effects. It is also important to evaluate the plans over a narrow

time frame, such as monthly, due to target audience forgetting, competitive advertising,

and other factors that reduce advertising carry-over effects. For example, annual reach-

frequency analyses will grossly overstate likely communication effects because it fails to

account for these factors (Lancaster, Kreshel, & Harris, 1986).


Table 13. Normative framework necessary to predict advertising communication effects

Items General Meaning


Target Audience The desired or intended audience for an advertising campaign.
Definition Usually defined in terms of specific demographic, purchase, or
ownership characteristics.

Vehicle Data Ratings of particular component of a media class, such as a
particular TV program or magazine, used for an advertising
campaign.

Message Data Ratings for actual advertisements, not the vehicles.

Advertising Total media expenses for an advertising campaign.
Expenditures

Media Evaluation Computer program that embodies normative theories of media
Model planning, including vehicle and message ratings and duplication
plus corresponding exposure distributions.

Narrow Time Due to target audience forgetting, competitive advertising, and
Frame other factors that reduce advertising carry-over effects, a narrow
time frame for analysis, such as monthly, is preferred.

Tracking Study Actual target audience response to the advertising campaign in
Results terms of the advertising goals established.

Media Plan Used Advertising media schedules during the period of the campaign.








A media evaluation model such as the beta binomial exposure distribution model

can be used to predict the campaign communication effects. Then, consumer tracking

studies will show actual consumer response to the campaign.

In this chapter, after introducing the brand and advertising campaign used in this

case study, each item in Table 13 will be explained again with reference to the information

available for the case.



Case Study




ASIANA Airline


A South Korean airline company is selected to test the framework of predicting

and measuring advertising campaign effects introduced above. The South Korean

domestic and international air transportation business has grown rapidly (about a 12

percent annual increase in passenger transportation volume) since Korea hosted the 1988

Olympics. Seoul's growing importance in international air traffic, good international

geographic location, congested traffic on the highways, and drastic increase in GNP (see

Table 6 in Chapter III) since 1988 accelerated the growth of the airline business in South

Korea.

Currently, there are two airline companies in South Korea. The Korean airline

(KAL) had been the only air carrier in South Korea until ASIANA started its business in

1990. In 1993, both KAL and ASIANA carried 15,655,000 passengers on their domestic

lines and 11,651,000 on international flights. Cargo services also have grown impressively








since the opening of American routes in 1971. The number of employees of ASIANA

airline is about 7,900.

Domestically, KAL has maintained a 70 percent market share since 1990, while

ASIANA has held only 30 percent market share for the last 7 years. Internationally, the

market share of both Korean air carriers was only about 50 percent until 1991 because of

dynamic foreign airline companies, such as Delta Airlines, TWA, United Airlines, British

Airways, and Singapore Airline. Many foreign airlines have been attracted to Seoul's

growing importance and have opened regular services to and from Korea. As of 1993,

about 26 foreign airlines are operating in Korea. However, the international market share

of both Korean air carriers has been increasing since 1991. While the KAL holds its

international market share at about 47 percent, ASIANA airline has been growing fast for

the last 5 years. The domestic and international market shares for both air carriers are

shown in Table 14.

Current consumer research conducted by SangAm advertising agency shows that

the ASIANA airline company has maintained a good company image on its service

evaluation because of perceived flight attendant service and kindness. However, many

consumers still favor KAL because of safety, tradition, and mileage bonuses. It is

assumed that promotional activities such as advertising and public relations play a major

role in the recent successful growth of ASIANA. Since advertising is an important means

of promotion for ASIANA airline, it is necessary to conduct consumer research to

examine and predict its current advertising effects and to help develop a more efficient

advertising campaign in the future. This case study also provides an opportunity to test

and advance normative advertising theory.








Table 14. Domestic and international market shares

Domestic market share


1990 1991 1992 1993 1994 1995 1996
(%) (%) (%) (%) (%) (%) (%)

Korean Airline 70 71 71 70 69 69 69

ASIANA Airline 30 29 29 30 31 31 31

Total 100 100 100 100 100 100 100



International market share


1989 1990 1991 1992 1993 1994 1995 1996
(%) (%) (%) (%) (%) (%) (%) (%)

Korean 52.3 45.3 44.8 43.8 47.3 47.4 47.1 46.9
Airline

ASIANA 2.0 6.0 8.9 12.1 15.3 17.4 17.8
Airline

Total 52.3 47.3 50.8 52.7 59.4 62.7 64.5 64.7

Source: adaptedfrom "SangAm advertising agency.


Target Audience and Advertising Campaign


The annual expense for ASIANA advertising was about US $5,920,000 in 1994

and ASIANA is one of the 150 biggest advertisers in South Korea (ranked 101 in 1994 in

'Kwango Yeonkam 95').








In 1996, ASIANA airline completed its new advertising campaign, "Her name is

ASIANA," focusing on ASIANA's better service and kindness. Since the new campaign

was performed by SangAm, one of the major Korean advertising agencies, most of the

information needed for this study, such as consumer survey data, were supplied by

SangAm.

According to the research conducted by SangAm (1996), the primary target

market of this campaign was defined as 20 to 40 year old men who have used any air

carrier at least once. However, there are some difficulties in using this target market

definition. First, previous use of any air carrier makes it difficult to estimate the described

target market size. Second, the geographic area of consumer research conducted by

SangAm was limited to Seoul, the capital city of Korea, where one fourth of South

Korean people live. Consequently, although the primary target audience is defined as 20

to 40 year old men who have used any air carrier at least once, the consumer survey was

conducted with 20 to 40 year old Korean men who live in Seoul. The target market size

used for the consumer survey is about 3 million (2,884,983 in 1995).

ASIANA airline first advertised its new campaign of "Her name is ASIANA" from

February 17 to May 16 in 1996. Two different types of messages were advertised on

television and in magazines. One was called "piano scene" and the other was called

"poetry situation." Both types of the actual magazine advertisements are in Appendix B.

The summary of the advertising insertions by media category and expenses of the

campaign is shown in Table 15.








Table 15. Advertising insertions by media category and expenses of "Her Name is
ASIANA" campaign from February 17 to May 16, 1996

Media Insertions Total Expenses
(US $)

KBS-2TV 96 $414,373
MBC-TV 137 477,035
SBS-TV 105 259,441


Weekly Magazines 33 60,750
Monthly Magazines 9 19,875


Grand Total 380 $1,231,474
Source: adaptedfrom marketing department of SangAm and MSK.
(S $1 = 800 Won)


Competitor


The market leader, KAL, which started its business in 1969, is a strong competitor.

The first name of KAL was Korean National Airlines and was managed by the

government. Later the name was changed to KAL and the management was turned over

to private hands. Because of the long tradition and previous monopoly, KAL has

maintained its market share at about 70 percent domestically and 47 percent

internationally.

However, compared with ASIANA's rapid growth, the market share growth rate

of KAL has been almost zero for the last four years (see Table 14). KAL's annual

advertising expense was US $11,932,000 in 1994 and KAL has been one of the 50 biggest

advertisers in South Korea (ranked 39 in 1994) (Kwango Yeonkam 95). KAL's








advertising expenses were approximately three times higher than ASIANA during the

period of ASIANA'S first "Her name is ASIANA" campaign. Overall advertising GRPs

and expenses from February through April of 1996 are provided in Table 16.


Table 16. Competitor GRPs and advertising expenses

Media Month GRPs Expenses (US S)


TV February 969.58 $499,130
March 1158.18 549,560
April 1072.82 524,167
Total 3200.58 1,572,857

Radio February 272.89 90,395
March 273.54 97,407
April 322.18 93,980
Total 868.61 281,782

Newspaper February 742.62 889,327
March 236.86 309,484
April 261.58 267,725
Total 1241.06 1,466,536

Magazine February 89.73 63,625
March 82.10 46,813
April 45.63 35,500
Total 217.46 145,938

Total February 2074.82 1,542,477
March 1750.68 1,003,264
April 1702.21 921,372

Grand Total 5527.71 $3,467,113
Source: adaptedfrom MSK (Media Service Korea).
(US$1 800 Won).








Media Data



TV Ratings

Television program ratings are collected from Media Service Korea Inc. (MSK)

which is one of the major media research institutes for ratings in South Korea. SangAm,

the advertising agency for ASIANA airline, provides advertising message ratings that are

also collected from MSK. Media Service Korea Inc., founded in 1987, is the only media

research institute in South Korea that uses people-meters to collect its ratings from a total

of 1,200 panel members who live in Seoul. Audience panels were established by MSK and

the panel members were randomly selected based on four critical demographic variables,

such as age, occupation, education, and income. Each panel member's daily television

watching data are gathered in the MSK central computer from 6 A.M to 1 A.M and all the

television program and advertising message ratings are reported daily after aggregating

them.

The ratings are also provided by important variables, such as gender and age. It is

possible to gather demographic information with the use of separate remote controls

distributed by MSK. Each panel member is assigned to a number and is supposed to input

the assigned number while he/she is watching television. For example, if a panel

household consists of four family members, such as a father, mother, son and daughter,

each family member is assigned to a number from 1 to 4 and is instructed to push the

number button on the remote control. Therefore, in this study, program and advertising

message ratings were gathered for the target audience group. According to MSK, these

ratings are significant with 95 percent confidence levels (p < 0.05). The average sampling








error is within + 0.025. Television program ratings, advertising message ratings, and costs

provided by the 'SangAm' agency are listed in Appendix C.

Television program and advertising message ratings are also collected from Lee's

PR which is another popular research institute for ratings in South Korea. These are often

used in comparison to MSK. While MSK uses panel members and people-meters, Lee's

PR conducts face-to-face interviews once a month to gather television ratings from a total

of 1,300 sample members. However, since MSK data have been used by SangAm, MSK

data are used in this study unless otherwise noted. According to the South Korean media

practitioner survey conducted by the author and reported in Chapter III, MSK is also the

most frequently used vehicle and message source in South Korea.


Magazine Ratings

Magazine advertising expenses of $80,625 were 6.5 percent of total media

expenses of $1,231,474 for the ASIANA campaign. Eleven weekly magazines and three

monthly magazines were used for the ASIANA advertising campaign. The circulation and

ratings for these magazines were collected from each magazine company because of the

lack of independent information about magazine ratings in South Korea. A Korean

advertising practitioner called each magazine company and collected the ratings of these

14 magazines. The ratings may be higher than actual ratings because they were gathered

from each magazine publication company. In the future, the magazine ratings should be

gathered from an objective media research institute for better prediction of advertising

campaign effects. Magazine circulation, ratings, and costs are listed in Appendix C.








Tracking Data


The tracking data used in this study were gathered from the advertising agency of

ASIANA Airline, SangAm. The main purpose of the survey research was to evaluate the

impact of the new ASIANA advertising campaign, "Her name is ASIANA" and to prepare

the next ASIANA advertising campaign After the three month advertising campaign from

February 17 to May 16, a post test with a total of 200 samples of the target audience was

conducted in May, 1996.


Sampling

Since the target audience was defined as 20 to 40 year old men who live in Seoul,

Korea, the sample was randomly drawn among the target audience. In this survey

research, "stratified sampling" was used, which separates samples from each of several

subpopulations. The subpopulations were defined by SangAm in advance on the basis of a

critical subject variable, such as age, that may influence scores on the dependent measures.

Rather than relying on random sampling, the target population was divided into

subpopulations on the basis of age to create a total sample by selecting the appropriate

proportion of subjects from each of the subpopulations. For example, if 20 percent of the

target population is between ages 20 and 29, the number of subjects that would represent

20 percent of the total sample was selected from that subpopulation. A total sample of

200 individuals was drawn from three different age groups. Forty percent of the total

sample was selected from the 30s age subpopulation, and another forty percent was drawn

from the 40s age subgroup. The age groups, number of sample individuals in each age

group, and sample characteristics are shown in Table 17.








Table 17. Sample characteristics

Number in (%)
Sample
Age 20-29 40 20.0
30 39 80 40.0
40 49 80 40.0
Total 200 100.0

Education Up to High School Education 43 18.5
College Education or More 157 81.5
Total 200 100.0

Marital Status Single 37 21.5
Married 163 78.5
Total 200 100.0

Occupation Specialist or Freelancer 19 9.5
Management Staff 42 21.0
Office Worker 110 55.0
Self-employed 29 14.5
Total 200 100.0
Source: adaptedfrom the marketing department of SangAm.


Measurement

Since the purpose of this survey was to evaluate the new ASIANA advertising

campaign effects, only a post test was conducted. In this survey, a face-to-face interview

was used to gather target audience information about the advertiser image, advertising

campaign message awareness, evaluation of the advertising message quality, preference,

and willingness to use in comparison with competitors. A total of 18 interviewers, who

were college students or housewives, were temporarily hired by the advertising agency of

ASIANA Airline to conduct the personal interviews for seven days beginning on May 25,

1996. Three questions that were asked of the respondents and which are directly related

to this study are: 1) "Are you aware of the new ASIANA advertisement, 'Her name is








ASIANA'?" 2) "Which airline company do you prefer?" 3) "Which airline are you going

to use for your next trip?" After data were completely collected, 20 percent of randomly

selected responses were verified again before editing, coding, and analyzing the data.

Then, the SPSS PC+ program was used to show frequencies and cross-tabulations. These

tracking data will be used to examine how accurately the normative media evaluation

framework of this study can predict advertising communication effects.



Media Evaluation Model Used in This Study



Beta Binomial Exposure Distribution Model

A variety of computerized media evaluation models have been developed and used

in both the advertising industry and academia over the past thirty years. The availability of

computers in the 1960s led to the development of a number of mathematical media

planning models. The first level of sophistication in the quantitative evaluation of a media

plan involves estimating a media schedule's reach and frequency. Another related concept

is gross rating points (GRPs), which refer to the total number of exposures generated by

the schedule. In 1961, Agostini developed a simple reach estimation model and in 1964,

Metheringham tried to estimate unduplicated audience measurement (Craig & Ghosh,

1986). They provided the impetus for the development of media models. After that,

media evaluation models have become much more sophisticated, including the beta

binomial models, beta matrix method, heuristic models, and variations or extensions of the

beta binomial models (Rust, 1986).








Most recent models deal with the sophisticated calculation of the exposure

distribution by the number of exposures of a given media schedule and they provide all

information relevant to the advertising media schedule. The general goal of these models

is to evaluate advertising media schedules by estimating reach, frequency, effective

frequency, effective reach, and exposure distributions.

Exposure distribution models are divided into two groups in terms of estimation

dimensions that include a number of parameters, such as univariate probability distribution

models and multivariate probability distribution models. Univariate probability distribution

models treat all vehicles as one univariate and composite vehicle and then average all

vehicle insertions of a schedule (Ju et al., 1990). On the other hand, multivariate exposure

distribution models provide individual vehicle audience information by preserving as many

estimation parameters as vehicles in the schedule. Though the complexity means that

multivariate models require more data manipulation and computation time, they are

considered to provide more accurate estimation than do univariate models. However,

since true multivariate probability distribution methods that estimate broadcast audiences

have not been developed, most multivariate models are magazine audience estimation

models that are tested in schedules with no more than two insertions in each vehicle

(Leckenby & Kim, 1992).

More than fifteen univariate probability distribution models have been developed

since Agostini's simple reach estimation model. Some include the beta binomial

distribution (BBD), Markov binomial, Morgenzstern, and Hofians geometric distribution

(HGD) (Leckenby & Ju, 1990). Among them, one of the most widely used and studied

exposure distribution models is the beta binomial distribution (BBD) model because of the








simple, parsimonious, and relatively sound estimation (Ju & Leckenby, 1990; Kreshel,

Lancaster, & Toomey, 1985).

The BBD model, which was first introduced by Metheringham in 1964, is a

statistical procedure for estimating the likelihood of a target audience exposure to a

schedule a given number of times. The binomial distribution estimates the number of

different ways a target audience can be exposed to a schedule from one to N total

insertions. The beta distribution estimates the average probability of exposure to the

schedule from one to N times (Lancaster & Katz, 1988).

The BBD model has been used as a basis of estimating the exposure distribution of

not only univariate but also multivariate models. The notable univariate variations of the

beta binomial distribution are BBD-DE, Hofmans BBD, and BBD simulation. The

multivariate analog of the BBD is the Dirichlet distribution.

Although the BBD has been accepted to be fairly accurate across a wide variety of

advertising media circumstances, the use of the BBD to evaluate an entire media category

schedule has some limitations. One of the major problems is that the use of BBD treats all

vehicles and pairs of vehicles in a schedule as a single average vehicle. Therefore, using

the BBD alone can mislead the estimation of reach and exposure distributions, particularly

in the case of disparate insertions and/or ratings among the vehicles in a schedule

(Lancaster 1992). It is also known that the BBD model often yields an over-estimation in

reach (Leckenby & Rice, 1985; Leckenby & Kim, 1992). However, the BBD model has

been used as a mathematical basis to predict advertising message exposure distributions of

advertising schedules because of the wide acceptance, proven reliability, and possible

variations over more than two decades (Leckenby & Boyd, 1984; Ju et al., 1990). For








example, due to the lack of pair-wise duplication data such as vehicle self-pair and cross-

pair ratings, many studies use a variation of the BBD model which is the beta binomial

distribution with limited information (BBD-L). The BBD-L model, developed by

Lancaster and Martin (1988), does not require pair-wise duplication data and was shown

to be accurate in predicting pair-wise duplication among vehicles.


Model Selection

Based on the BBD model, a computerized media evaluation program, ADplus, is

used in this study because the program is comprehensive, widely available, and also can

handle vehicle and message reach and frequency.

ADplus, originally developed by Lancaster in 1990, is an easy-to-use program that

helps predict advertising campaign effects and evaluate advertising media plans. The

program is currently distributed worldwide by Telmar Information Service Corp. of New

York (http://www.telmar.com) and is used by thousands of media marketing professionals

worldwide. It is possible to evaluate all major advertising media categories and

subcategories with ADplus. It also helps advertising media planners find optimum

schedules involving multiple media categories.

As input data, ADplus needs advertising media vehicle ratings, target market size,

advertising cost, and number of insertions. As shown in Figure 4, ADplus results provide

media planners with frequency distributions, reach, effective reach, gross rating points

(GRPs), average frequency, gross impressions (GIs), cost-per-thousand (CPM), and cost-

per-rating point (CPP). ADplus also provides two types of frequency distributions,

including vehicle and message frequency distributions. For the magazine example in








Figure 4, the message effective reach (3+) is 0.1 percent. Therefore, if it is assumed that

consumers are aware of the new advertising message after three exposures to the

magazine advertisement (effective frequency), only 0.1 percent of the target audience is

estimated to have been exposed at this level.

Although the typical vehicle and message effective reach of 3+ is shown in the

ADplus results, media planners also can examine other definitions of effective frequency

and reach that may be more fitting with their situation including, for instance, effective

frequency and reach of 4+ or 5 through 10.

ADplus can overcome the problems of the BBD model which treats all vehicles

and pairs of vehicles in a schedule as a single average vehicle. With the estimation of

vehicle self-pair reach, ADplus can estimate separate vehicle exposure distributions. A

matrix of individual vehicle exposure distributions is then used to estimate multiple vehicle

exposure distributions in conjunction with vehicle cross-pair reach data. This procedure is

referred to as a beta binomial matrix exposure distribution (BBMD) model (Lancaster,

1993). The technical procedures used to implement the BBMD model are described in

detail by Lancaster (1993) and important dimensions of this model are presented in

Appendix D.

The ADplus program also helps media planners find an optimum combination of

media vehicles and media categories within desired budgets and recommends media

budgets for desired communication goals.











ADplus(TM) RESULTS: MAGAZINES

By: Hyunsoo Park
University of Florida --
For: ASIANA Airline

Typical monthly schedule
f
Target: 2,884,983
Korean men 20-49 years 0
old who live in Seoul 1


Message/vehicle


45.0%


Frequency (f) Distributions


Vehicle

% f % f+

65.1 100.0
28.4 34.9
5.8 6.5
0.7 0.8
0.1 0.1
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0


Message

% f % f+

82.6
15.9 17.4
1.4
0.1
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0


Summary Evaluation

Reach (1+) 34.9% 17 %
Effective reach (3+) 0.8% 0.1%)
Gross rating points (GRPs) 42.2 190
Average frequency (f) 1.2 1.1
Gross impressions (000s) 1,266.0 569.7
Cost-per-thousand (CPM) 21.23 47.17
Cost-per-rating point (CPP) 637 1,415

Vehicle List Rating Ad Cost CPM-MSG Ads Total Cost Mix


Sisa Journal
Han Kyurei 21
Jugan Chosun
News Plus
Jugan Hankuk
News Week
News people
News Maker
Jugan MaeKyun
Economist
HanKyung Busi
Shin Dong-Ah
WalKan Chosun
Win


2,250
2,000
2,250
1,875
1,500
2,250
1,500
1,500
1,750
1,750
1,625
1,875
1,875
2,875


25.25
33.67
30.86
49.60
69.44
59.52
55.56
65.36
81.02
129.63
240.74
26.21
26.21
177.47


2,250
2,000
2,250
1,875
1,500
2,250
1,500
1,500
1,750
1,750
1,625
1,875
1,875
2,875


8.4%
7.4
8.4
7.0
5.6
8.4
5.6
5.6
6.5
6.5
6.0
7.0
7.0
10.7


Totals: 47.17


26,875 100.0%


Figure 4. ADplus Results








SMdel Modification

As shown in Appendix C, the media evaluation model used in this study needs

regression equations for the estimation of R,, and R,,. R2, is reach as a percentage of the

target audience of one typical insertion in vehicle i (self-pair or two-use reach), while

R,, is cross-pair ratings in vehicle i and. The estimated regression equations of R2, and

R, for the South Korean target audience may be different from US equations. In 1992,

Korean Gallup reported the regression equations for the estimation of R2, (n = 528) and

R, (n = 560,292) with South Korean target audience groups (see table 18).

Table 18. R, and R, estimation equations for network television

US Equations (for overall audience)

R, Estimation R, = -0.0065 + 1.795 R, 0.331 (R,)2

R,, Estimation R, = -0.0015 + 1.0095 (R,+ R,)- 1.5R, R,

Korean Equations (for 20 49 year old men)

R, Estimation R2,= 0.000525 + 1.848325R,- 1.432(R,)2

R, Estimation R, = 0.00005 + 1.003 (R,+ R,)- 1.87755R, R,
American source: adapted from Kent Lancaster (1993).
South Korean source: adaptedfrom Korean Gallup (1992).



Since the Korean Gallup reported the estimated regression equations with gender

and age variables, in this study, the equations for only the target audience group of

ASIANA campaign are estimated. Then, the media evaluation model will be modified

with the regression equations calculated from South Korean target audience which is 20 -








49 year old men. The estimated American and South Korean regression equations are

shown in Table 18.



Error Estimation for the Media Evaluation Model


The BBD model has been widely accepted in both academia and industry because

of its simplicity, aggregating power, and good empirical fits. This approach has been

improved by the availability of household-level exposure records obtained from people-

meters and the increasing computation power of computers. However, this approach is

also assumed to exhibit error in its estimations. It is necessary to estimate how much error

the model is likely to produce.

Tests of the performance of media evaluation models have been undertaken in the

US. These studies were conducted to compare the performance of the various media

evaluation models. Recently, studies tried to compare the accuracy of media evaluation

models with different error definitions. Frequently used error definitions are average error

in reach (AER), average error in levels of exposure (ALE), and average percentage error

in exposure distribution (APE). The error terms are defined as the difference between

observed (tabulated) and estimated values in a schedule. For example, AER is the sum of

the absolute differences between observed and estimated reach in each schedule divided by

the total number of schedules. Sometimes, each error term produced different results that

made it difficult to decide which model's performance is superior to another. However,

AER, APE, and ALE have been frequently used in these studies.

Leckenby et al. (1990) compared four different models' performance in terms of

error estimation with about 500 magazine schedules tabulated from SMRB (Simmons








Market Research Bureau). This study showed that the BBD produces the smallest

percentage of error for reach estimation among the models used in the study. The study

also revealed that the larger the schedule size is, the more accurate the estimations made

with the models.

The beta binomial distribution (BBD) is an univariate probability distribution

model which treats all vehicles as one univariate and 'composite' vehicles and then

averages all vehicle insertions of a schedule. Therefore, the model often produces over-

estimations in reach. Because of the problem with univariate models, the development of

multivariate models has been attempted. However, because of the excessive

computational difficulties involved in using all of the vehicle numbers together, complete

multivariate models have not been developed. Ideally, multivariate media evaluation

models should be developed because they may allow the most realistic view of complex

media evaluation processes without simplifying or averaging assumptions which are

currently used with univariate models (Leckenby et al., 1993).

An alternative is "sequential aggregation." Although sequential aggregation is not

a true multivariate method, studies showed that the models using sequential aggregation

are better for estimating reach (n+) than those of univariate models. With sequential

aggregation, two-vehicle bivariate distributions are collapsed to form a new composite

vehicle which is then combined with a third vehicle and this procedure is continued until

all vehicles are aggregated into the distribution of the schedule. The BBMD model used

in this study uses sequential aggregation.

Leckenby et al. (1993) tried to test the performance of several media evaluation

models including model using sequential aggregation. In this study, about 480 magazine








schedules were tabulated from SMRB. The average error in reach (AER) with BBD

model was about 6.4 percent, while the AER with the model using sequential aggregation

based on the BBD model [Morgenszten Sequential Aggregation Distribution model

(MSAD)] was 3.1 percent. The error estimation of the model used in this study (BBMD)

is similar to MSAD because both models are based on the BBD and use sequential

aggregation.

According to Leckenby et al. (1993), the model which uses sequential aggregation

is also the most accurate in predicting each level of the exposure distribution. The average

total error percentage in the exposure distribution (APE) was about 15 percent with

MSAD, while that of the BBD was 33 percent. Therefore, it can be assumed that about 3

percent error is likely to be made for reach estimation and about 15 percent error is

expected for the total exposure distribution for models which use sequential aggregation

based on the BBD. According to a selected example of observed and estimated exposure

distributions with their large schedule sample in the study, the error estimation for reach

(3+) was about 1.9 percent which is smaller than the error estimation of reach (1+) of 3

percent.

However, these error estimations also have limits. First, all of the studies

mentioned above used vehicle data instead of advertising message data. Since the model

used in this study will predict reach/frequency estimation with advertising message data,

the error estimation with message data would be smaller than that of vehicle data.

Another important point to be considered is that all the studies mentioned above used only

magazine advertisements. No study has been conducted for the error estimation of media

evaluation models for TV advertising. Since the ASIANA campaign mostly used








television advertising (about 93.5 percent of total campaign expenses), the error

estimation for television advertising is suggested for a future study. Additionally, this type

of study should be conducted with the actual BBMD model for better error estimation and

it is also recommended that the error estimation studies should be conducted with South

Korean data because the error estimation with US data may be different from South

Korean error.

It also should be considered that all of the observed (tabulated) data of previous

studies are based on the sample estimates. Although the possible error can be very small

with proper sampling methods, all samples must have sampling errors. The studies

mentioned above seem to have ignore the possible sampling errors for observed

(tabulated) data to examine error estimation of media evaluation models. The application

of sampling errors should be considered for better error estimation of media evaluation

models.

The model used in this study will be examined to show how accurately the model

can predict the reach/frequency estimation. If a critical amount of error is involved with

the model estimation of reach (n+) in this study, the study results will not demonstrate

significant (accurate) predictions. Therefore, the error estimation may not be included to

test hypotheses in this study. Whatever amount of error occurs from the model used in

this study, the error should be a part of the model performance.

In addition to the error estimation of the model used in this study, additional errors

are likely to be produced because of the unknown factors in South Korea, such as carry-

over rate and magazine ratings. However, the normative framework combined with a




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EATWDW5SS_ZOA7SJ INGEST_TIME 2013-10-10T03:08:24Z PACKAGE AA00017634_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


99
/V
Where:
Z = Test statistic,
A
ft = Percent of the sample indicating preference or willingness to use,
ft$ = Minimum predicted percent of consumer preference or
willingness to use,
CJ
ft
Standard error of sample consumer preference or
willingness to use, where a. = )x0(\ - tt0) / n .
Source: adapted from Agresti and Finlay (1986).
If the calculated Z-scores are big enough to accept the research hypotheses ( Ha :
n > n 0 ) at 90 - 95 percent confidence levels, the second and third research hypotheses
are statistically supported. In other words, the framework will support the second and
third hypotheses if the estimated reach 4+ and 5+ respectively are higher or equal to the
actual consumer preference and willingness to use.
Calibrating Forecasting Procedures
Since this study lacks some necessary information, such as the message/vehicle
ratio for magazine advertising and advertising carry-over rate for an airline service in
South Korea, the normative framework may show a gap between what is predicted and
what is found in the tracking study. In this case, the effective frequency levels (n+) used,

FORECASTING ADVERTISING COMMUNICATION EFFECTS
USING MEDIA EXPOSURE DISTRIBUTION MODELS:
TEST MARKET RESULTS IN SOUTH KOREA
By
HYUNSOO PARK
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1998

Copyright 1998
by
HYUNSOO PARK

ACKNOWLEDGEMENTS
First of all, heartfelt thanks goes to my adviser, Dr. Kent M. Lancaster. I really
appreciate his kind guidance in developing the research design and the normative
framework, and in writing this dissertation. Without his scholarly advice, guidance, and
great help during my whole Ph.D. study, this dissertation would not have been possible. I
also would like to thank my committee professors, Dr. Joseph Pisani, Dr. Leonard Tipton,
Dr. Marilyn Roberts, and Dr. Richard Scheaffer for excellent suggestions, encouragement,
and help.
Acknowledgments are due to several professionals and professors in Korea.
Special gratitude goes to the CEO of SangAm agency, Dr. Tong-Yeol Sung, and Ms.
Nan-Hee Choi, Vice-chief of the marketing department, for their considerable efforts in
providing valuable data. Mr. Hyun-Tae Kim, Dr. Jung-Sik Jo, and Dr. Yang-Ho Choi also
provided valuable help.
Appreciation is also extended to June-Hyung Ji and my brother-in-law, Yeon-Soo
Kang, for sending me precious Korean papers and data for this study. Special thanks go
to my brothers and friends, Yoon-Soo Park, Jeong-Soo Park, Kwang-Ho Lee, Dong-
Hyun Han, Yoon-Sung Hwang, Il-Joon Yoo, Seung-Kwan Ryu, and Ho-Cheol Kim
whose encouragement and friendship made this dissertation possible.
I am really grateful to my lovely wife, Yeon-Mi, whose self-sacrificing support and
help during my Ph D study have given me the inspiration, confidence, and balance
necessary to complete this endeavor.
Finally, I would like to dedicate this dissertation and my Ph D. degree to my
parents, Jong-gug Park and Bum-Rok Lee. “Thank you for life, love, and help every
moment of my life. I love you.”

44 o| m 4 414 2-1 *4 *444 *14 4 4 4
*2* 444 $24 *24*11 2444. 4 *14 2*4 44* a
4* 35. 1 * 4424 4 *14 ?>! * 4* 2* 44! 1*4
2^44.
* *14 441 444 41*4 §44 4 4 4 45.1- 414
Dr. Kent M. Lancaster 2*4*112 4* 441- 2¡¡J44. 44 444
21 *444 *4 2*4 14 4442. 44 45!*4 4. o>i¡h *
*1! 44«fl *4 2 4* 424- 441 444 S-5-4 Dr. Joseph
Pisani, Dr. Leonard Tipton, Dr. Marilyn Roberts, Dr. Richard
Scheaffer 2*4-¡H|2 444 444 2=J44.
4- *44 El> a sin 4* *14 2*22 444 *
44*44. 44 44 444 44*4 444 4* 51*1 *4
(1)1444 43.44 4 *4 444, 4S. *41 44 44*4 4
4*114, 4 44 444 :14s. *4 44^4 2 44 144, 24
4424 4 4444 4* 44 2444. *1 4*11 t°}
=11* 44444 4 *4 *4 44 4 4*4512 244* 41
444, 2 44 University of Florida 2* 51*414 4414 I’ll
411 4*2 4*44.
5444 1444 14 44 42 44444 441 24* 1*
44 1*8 4* 2431 4 42, 4 14, 4 14, 1 4*, 1 *4, 4
il 14 4*14 *42 £1 2*4 44*44.
4144 44441 444 4 424 4*8! 4444 if!
444* 44 1444, 414 44! 14 14 4!, 42*42 4!
441 2444. *22, 4*4 *4! 44 44 *44 144 444
44414 44! 244, 44 44 *14 44! 4*il 4414 21
*! IV

TABLE OF CONTENTS
page
ACKNOWLEDGEMENTS iii
LIST OF TABLES ix
LIST OF FIGURES xi
ABSTRACT xii
CHAPTER I. INTRODUCTION 1
Advertising Campaign Effects 1
Complexity of the Problem 2
Media Exposure Distribution Models 3
Objectives of this Study 5
Rationale 6
Organization of the Dissertation 8
CHAPTER II. MEASURING AND FORECASTING ADVERTISING EFFECTS 9
Primary Reason for Evaluating and Forecasting Advertising Campaigns 10
Problems in Advertising Campaign Measurement 11
Advertising Effects on Sales 13
Difficulties in Measuring Advertising Effects on Sales 14
Advertising Communication Effects 16
Awareness 17
Previous Research on Awareness 19
Recall 20
Liking and Attitude Toward Advertisement 21
Preference 23
Purchase Intention and Actual Purchase 24
Achieving Desired Communication Effects 25
Effective Exposure to Advertising 28
Advertising Exposure 28
Reach 29
Frequency and Frequency Distribution 30
Effective Frequency and Reach 32
Problems with Using the Concept of Effective Frequency 34
How to Set Effective Frequency Levels 37
v

CHAPTER m. BACKGROUND ON THE SOUTH KOREAN ADVERTISING
INDUSTRY 42
Introduction 42
South Korean Economic Development 42
The Growth of the Advertising Industry in South Korea 43
Advertising Media in South Korea 45
Recent Changes in the South Korean Advertising Industry 47
Studies on Measuring and Forecasting Advertising Effects in South Korea 48
Media Planning and Evaluation Procedures of South Korean Advertising Practitioners ..51
Purpose of the Studv 52
Methods 52
Results 53
Definition of Reach and Effective Reach 56
Communication Effects Used 57
Message Weighting and Reasons of Not Weighting 59
Needed Improvements 60
Discussion and Implications for Further Research 61
CHAPTER IV METHODOLOGY 64
Case Study 66
AS IAN A Airline 66
Target Audience and Advertising Campaign 69
Competitor 70
Media Data 72
TV Ratings 72
Magazine Ratings 73
Tracking Data 74
Sampling 74
Measurement 75
Media Evaluation Model Used in This Study 76
Beta Binomial Exposure Distribution Model 76
Model Selection 79
Model Modification 82
Error Estimation for the Media Evaluation Model 83
Suggestions for Improving Campaign Efficiency and Budget Levels 87
Sophisticated versus Naive Comparison 88
Summary of the Normative Framework 89
Research I lypotheses 90
Testing Procedures 93
Statistical Procedures 97
Calibrating Forecasting Procedures 99
vi

CHAPTER V. RESULTS
104
Introduction 104
Test Market Results 106
Media Evaluation Model Prediction 106
Tracking Study Results 109
Testing Hypotheses 111
Awareness of the Campaign Message Ill
Preference Ill
Willingness to Use 112
Calibrating Procedures 113
Critical Analysis of the ASIANA Advertising Campaign 120
Suggestions for Future ASIANA Campaigns 125
Selecting Appropriate Vehicles 125
Appropriate Budget for Monthly Schedule 127
Sophisticated versus Naive Comparison 132
Pooled Approach 132
Vehicle versus Message 133
CHAPTER VI. SUMMARY, CONCLUSIONS, AND IMPLICATIONS 135
Summary and Conclusions 135
Implications 139
Limitations and Suggestions for Future Study 140
APPENDICES
A. SURVEY QUESTIONNAIRE IN BOTH ENGLISH AND KOREAN 143
ENGLISH VERSION 144
KOREAN VERSION 151
B. TWO TYPES OF MAGAZINE ADVERTISEMENTS 158
C. TV PROGRAM RATINGS, ADVERTISING MESSAGE RATINGS,
MAGAZINE RATINGS, AND COSTS FOR THE ASIANA
ADVERTISING CAMPAIGN IN BOTH ENGLISH AND
KOREAN 161
ENGLISH VERSION 162
KOREAN VERSION 166
D GENERAL FORM OF THE BBMD MODEL 170
E ADPLUS RESULTS FOR NETWORK TELEVISION IN FEBRUARY.... 176
F. ADPLUS RESULTS FOR NETWORK TELEVISION IN MARCH 179
vii

G, ADPLUS RESULTS FOR NETWORK TELEVISION IN APRIL 182
H. ADPLUS RESULTS FOR MAGAZINE ADVERTISEMENTS 185
I ADPLUS RESULTS FOR NETWORK TELEVISION AND
MAGAZINES COMBINED IN FEBRUARY 187
J. ADPLUS RESULTS FOR NETWORK TELEVISION AND
MAGAZINES COMBINED IN MARCH 191
K. ADPLUS RESULTS FOR NETWORK TELEVISION AND
MAGAZINES COMBINED IN APRIL 195
L. ADPLUS RESULTS WITH FLOWCHART 198
REFERENCES 203
BIOGRAPHICAL SKETCH 211

LIST OF TABLES
Table page
1. Common measures for advertising communication effects 17
2. Relationship between common communication measures and needed
exposure frequency in a normal marketing, copy, and media situation 27
3. Example of frequency distribution 32
4 Marketing, copy and media factors that affect effective frequency 38
5 Suggested additional points and modified effective frequency levels
for brand X 41
6. The progress of the South Korean advertising industry 45
7. South Korean advertising media, total expenses, and costs 46
8. Sung’s independent and dependent variables and regression equations 50
9. Systems used to develop advertising media plans 54
10. Factors used in evaluating media plans 55
11. Representative definitions of effective reach in comparison with two
previous American studies 58
12. Variables used to evaluate advertising communication effects 59
13. Normative framework necessary to predict advertising communication effects 65
14 Domestic and international market shares 68
15. Advertising insertions by media category and expenses of
“Tier name is ASIANA” campaign from February 17toMay 16, 1996 70
16. Competitor GRPs and advertising expenses 71
17. Sample characteristics 75
18. R2i and R.. estimation equations for network television 82
19. Summary of the normative framework applied to the ASIANA case study 89
ix

20. Suggested points and modified effective frequency levels
for ASIANA Airline campaign effects 91
21. Adjustment options, expected effects, and decisions in this study 102
22. Consumer tracking study results and model predictions 110
23. Adjustment options and changes in predictions 116
24. Consumer tracking study results and model predictions with calibration 118
25. CPM-MSG figures for the most efficient 11 television programs (vehicles) 122
26. CPM-MSG figures for all magazine advertisements 124
27. Relatively high CPM-MSG vehicle list in the ASIANA campaign 126
28. Relatively low CPM-MSG vehicles added for schedule modification 126
29. Suggested monthly budget levels and estimated message reach 3+ 131
x

LIST OF FIGURES
Figure page
1. The effects of top-of-mind awareness of the advertisement 20
2. The shape of a typical message reach curve 30
3. Sung’s model to estimate advertising effects 50
4. ADplus results 81
5. Media evaluation model prediction 108
6. Media evaluation model prediction with calibration 117
7. Estimated frequency distributions for the three-month ASIANA campaign 119
8. Budget allocation among television programs 120
9. Budget allocation among magazines 121
10. Reach percentage changes with budget increases for February vehicles,
television and magazines combined 128
11. Reach percentage changes with budget increases for March vehicles,
television and magazines combined 128
12. Reach percentage changes with budget increases for April vehicles,
television and magazines combined 129
13. Estimated reach (3+) for each monthly advertising schedule 130
14. Estimated message reach (3+), naive versus sophisticated approaches 133
15. Estimated frequency distributions with vehicle and message data 134
16. Ideal procedures for predicting and evaluating advertising effects using media
exposure distribution models 136
xi

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy
FORECASTING ADVERTISING COMMUNICATION EFFECTS
USING MEDIA EXPOSURE DISTRIBUTION MODELS:
TEST MARKET RESULTS IN SOUTH KOREA
By
Hyunsoo Park
May, 1998
Chairman: Dr. Kent M. Lancaster
Major Department: Mass Communication
The primary objective of this study is to demonstrate how to predict advertising
communication effects using a variation of the beta binomial exposure distribution (BBD)
model. A computer based model which embodies important aspects of normative theories
of media planning is used to forecast the communication effects of a South Korean
advertising campaign.
Due to the lack of systematic study of advertising effects in South Korea, the
South Korean advertising industry needs objective and scientific methods to develop a
model that is applicable to the South Korean advertising environment. Therefore, this
study applied a media exposure distribution model that is used in the United States and
worldwide to the South Korean market to probe how accurately such models and
procedures can predict the communication effects of a South Korean advertising

campaign. A consumer tracking study is used to determine the accuracy of forecasting
advertising campaign effects.
Prior to the application of the media evaluation model and procedures, survey
research was conducted among South Korean media practitioners. The goal was to show
how the media practitioners view the current state of reach/ffequency estimation methods
and which communication effects are used to predict or measure potential advertising
impact. Based on the survey results, previous studies as well as marketing, copy, and
media factors, an effective exposure frequency level is estimated to predict target audience
awareness of a South Korean advertising campaign message.
In addition to advertising message awareness, this study also examines target
audience preference and willingness to purchase for the brand used as a case study.
Minimum frequencies of consumer exposure to the advertising campaign necessary to
achieve preference and willingness to purchase also are estimated with the same normative
framework. Then, the minimum effective frequency levels are analyzed in comparison
with actual consumer tracking data to determine whether more accurate predictions would
have been possible.
The study results showed that the normative framework and media exposure
distribution model used in this study can be accurately used in predicting target audience
awareness of the campaign message. The prediction for target audience awareness of the
campaign message was close to the tracking study results. With additional calibrating
options, the predicted awareness of the campaign message could be exactly matched with
the tracking study results.
xiii

This study also suggests that there may be a hierarchy in consumer processing of
advertising messages. The needed target audience exposure to achieve preference was
significantly higher than the needed exposure for awareness. The needed message
exposure for willingness to use was also significantly higher than the target audience
exposure for awareness and preference. Though many more factors should be considered
for accurately predicting or evaluating advertising effects in terms of preference and
willingness to use, this study still provided useful results for the estimation of advertising
effects on preference and willingness to use.
The advertising campaign used as a case study also was critically analyzed to
suggest appropriate monthly budgets and vehicle selections for future campaigns. The
sophisticated approach used in this study was compared with typical naive approaches to
show how the potential gaps can be misleading.
In this study, the normative frameworks, media exposure distribution models,
evaluation methods, including many influential factors, and calibrating procedures provide
useful guidance for forecasting advertising communication effects.
xiv

CHAPTER I
INTRODUCTION
Advertising Campaign Effects
Advertising, an important promotional tools of modern marketing management,
has been used worldwide to present information to potential buyers and reinforce
consumer preference for a company’s products or services (Liben, Kotler, & Moorthy,
1992). Every year billions of dollars are invested by companies around the world to test,
produce, and implement advertising campaigns. According to Aaker, Batra, and Myers
(1992), estimated advertising expenditures will be $780 billion worldwide and $320 billion
in the United States alone by the year 2000. Because of these enormous expenditures,
advertisers want the advertising planners and managers to identify exactly what are the
results obtained from the advertising investment and to provide evidence of the return on
investment. Therefore, industry practitioners and social scientists throughout the world
continue to predict and evaluate the effects of advertising campaigns not only prior to
campaign execution but also after advertising is aired or printed.
1

2
Complexity of the Problem
The effects of advertising campaigns are evaluated according to campaign
objectives. No matter what objectives an advertising campaign has, such as an increase in
sales, market share, brand awareness, advertising message awareness, recall, preference,
or something else, there seem to be limited means of accurately measuring advertising
campaign effects. The main reason is that advertising effects typically interact with other
elements in the marketing mix, such as an advertiser’s image, price, competitor
promotional activity, distribution, product quality, and other promotion activities.
Although technology is making the prediction and evaluation of advertising effects easier,
skepticism about the possibility of accurately predicting or evaluating advertising effects
has continued over the years (Lancaster & Helander, 1987).
Another problem in predicting or evaluating advertising effects is what is to be
measured? Two frequently used approaches are advertising communication effects and
sales. Since one of the main purposes of advertising is to communicate persuasive
messages, measuring communication effects, such as brand and advertising message
awareness, recall, attitude toward brand, preference, and willingness to purchase, is a
natural way to evaluate advertising campaign effects (Schultz, Martin, & Brown, 1984;
McDonald, 1995; Woodside, 1996).
Measuring actual behavior of consumers, such as sales or purchases, is another
frequently used approach to measure advertising campaign effects. Since the ultimate goal
of advertising may be the increase of sales, the amount of sales has been used to measure
advertising effects (Leone, 1983; Assumus, Farley, & Lehmann, 1984; Holstius, 1990;

3
Schreiber, & Appel, 1991; Lodish et al., 1995). However, since sales are the result of
various marketing efforts, it is very difficult to measure the absolute effects of an
advertising campaign in terms of sales.
Advertising communication effects can be used more efficiently as a method of
evaluating the effects of advertising campaigns. Brand awareness, advertising message
awareness, recall, preference, and willingness to purchase are the most frequently used
measures in terms of communication effects. For example, advertising campaign message
awareness can be measured with much less interaction with other marketing factors than
the measure of sales. However, the communication measures should be explained in
relation to the final objectives, such as increase in sales or market share. Many previous
studies in the United States suggest that there is a positive relationship between brand or
advertising message awareness and preference, sales, and market share (Burke &
Schoeffleer, 1980; Wilson, 1981; Zufryden, 1996; Woodside, 1996).
What is best among various communication effects is up to the planners to decide
in relation to campaign objectives. However, no matter what kind of approach is used to
measure advertising campaign effects, it is still difficult to identify actual advertising
effects while completely controlling for all other important controllable or uncontrollable
dimensions of the marketing mix, such as pricing behavior, distribution intensity, product
quality, competitor target market selection, and other promotion activities.
Media Exposure Distribution Models
Because of the enormous advertising expenditures and the complexity of the
problem, studies of advertising effects have been active in the advanced countries like the

4
United States in an effort to overcome the problems explained above (Leckenby & Kishi,
1982a; Lancaster & Helander, 1987; Rust & Klompmaker, 1981; Rice, 1985; DeKimpe &
Hanssens, 1995a; Hansen, 1995). Models have been developed to predict and measure
advertising effects.
One of the most frequently used models in the United States is the beta binomial
exposure distribution (BBD) model (Leckenby & Kishi, 1982), which was developed by
Metheringham in 1964. The general goal of the BBD model is to estimate the probability
and distribution of consumer exposures to advertising media schedules. The model
incorporates both beta and binomial distributions of schedule exposures. The beta
distribution is used to estimate the probability of consumer exposures to any number of
advertising vehicles or messages included in a schedule, such as one, two, three, or up to
any number of exposures to advertising vehicles or messages. These probabilities are then
used with the binomial distribution which estimates the proportion of consumer exposures
to advertising vehicles or messages, for example, once, twice, or three times (Lancaster &
Katz, 1989).
Computer-based models also have been developed based on the beta binomial
exposure distribution (BBD). One such microcomputer program is ADplus, which is used
to evaluate advertising schedules based on a variation of the beta binomial exposure
distribution model. It uses procedures that vary by media category to account for
differences in available audience and cost data (Lancaster, 1993).
Three test market studies in the United States (Lancaster & Helander, 1987;
Lancaster & Katz, 1989) demonstrate the ability of the predecessor of ADplus to predict

5
the effects of advertising campaigns, including communication effects such as advertising
exposure and brand awareness.
Objectives of this Study
While the study of advertising effects has been active in advanced countries, it has
not been widespread in some other countries such as South Korea. Even though the
South Korean advertising industry has in recent years grown into the world’s 13th largest
market (US $6.1 billion in 1995), its research remains at an immature level. Only a few
studies have been conducted about measuring advertising effectiveness in South Korea
(Sung, 1989; Korean Gallup, 1992).
Because of the lack of systematic study of advertising effects in South Korea, the
development of an advertising effects model applicable to the advertising environment of
South Korea is relatively poor. Therefore, a model of measurable advertising effects that
meets the advertising environment of South Korea is needed. Such a model should be
developed by means of an objective and scientific method for forecasting and evaluating
advertising effects. Much more research should be conducted in South Korea concerning
the evaluation of desired communication effects.
Therefore, one of the primary objectives of this study is to demonstrate how to
predict the effects of an advertising campaign using a variation of the beta binomial media
exposure distribution model in the South Korean market. The computer-based beta
binomial exposure distribution model ADplus will be used to forecast the effects of an
advertising campaign. These predicted advertising effects will be compared with actual
consumer survey results in the South Korean market. Consequently, the computer-based

6
BBD model developed in the United States will be used to examine how accurately the
model can predict the effects of a South Korean advertising campaign.
Little information is available about how South Korean media practitioners view
current media evaluation terms and which communication effects are frequently used to
predict or measure potential advertising impact. Therefore, a survey was conducted prior
to the application of the media evaluation model and procedures. Based on the survey
results and previous American and Korean studies, optimal advertising exposure
frequencies will be estimated to predict desired communication effects, such as consumer
awareness of an advertising campaign message, preference, and willingness to purchase.
Then, a consumer tracking study will be used to determine how accurately the model
predicts the potential advertising impact.
This study also will examine methods used to measure advertising effects as well
as factors that may influence advertising communication effects, such as advertising
message quality, market share, and competitor promotional activities. For the brand used
in this study, these factors will be used to determine the level of effective frequency, which
is a predetermined level of consumer exposure repetition to an advertising campaign that
is needed to achieve desired communication effects. For example, the brand that has
strong competitor advertising, low market share, low brand loyalty, or low advertising
message quality may need a relatively high effective frequency level.
Rationale
Advertising planners have experienced difficulties in predicting and measuring the
effects of advertising campaigns over the years. One of the major challenges in evaluating

7
advertising effects is the inability of researchers to estimate the accurate functions between
advertising and consumer responses. Predicting advertising effects is certainly challenging
due to the difficulty in controlling other important dimensions of the marketing mix
(Lancaster & Helander, 1987). Therefore, this study will attempt to overcome some of
these problems. Advertising expenditures will be analyzed by their essential message and
media characteristics, instead of examining the amount in relation to only sales or
communication variables. Actual advertising media schedules and messages gathered from
the South Korean market will be analyzed from the perspective of media planners in terms
of the likely effects on the intended target consumers.
Since the reliability of the beta binomial exposure distribution model has been
proven over decades in the United States (Leckenby & Boyd, 1984; Leckenby & Kishi,
1982; Kreshel, Lancaster, & Toomey, 1985; Rust 1986; Lancaster & Helander, 1987), it
will be worthwhile to apply the BBD model to a totally different market, such as South
Korea, to examine whether the model works as it does in the American market.
The total expenses for advertising in South Korea increased more than 70-fold
during the last two decades. However, in spite of the recent rapid growth of the South
Korean advertising industry, there have been few Korean studies on measuring and
predicting advertising effects. Studies should be conducted to develop a model that is
suitable for the Korean advertising environment. This study, which applies a model used
in the United States and around the world, such as Europe and South Africa, to the South
Korean market, will help South Korean advertising planners develop scientific models.
These can then be used to evaluate and forecast advertising results in terms of
communication effects.

8
Organization of the Dissertation
This dissertation is organized into six chapters. The introductory chapter provides
the conceptual framework of this study, including the objectives and rationale. The
second chapter presents a literature review related to the purpose of this study. The
review includes many problems of measuring advertising effects as well as previous studies
on predicting and measuring advertising communication effects. Chapter II also presents
the concept of effective exposure to advertising campaigns from the prospective of media
planners.
Chapter III presents South Korean economic development, the growth of the
South Korean advertising industry, the advertising media, recent changes in the advertising
industry, and previous studies measuring and forecasting advertising effects. Survey
results of 61 South Korean advertising agencies are also included in Chapter III. This
survey research demonstrates how South Korean media practitioners perceive media
evaluation terms and the current state of reach and frequency methods.
Chapter IV, Methodology, explains the brand used as a case study, the framework
used to analyze this case study, data collection, research hypotheses, testing procedures,
and calibrating options. Chapter V concerns the results of hypothesis testing and the
application of a commercial model to the South Korean market. Test market results are
provided with comparison to the model predictions of advertising communication effects.
Finally, a summary, conclusions, implications, limitations and suggestions for
future research are presented in Chapter VI.

CHAPTER II
MEASURING AND FORECASTING ADVERTISING EFFECTS
Over the past decades, the research of advertising effects has evolved with great
detail and diversity of theoretical, methodological, and technical approaches. These
research efforts suggest that a possible solution to the ultimate problems of measuring or
predicting advertising campaign effects may be within reach (Dutka, 1995). In this
chapter, the reasons and problems with forecasting and measuring advertising effects, and
publications conducted to help measure or forecast advertising effects are reviewed,
focusing on recent developments. The review of literature will help develop appropriate
methods to conduct this study.
Advertising target audiences should be exposed to the advertising campaign with
optimal frequency to achieve the desired advertising effects. Since there is little agreement
on precisely how many exposures to an advertising campaign message are sufficient to
achieve the advertising campaign goal, studies have been conducted to examine this issue
(Krugmen, 1972; Craig, & Ghosh, 1993; Dekimpe, & Hanssens, 1995;).
In this chapter, in addition to the previous studies on predicting or measuring
advertising effects, the factors that may influence the number of sufficient exposures are
9

10
also reviewed. This study will attempt to integrate the possible influential factors to
predict and measure advertising effects from the perspective of media planners.
Primary Reasons for Evaluating and Forecasting Advertising Campaigns
The decision process for advertising typically begins with some form of planning
and either an implicit or explicit strategy or objective. After setting or establishing
measurable objectives and a budget for the various stages and elements in the advertising
campaign, creative strategies and media plans are prepared. Consequently, the planners
need to know whether the advertising campaigns achieved the goals established. Because
of enormous expenditures on advertising campaigns, advertisers want advertising planners
and managers to identify exactly what are the results to be obtained from the advertising
investment and to provide evidence of the return on investment. The advertising planner’s
role is to be able to answer the question of whether advertising is the best way to use the
company’s resources. Some form of advertising research, such as advertising evaluation,
is required to answer the question.
Most advertising research can be divided into two forms. One is used to predict
what might occur in the real-world marketplace On the other hand, evaluation research
also can be used as a basis for future actions. In this case, its basic purpose is to measure
what occurred as a result of the advertising campaign and what was returned on the
advertising investment.
According to Schultz and Barnes (1994), there are three primary reasons for
evaluating an advertising campaign. First, advertising campaign evaluation can be used to
determine whether or not the overall advertising campaign objectives are accomplished.

11
Individual objectives for the various elements of the campaign also can be evaluated and
then, based on these evaluations, the planners can determine why failure may have
occurred.
Second, the return on the campaign investment can be quantified and justified by
knowing what is accomplished. Management also can use the information to explain the
relationship with other potential uses for funds and determine the cost effectiveness of the
campaign.
Third, the measurement information also can be used to change, add, or correct
the campaign plans for future campaigns. Since advertising campaigns can always be
improved, evaluations of previous campaigns and predictions of current campaigns are a
great help to improve the elements and makeup of future campaigns.
Problems in Advertising Campaign Measurement
Although it is important to measure and forecast advertising campaign effects, it
has been difficult for advertising campaign planners to accurately evaluate and predict the
effects of advertising campaigns over the years.
One of the major challenges in measuring advertising campaign effects is what
should be measured. Advertising communication effects, such as consumer awareness of
a brand, or actual behavior of consumers, such as sales or purchases, can be used to
measure and evaluate advertising campaign effects. According to the campaign objectives
established, advertising planners should decide what measures are best. However, no
matter what objectives were set for the campaign, such as sales, market share,

12
communication effects, or some other, there are problems inherent in the measurement
process.
Schultz and Barnes (1994) summarized the problems in measuring advertising
campaign effects. First, there is a problem in differentiating between what is advertising
and what is not. In most cases, advertising is one part of the whole marketing effort.
Certainly advertising and all other marketing efforts, such as sales promotion and public
relations, do not work in isolation. For example, if a customer sees advertising sales
messages carried on the manufacturer’s coupon and remembers it, but not the actual
advertising media message itself, this calls into question whether the advertising campaign
should be credited for the consumers’ awareness of the advertising sales message. It is
very difficult to identify actual advertising effects while controlling for all other important
controllable or uncontrollable dimensions of the marketing mix, such as pricing behavior,
distribution intensity, product quality, competitor target market selection, and other
promotion activities. Usually advertising effects can not be separated from other
marketing efforts.
Second, since most campaigns are advertised over several months or even a year, it
is hard to identify the exact effects of one campaign. In most cases, the advertising
campaign effects are built over time and there are lagged effects of advertising campaigns
Third, the goals of most advertising campaigns are subjective to interpretation.
For example, when the advertising objectives are communication effects, such as
awareness, knowledge, preference, or recall, the measurement of advertising effects often
can not be as precise as those for direct marketing in which consumer response is
measured more accurately.

13
The fourth problem is attributed to human memory. No one can remember all the
advertisements he or she has seen or been exposed to. Due to the lack of information
about how human memory and information storage works, it is certainly difficult to say
exactly what should be measured to evaluate advertising campaigns.
As specified above, skepticism at the possibility of accurately measuring and
forecasting the effects of advertising campaigns has grown over the years. However, in
spite of all these problems, the results of advertising campaigns must be measured and
predicted because of the advertisers’ and advertising planners’ needs and requirements for
accountability. Two frequently used approaches are sales and communication effects.
Advertising Effects on Sales
The purpose of advertising is to provide information that changes consumers’
mental and behavioral responses in a manner favored by the advertiser. However, it is
difficult to measure or predict advertising effects accurately because, typically, there is
more than one response resulting from an advertisement. Therefore, advertisers should
decide what is the best consumer response describing the effectiveness of the
advertisement (McDonald, 1995).
Generally, advertising can be measured and predicted at two levels, either
communication effects or sales effects. Sometimes, advertising effects are measured on
the behavior level, which includes not only purchase behavior but also information seeking
behavior before making a purchase. Direct advertising, which is general for industrial
marketing, makes it easy to measure advertising behavior effects. For example, customers
can receive detailed information about a certain product or service, if only they send a

14
coupon, which is available on a certain part of the advertisement. However, academically
and practically, advertising effects are usually evaluated in terms of communication effects
or sales effects.
Since the ultimate goal of most advertising campaigns is either to increase sales or
market share or to defend achieved positions, many researchers have studied the
advertising-sales relationship (Assumus, Farley, & Lehmann, 1984; Holstius, 1990;
Schreiber, & Appel, 1991; DeKimpe, & Hanssens, 1995a). However, in spite of all the
efforts that have been made to track advertising's effect on sales, it still seems to be an
elusive goal for researchers. Multiple effects exist in advertising and the other
components of the marketing mix. The promotional activities of competitors also easily
confound the measurement of advertising effects on sales (Holstius, 1990).
Difficulties in Measuring Advertising Effects on Sales
One of the major strengths of the evaluation of advertising effects on sales is that
advertising is evaluated in terms of its ultimate objective. However, studies show that
there are some limitations in measuring and predicting advertising effects on sales
(Schreiber, & Appel, 1991; Hoistius, 1990; Lodish et al., 1995). First, although it is true
that the ultimate long-term objective of advertising is to increase sales, it is also
incontestable that advertising frequently has other important short-term objectives, such as
an increase in awareness, knowledge, or recall. It may be true that ultimately all of these
objectives should lead to an increase in sales. However, it is still practically important to
recognize whether these intermediate advertising objectives are achieved or not. The sales
effect measurement approach can not provide the information about the intermediate

15
objectives of advertising. It should not be underestimated that the effectiveness of each
advertising program element can be evaluated solely in terms of its contribution to sales in
the complexity of the typical advertising efforts. In other words, the measured sales
effects only provide whether or not a particular program element meets the advertising
objectives. However, unlike the intermediate factors, the measured sales effects are not
able to explain why the results occurred and what corrective action is needed.
Second, sales are influenced by so many different factors that it is difficult to
isolate the sales effects provided by one particular element of an advertising program from
those provided by other factors. It is also difficult to measure sales response because of
the lag time between the advertising and the resulting sales. Lodish et al. (1995) indicated
that increased advertising weights increased sales of established brands in only 33 percent
of cases.
Therefore, desired advertising effects are mostly stated in terms of communication
effects in both South Korean and American advertising industries. According to Kreshel,
Lancaster, and Toomey’s study (1985), most (79 %) of American advertising practitioners
use communication effects when evaluating advertising media plans. The survey
conducted by the author (1996) also revealed that the majority (90.3%) of the South
Korean advertising practitioners use advertising communication effects to measure or
predict the advertising impact. Only 4.9 percent of media practitioners answered that they
do not use communication effects to predict or measure advertising campaign results.

16
Advertising Communication Effects
Advertising effects are more often measured on cognitive and attitudinal levels
than sales so that they may be better understood. This cognitive and attitudinal method, of
course, provides the advertiser with more detailed information about the performance of
its communication activities (Holstius, 1990). Alternative measures such as awareness of
a brand or advertising messages delivered, attitude toward the brand advertised,
preference, or purchase intention can be used to evaluate the effects of advertising
campaigns.
Sometimes the purpose of advertising is to create awareness and favorable
attitudes that will ultimately change predispositions to companies and their products.
Most techniques designed to measure this kind of effectiveness are based on theories
about the ways in which consumers process information. These techniques help
advertising planners to determine whether the advertising communication was received
and what the impact was on consumers. For example, if an advertisement successfully
increased peoples’ awareness of a product or service, then it would be evaluated to be
effective. In this case, the change in consumer awareness of a product or service is used
to measure advertising effectiveness, and it is presumed that the increase in consumer
awareness ultimately increases sales of the product or service (Cohen, 1988).
The common measures for the measurement of advertising communication effects
are shown in Table 1. Frequently used measures for advertising communication effects
and previous studies are reviewed in the next section.

17
Table 1. Common measures for advertising communication effects
Stages of Information
Process
Common Measures
Coenition
Awareness/recognition
Comprehension
Recall
Affect
Attitudes toward product
Liking
Preference
Conviction
Conation
Intention to try product
Intention to purchase
Actual product purchase
Awareness
Providing either awareness of the brand being advertised or the advertising
message itself is the first level of advertising’s communication effects and the first
objective of an advertising campaign. Especially when the product is new or unknown,
the usual objective of the advertising campaign is to increase the consumers’ awareness of
the brand being advertised. Therefore, awareness of either the brand or the sales message
is the simplest and the most widely used measure of advertising communication effects
(Woodside, 1996). In a typical measurement of advertising communication effects,
awareness is obtained both for the brand and its advertising message. For an existing and
well-known brand, the usual advertising goal is to develop awareness within the target

18
market about a specific benefit of the brand or the consumer problem that the brand can
solve. At this stage, the advertiser does not merely attempt to determine whether the
consumers are aware of the brand. If consumers are aware of an advertising message that
communicates specific benefits of the brand, advertising planners may determine that the
advertising achieved its basic objective.
Advertising communication effects also can be measured in terms of advertising
message awareness, instead of brand awareness. For an established brand that is well
known prior to the beginning of an advertising campaign, the awareness of the new
advertising campaign message can be measured at any time after the campaign begins.
The measurement of awareness is generally accomplished on both an unaided and
aided basis. In dealing with the relative utility of measurements between unaided and
aided awareness, Donius (1986) provides a helpful perspective that focuses on the
function of market fragmentation.
If a brand is in a fractionated, highly competitive market, or if it is new, aided
awareness may be the most useful, sensitive measure for that brand. If, on the
other hand, the brand is a household word, dominates its category and/or is in a
highly advertised category with few competitors, spontaneous awareness is
probably going to prove more sensitive and useful.
According to research conducted by Lowe (1984) in the United Kingdom, the
interrelationships among three measures of awareness, such as top-of-mind awareness,
spontaneous awareness, and aided awareness, were examined with over 16 waves of a
tracking study for two leading oil brands. The results suggest that the three measurements
of awareness are, to a considerable extent, measuring the same thing because the three
measurements tend to move together and at much the same rates. The correlation
coefficients of three awareness measurements were over 0.95.

19
Previous Research on Awareness
To use consumer awareness as a measurement tool of advertising communication
effects, evidence should be presented as to whether consumer awareness of brand or
advertising campaign messages is related to consumers’ preference and purchase of the
brand. In a recent study of advertising for box office performance of new film releases
(Zufryden, 1996), the most significant predictors of consumer intention to see a film is
awareness. This study sequentially establishes a connection between advertising
expenditure and consumer awareness, intention to watch the movie and expected ticket
sales at the box office.
According to Gallup research (Mehta & Purvis, 1994) conducted to learn
consumers’ top-of-mind awareness levels of competing brand advertising, top-of-mind
awareness is also related to preference The empirical results showed a strong and
positive relationship between unaided top-of-mind brand awareness and brand preference
as Woodside (1996) stated.
Determining the size of a brand’s share of consumers who identify that brand’s
advertising in the top-of-mind awareness position may be one useful indicator of
effectiveness of the brand’s advertising relative to the advertising of competing
brands.
In Woodside’s study, the tests of association indicated that top-of-mind awareness
of an advertising message is related to “brand-mind-position” for three brands tested and
also related to preference for four of the brands investigated. These findings suggest that
a positive relationship may occur between top-of-mind awareness of advertising message
and consumer brand preferences. This study also suggests that advertising exposure leads

20
to top-of-mind awareness that may affect brand purchase behavior through brand
preference. These schemes are depicted in Figure 1.
Figure 1. The effects of top-of-mind awareness of the advertisement
Source: adaptedfrom Woodside (1996).
Recall
Advertising planners can use recall to determine whether or not consumers are
exposed to the advertising campaign. Those who are exposed to an advertising campaign
might repeat or play back certain portions of ideas they have seen or heard.
Although recall of sales messages may have some effect on the consumers’ future
purchase behavior, the relationship between recall and future consumer purchase seems to
be somewhat tenuous. According to a recent study by Lodish et al. (1995), the

21
examination of 389 real-world split cable television advertising experiments shows that
recall is not strongly related to sales.
Similar to awareness tests, two types of recall, such as aided recall and unaided
recall, are used in evaluating advertising campaigns. However, since the respondents are
asked only about the product category and are expected to spontaneously remember the
advertising, the unaided recall is believed to be the more useful measure.
When the marketing or advertising manager’s objective is to determine the extent
to which consumers have learned or remembered advertising content, “day-after-recall
(DAR)” has been used as a measure of advertising effectiveness (Higie & Sewall, 1991).
DAR is perceived as an important factor to increase a brand’s probability of becoming a
member of a consumer’s evoked set (Stewart, 1986 and 1989). Therefore, in spite of the
arguments against the use of DAR as a measurement tool of advertising effectiveness,
DAR is frequently employed by advertising managers (Walker & Von Gonten, 1989).
Liking and Attitude Toward Advertisement
Liking assumes that consumers are aware of the brand and have some knowledge
of the brand either from the advertising or from actual product use. Usually, it is difficult
to separate advertising effects on consumer liking from experience, which makes
measurement more difficult. In the liking measure, it is assumed that there is an
advertising effect on the consumer’s mental condition or some kind of attitude change.
The consumers move from the awareness or knowledge stage and form a positive
opinion about the product or service. However, liking does not necessarily mean that the
consumer will purchase the product. It simply means that there exist positive feelings or

22
impressions about the product or service. For example, liking can be measured by having
a consumer name several acceptable brands of products in a category. The assumption is
that if the products are acceptable, they are liked. Although liking is an important step to
evaluate the results of an advertising campaign, the liking measure still does not assume an
actual purchase action. Consumers may like many products but only purchase one or two
(Schultz & Barnes, 1994).
On the other hand, an individual’s evaluation of an advertisement, as measured by
his/her attitude toward the advertisement, is thought to be an important mediator of the
advertising effects on brand attitude. Recently, there has been considerable research
conducted to understand the antecedents of the relationship between attitude toward the
advertisement and brand attitudes (Homer, 1990).
Research shows that attitude toward an advertisement is affected by a brand or
non-brand processing set, such as advertising exposure level (Burke & Edell, 1986; Cox &
Cox, 1988), message involvement (Park & Young, 1986; Muehling & Laczniak, 1988),
advertising message quality/content (Hastak & Olson, 1989), and effective responses
generated during advertising exposure (Burke & Edell, 1989; Machleit & Wilson, 1988).
Attitude toward an advertisement has also been observed to be related to the recall
of an advertisement (Pieters & Klerk-Warmerdam, 1996), purchase intention (Mitchell &
Olson, 1981), and attitude toward the act of buying the brand (Mitchell, 1993).
Biehal et al. (1992) suggested that attitude toward the advertisement also had a
direct and positive effect on brand choice. This may be an important issue for advertisers,
who should be concerned that attitude toward advertisements affect not only attitude
toward the brand and purchase intention but also brand choice.

23
Preference
Preference is defined as one brand being selected over the others among a number
of brands within a certain product category. In advertising campaign evaluation, it is
assumed that the advertised brand is preferred and likely purchased by consumers among
available alternatives, because advertising messages create a level of acceptance for a
specific brand.
Therefore, brand preference is another frequently used measure of advertising
effectiveness and has been of interest to marketers. If the major goal of advertising is to
create product preference, whether or not any actual sales results, the advertising
campaign is usually considered successful if the advertising only creates warm feelings
about the brand. Once a brand is preferred to other brands, it is usually believed that the
advertising or sales messages can do more to trigger the purchase decision. However, the
consumers may or may not purchase the product because there are still many market
factors, such as price, availability, and sales promotion that may influence consumer
buying decisions (Clark, Brock, & Stewart, 1993).
A commercial’s ability to affect brand preference usually is measured as the
difference between before and after exposure to the advertisement (Fletcher & Bowers,
1988). However, only a few studies have examined the reliability of preference. In one
study, Stewart and Furse (1986) estimated fairly good r-squared (0.86) for the use of
preference, using test-retest methods.
In another study, Higie and Sewall (1991) suggested that although measuring
preference is more closely related to actual purchase, preference is less reliable than day-

24
after-recall or other communication effects. They also indicated that it is difficult to
measure significant differences in preference based on a single exposure to an
advertisement.
Purchase Intention and Actual Purchase
If an advertising message is successfully communicated with consumers and other
marketing variables are favorable, the final step of consumer reactions is purchase
intention or actual purchase of the brand Like other measures explained earlier, it is
assumed that the target audience is exposed and reacts to the advertising campaign
messages. The consumers’ awareness of the brand advertised may help develop a liking,
move through preference, and drive to the final step, such as purchasing the brand.
Possible measurement of advertising effects in terms of purchase intention is to identify
advertising as the force that creates a consideration of future brand switching or purchase
(Dutka, 1995).
According to Poiesz and Henry (1994), studies for advertising communication
effects are limited to the investigation of psychological dependent variables, such as
awareness, recall, or attitude toward a brand that are only assumed to precede choice
effects. Although the current availability of product scanner data and other point-of-
purchase systems can provide behavioral data on individual households, purchase
behaviors may be the most difficult concepts to measure because of the many variables
involved, like advertising effects on sales (Poiesz & Henry, 1994).
Nedungadi, Mitchell, and Berger (1993) also argued that although advertising
may have a variety of effects on consumers, an advertising campaign that increases

25
awareness or improves attitudes is only partially successful if the campaign does not
ultimately influence brand choice. Most evaluation questions on this level deal with what
was purchased or what will be purchased in the future. If change, such as brand
switching, did occur, or if there is an indication of future brand switching, research can be
attempted to relate the behavior change to the advertising exposure.
Achieving Desired Communication Effects
The common measures and previous studies of advertising communication effects
have been explained in this chapter. The desired communication effects can be consumer
awareness of a brand or advertising message, recall, attitude change, preference, or
purchase intention. No matter what communication goals an advertising campaign has,
target consumer’s cognitive and attitudinal process is initiated by advertising exposure.
Therefore, one of the fundamental concerns of advertising planning and practice is to
develop and evaluate advertising media schedules in terms of advertising exposure
The desired communication effects typically can be accomplished when target
audiences are exposed to the advertising campaign with optimal frequencies. For
example, if an advertising campaign goal is to achieve 50 percent awareness of the
advertising campaign message, the advertising planner should estimate how much
repetition can help accomplish the goal. Although no study was attempted to examine
how many times a consumer should be exposed to an advertising campaign to achieve
each level of the communication hierarchy described above, it is assumed that the optimal
frequency may vary with the desired communication objectives established. For example,
conative objectives such as purchase intention may need more consumer exposure than

26
cognitive goals such as awareness, because the target audience should first be aware of the
existence of a brand with about 3 to 5 exposures to the advertising message and then,
more advertising exposures after awareness may develop purchase intention through
preference and positive attitude toward the brand. It is assumed that consumers need
more information about the brand to reach purchase intention than awareness. The
additional exposures to advertising message will provide more information about a certain
product or service with the target audiences.
Affective and conative measures also may be influenced more by marketing, copy,
and media factors, such as product quality, advertiser image, and message creativity, than
cognitive measures. In spite of enough exposures to advertising messages, consumers
may not reach purchase intention, for instance, if a brand has poor quality and a weak
advertiser image. Table 2 shows the possible relationship between common
communication measures and needed exposures to achieve the goals in a normal
marketing, copy, and media situation. In other words, when ignoring all other influential
conditions such as marketing, copy, and media factors, optimal frequency of consumer
exposure to advertising to achieve cognitive goals may be lower than affective and
conative objectives. For example, if a brand needs consumer exposure to the
advertisement at least three times to gain 50 percent of awareness of the brand, consumers
might need to be exposed six times to achieve preference, and ten times for purchase
intention. In this case, the minimum numbers of consumer exposures to the
advertisement to achieve awareness (a+), preference (b+), and purchase intention (c+) are
3+, 6+, and 10+ in a normal marketing situation. However, until further study identifies a
scientific method to determine the needed exposure frequency for each level of consumer

27
information processing, it is up to advertising planners to decide how many exposures are
needed to achieve each communication goal established
Table 2. Relationship between common communication measures and needed exposure
frequency in a normal marketing, copy, and media situation.
Common Measures
Minimum Exposure
Frequency Needed for
Average Consumer
Cognitive
Awareness/Recognition
Recall
a+
Affective
Attitude toward brand
Liking
b+
Preference
(a < b)
Conative
Purchase Intention
c+
Actual purchase
(a < b < c)
The concepts shown in Table 2 will be integrated with marketing, copy, and media
factors later in this chapter to estimate optimal frequency level for a selected case study.
Since this study is to estimate the optimal frequency of advertising exposure from the
prospective of media planners, the terms used to explain effective exposure to advertising,
such as reach, frequency, effective reach and effective frequency are reviewed to illustrate
what should be achieved to generate the desired advertising communication effects.

28
Effective Exposure to Advertising
Advertising Exposure
The meaning of exposure is “open eyes or listening ears” facing the advertising
vehicle or message (Sissors & Bumba, 1996). The most common standard to evaluate the
performance of an advertising media vehicle is its rating. Rating can be defined as the
percentage of people exposed to a media vehicle or message and expressed as a
proportion of a selected population base. The rating has been widely used to measure the
performance of media vehicles in terms of advertising vehicle delivery.
However, since audiences exposed to a media vehicle are not automatically
exposed to the actual advertising message, the rating can not indicate what percentage of
the target audience that is exposed to an advertising message within the vehicle.
According to Sissors and Surmanek (19S6),
Vehicle exposure represents an opportunity-to-see (OTS). The number of
opportunities-to-see that a media vehicle develops does not guarantee any
exposure to advertising.
Since it is possible to be exposed to the vehicle without exposure to the advertising
message, generally, the vehicle audience is bigger than the advertising audience. In spite
of the difference between vehicle and message exposure, due to the lack of advertising
exposure data and the technical difficulties to measure accurate advertising exposure,
many media practitioners have frequently used vehicle exposure instead of advertising
exposure (Kreshel, Lancaster, & Toomey, 1985).
However, it is certain that the cognitive process explained earlier is initiated by
exposure to an advertising message, not exposure to the vehicle. Therefore, it is the

29
advertising exposure that should be measured and used for the estimation of media
evaluation factors such as reach, frequency, effective reach, and effective frequency.
In this study, advertising exposure represents exposure to advertising message, not
exposure to the vehicle. The estimation of media evaluation factors is explained based on
the exposure to advertising message in order to measure or predict advertising campaign
effects.
Reach
Recent developments suggest that an important element of the advertising media
decision-making process should be the evaluation of reach and frequency in terms of
advertising exposures or communication effects (Kreshel, Lancaster, & Toomey, 1985).
Reach (sometimes called “coverage") is a measurement of the proportion of the
population that is exposed to an advertising campaign at least once within a given period
of time. Reach is generally explained as a percentage, such as 20, 30, or 40 percent. For
example, if the target market of brand X is 4 million and 2 million of them are exposed to
the brand X’s advertisement at least once within a given period of time, the reach would
be 50 percent. However, since reach is accumulated, the duplication should be subtracted.
Suppose only two magazine advertisements were used for brand X’s advertising
campaign. If the two advertisements were read by 30 and 20 percent of target audiences,
and 10 percent of them were exposed to both advertisements, the reach of this schedule
would be 40 percent [ (30 + 20) - 10 = 40) ].
Reach is usually built in a fairly consistent pattern over time. Using television as
an example, the first time an advertisement is telecast, it accumulates the largest number of

30
viewers. The second time it is telecast, many of the viewers are repeat viewers. For the
third time, even fewer new viewers may see the advertisement (Sissors & Bumba, 1996).
The shape of a typical reach curve is convex as shown in Figure 2.
Households
Reach %
50
40
30
s'
20
f
10
Week
1 2
3
4
5
Figure 2. The shape of a typical message reach curve
Source: adaptedfrom Sissors and Bumba (1996).
Frequency and Frequency Distribution
Frequency is a companion statistic to reach and indicates the average number of
times that target audience members are exposed to an advertising message Whereas
reach is a measure of message dispersion, indicating how widely the advertising message
may be received in a target market, frequency is a measure of repetition, indicating how
many times a target market is exposed, on average, to the advertising message (Sissors &
Bumba, 1996).

31
For the prediction or measurement of the desired advertising communication
effects, such as awareness, recall, preference, or purchase, these two concepts, reach and
frequency, have been widely used (McDonald, 1995). An interesting example was
provided by a research agency, Millward Brown (Brown, 1993), claiming that the increase
in advertising message awareness can be produced by every 100 TVRs (Television Rating
Points).
The average frequency is calculated by dividing the gross rating points (GRPs) by
reach. For example, if the gross rating points of twenty TV advertisements is 80 and the
percent of the target audience exposed to the advertising schedule at least once (reach) is
40 percent, the average frequency would be 2 (80 + 40 = 2).
A frequency distribution is the percentage of the target audience exposed to each
frequency level of an advertising schedule. Since this study uses advertising message
exposure, instead of vehicle exposure, frequency exposure can show the distribution
patterns of message reach at each frequency level. For example, if an evaluation of an
advertising campaign predicted the frequency distribution shown in Table 3, it is assumed
that 70 percent of the target market is exposed to the advertising message at least once
and 30 percent of the target market is exposed to the advertising message more than two
times with the plan.
To achieve desired communication effects, target audiences need to be exposed to
the advertising message a certain number of times, such as three or four times. Since the
required number of exposures may vary with each advertising campaign or plan, frequency
distributions demonstrate more pertinent information than do individual reach and

32
frequency when comparing alternative plans. Frequency distributions also provide
planners with a method of determining the pattern of repetition that the plan may provide
Table 3. Example of frequency distribution
Frequency of Message Exposure
Advertising Plan
1
24%
2
16%
3
9%
4-5
6%
6-8
5%
9-12
5%
+13
5%
Average Frequency
4
Total Message Reach
70
Message Reach at 3+ Level
30
Source: adaptedfrom Sissors and Bumba (1996).
Effective Frequency and Reach
In recent years, one of the most significant changes in advertising media planning
and evaluation is the development of the concepts of effective reach and frequency
(Sissors & Bumba, 1996). With these concepts, media planners can understand a number
of important factors that help decide which plan is better among alternatives or how much
repetition is needed to accomplish the expected advertising communication effects.
Effective frequency is the number of advertising exposures needed to accomplish
desired communication effects on the target audience. Since effective reach is defined as
the percentage of the target market exposed to an advertising campaign schedule at the

33
level of the effective frequency, determining the effective frequency level is one of the
major steps in predicting or measuring advertising campaign effects. Advertising media
planners may decide the effective frequency level needed through test marketing and study
responses to previous advertising schedules. However, sometimes, there is no data by
which to determine objectively how much repetition is necessary to achieve the desired
communication effects In this case, the planner’s estimation may have to be based on
either experience, or specialized research.
Much of what is known about the effects of frequency can be traced to
psychologically trained researchers who explored the subject in a laboratory environment.
One of the major psychologically trained researchers is Krugmen. According to his “three
hit theory” (1972), the first exposure to an advertisement is only perceived by audiences to
raise questions about “What is it?” After the second exposure, the audiences may ask
“What of it?” Then, the second exposure makes audiences react to the advertisement and
begin to compare with alternative brands. The third exposure is a reminder of the other
two exposures and may be the optimal advertising frequency.
Many other studies (Craig & Ghosh, 1993 and Sterling, 1997) have hypothesized
that audiences begin to respond more effectively to advertising from the third exposure to
tenth exposure.
According to Sterling (1997), reach and frequency analysis does not tell us enough
about whether the advertising has been sufficiently exposed to the target audience. The
use of reach (the percentage of target audience members potentially exposed to an
advertising message at least once) is misleading because it often suggests that most any
media plan successfully reaches the vast majority of its intended target audience.

34
Consequently, Sterling suggests that effective frequency can solve such a problem because
effective frequency is the minimum number of times that target audience members should
be exposed to the advertising campaign in order for the advertising campaign to
accomplish its goal. He said that although the number varies from campaign to campaign,
common levels of message effective frequency are between 3+ and 6+. Threshold effects
tend to occur below three exposures. It means that there will be no or little advertising
effects below the threshold level. An S-shaped response ftmction indicates that there is a
threshold effect.
On the other hand, Naples (1979) suggested that the optimal exposure frequency
should be at least 3+ within a purchase cycle and a “convex-shaped” response curve is
expected after three exposures. He also indicated that advertising copy, market share, and
total media expenses are significant factors to decide exposure frequency levels.
According to him, too many exposures can occur negative responses to the brand.
Sissors and Bumba (1996) also argued that many researchers of advertising
response curves have not observed the S-shape, which is the only shape that can explain
the threshold effect. In fact, “convex-shaped” curves have been frequently found. The
advertising effective frequency level may also vary in response to many other marketing,
copy, and media factors, such as consumer involvement with the advertising message or
product, extent of brand loyalty, product life cycle, purchase cycle, creative execution,
extent of clutter, and other specific market or brand conditions (Ostrow, 1982; Sissors &
Bumba, 1996).

35
Problems with Using the Concept of Effective Frequency
Although many studies of the concept of effective frequency (Elliott, 1985;
Kreshel, Lancaster, & Toomey, 1985; McDonald, 1995; Leckenby, & Kim, 1994) showed
that effective frequency is widely accepted among media practitioners in the United States,
there have been a number of questions about the use of the effective frequency concept.
The first problem is the differentiation within product categories. Schumann et al.
(1990) indicated that when subjects were exposed to the advertisements only under
conditions of high rather than low involvement, repeated advertising exposure led to more
positive attitudes toward the advertised brand. Recently, a growing number of media
planners suggest that there should be differences among various product category
frequency levels. Although there is some speculation that, for instance, low-involvement
products may need more frequency than high-involvement products, there is still no
practical measurement to differentiate the frequency levels needed for different kinds of
products (Sissors & Bumba, 1996).
Another problem with the concept of effective frequency is whether a “threshold
effect” exists. As Krugmen suggested, many media planners in the US still believe that
there is a threshold effect and that advertising begins to be effective with the third
impression. However, according to Simon and Arndt (1980), most results of studies
concerning the threshold effect showed that the advertising response curves are mostly
convex, not S-shaped. Furthermore, many direct marketing practitioners objected to the
threshold, arguing that there is no answer about the tremendous consumer responses to
their first impressions (Sissors & Bumba, 1996).

36
The relationship between advertising frequency and its wear-out is another
problem. Studies show that repeated exposure to an advertisement after a certain number
of initial exposures might actually produce negative feelings toward the brand (Calder and
Sternthal 1980 ; Pechmann, & Stewart 1988 ). On the other hand, according to Russell &
Verrill (1986), each additional frequency increases advertising communication effects,
such as recall, recognition, and purchase up to 20 exposures. Therefore, the main problem
that media planners have is when, and under what circumstances, the frequency effects
wear-out. Until further research provides the answer, planners seem to be responsible for
the judgement based on their subjective evaluation.
McDonald’s study (1982) also recommended that product-purchase cycle should
be considered to decide the effective frequency level. However, since each customer in a
product category has a totally different purchase cycle, it may not be useful to help
planners decide an effective frequency level. For example, if a product has a purchase
cycle of two months, each customer may purchase the product on different days during
the 60 days, such as the first day of the first month, or the last day of the second month
(Sissors & Bumba, 1996).
There are more problems to be considered. The quality of advertising message
also may affect the effective frequency level because uninteresting copy may require much
more frequency. As explained earlier, the difference between vehicle and message
exposure also may confuse the decision of an effective frequency level. According to
Lancaster et al. (1986), due to the lack of information or technical difficulties, more than
half of advertising practitioners in the United States use vehicle exposure, in spite of their
understanding of the difference between vehicle and message exposure. However, since

37
the difference between vehicle and message is apparent (Joyce, 1984; Sissors &
Surmanek, 1986; Lancaster & Katz, 1988; Lee, 1996), for the media planners to use the
effective frequency concept, they should focus on how much message frequency is
necessary to achieve the desired advertising exposure. It also might be necessary for
media planners to understand the ratio of message to vehicle exposure.
How to Set Effective Frequency Levels
One of the difficult issues facing any advertising planner is that of deciding how
much effective frequency is enough for an advertising campaign. Since effective frequency
levels vary with marketing, copy, and media factors, those factors should be considered
when choosing an effective frequency level for each advertising campaign.
One suggestion about how to set a frequency level was specified by Ostrow in
1982. According to Ostrow,
The right level of frequency for an advertising campaign is the point at which
effective communication takes place. For example, getting consumers to
understand the advertising message; helping consumers become more positive
about a product or influencing the purchase decisions directly.
He itemized a number of conditions that should be considered based on marketing,
media, and creative strategy factors. For example, new brands need higher frequency
levels than established brands, because new brands need to be learned by the target
market. Planners can use Ostrow’s items (see Table 4) by adding or subtracting points
from a base such as 3+ or 4+. The base levels can be chosen according to the
communication goals established. For example, the base may be 3+ for consumer
awareness of a brand while the base may be increased to 6+ to achieve consumer

38
preference. In Table 4, Sissors and Bumba (1996) suggested specific points with
Ostrow’s categorized items. However, these items and points are only suggestions.
Table 4. Marketing, copy, and media factors that affect effective frequency
Marketing factors that affect effective frequency.
Established brands
(y)-.l +.1 +.2
New brands
High market share
-.2 -.1 +.1 (T|)
Low market share
Dominant brand
in market
-.2 -.1 (+) +.2
Smaller, less well-known brands
High brand loyalty
-.2 -.1 +.1 (2)
Low brand loyalty
Long purchase cycle
-.2 -.1 +.1 +.2
Short purchase cycle
Products used daily
-.2 -.1 +.1 (T2)
+.1 o
Products used occasionally
Needed to beat competition
Copy factors that affect effective frequency.
Simple copy
-.2 (TT)+.l +.2
Complex copy
Copy more unique than
-.2 (TT)+.l +.2
Copy less unique than competition
competition
Continuing campaign
-.2 -.1 +.1 (CT)
New copy campaign
Product sell copy
-.2 -.1 +.1 (2)
Image type copy
Single kind of message
+.1 +.2
More different kinds of messages
To avoid wearout:
New messages
0-1 +.1 +.2
Old messages
Large ad units
0-1 +.1 +.2
Small ad units

39
Media factors that affect effective frequency.
Lower ad clutter
-.2 QjJ+.l +.2
High ad clutter
Compatible editorial
environment
-.2 -.1 +.1 +.2
Non-compatible environment
Attentiveness high
-.2 -.1 +.1 +.2
Attentiveness low
Continuous advertising
^2)-.l +1 +.2
Pulsed or flighted advertising
Few media used
-.2 (TT)+.l +.2
Many media used
Opportunities for media
repetition
£T)-.l +.1 +.2
Fewer opportunities
Source: adapted from Sissors & Bumba (1996) p251-52.
Circled points are applied to the example of brand X described below, while the
remaining factors are assumed to be neutral.
For example, the effective frequency for a certain brand X can be estimated with the
factors in Table 4. If brand X is an established brand which has a strong market leading
competitor with low brand loyalty and occasionally used, the sum of the scales with
marketing factors will be (+0/7). For the copy factors, if brand X has a new advertising
campaign with simple image type copy that is more unique than the competition, single
kind of new message, and large advertising units, the sum of scales applied with copy
factors will be (- 0.41. Finally, media factors can be applied. If the advertising campaign
for brand X is continuous and also has low advertising clutter, high opportunities for media
repetition, and a few advertising media, the sum of scales applied to the example with
media factors will be (- 0.6T When the remaining factors are assumed to be neutral, the
sum of scales applied to brand X with three factors will be (- 0.3), [ (+ 0.7) + (- 0.4) +
(-0.6) = (-0.3)].

40
Therefore, if the advertising campaign goal is to gain 50 percent consumer
awareness of brand X and the media planner decided the base is 3+, the estimated effective
frequency would be 2.7 or 3+. In this case, the base 3+ indicates that at least three
exposures are needed to accomplish the campaign goal when ignoring marketing, copy, and
media factors.
However, different items or scales may be applied for the estimation of effective
frequency levels as the situation dictates. If the brand X company has a relatively good
image (-.1) and high advertising message quality (-.1) but more active competitor
promotional performances (+.2) and relatively poor product or service quality (+.2), the
estimated sum of points will be (-. 1). Table 5 shows the additional points which are
suggested by the author and applied to brand X. It is assumed that the effective frequency
base for advertising message awareness is 3+, then the minimum bases for preference (b+)
and purchase intention (c+) can be estimated as follows:
b > 3 or at least 4+
c > 4 or at least 5+
In addition to Ostrow’s items and Sissors and Bumba’s point scales, the items suggested
by the author can be applied to the bases of awareness (3+), preference (b+), and purchase
intention (c+). Therefore, the modified effective frequency levels for awareness,
preference, and purchase intention estimation are shown in Table 5.

41
Table 5. Suggested additional points and modified effective frequency levels for brand X
Suggested Additional Points
Good advertiser image
Product or service quality
(relatively poor)
-. 1 Advertising message quality (high) -. 1
+.2 More competitor promotional activities +.2
Sum of Points from the Ostrow’s items = (-0.3)
Sum of Points from the additional items = (+ 0.2)
Total Sum of Points = (-
0.1)
Communication Goal
Effective
Frequency
Base
Modified Effective
Frequency Level
Awareness of Brand X
3+
3.0-0.1 =2.9 or 3+
Preference
b+
(b > 3 or at least
4+)
4.0-0.1 =3.9 or
at least 4+
Purchase Intention
c+
(c > 4 or at least
5+)
5.0-0.1 =4.9 or
at least 5+
Since one of the main goals of this study is to apply a frequently used American
media evaluation model to the South Korean market, background information on the
South Korean advertising industry will be reviewed in the next chapter. Survey
research conducted by the author and presented in Chapter III also shows how South
Korean media practitioners perceive the current state of reach and frequency
estimation.

CHAPTER in
BACKGROUND ON THE SOUTH KOREAN ADVERTISING INDUSTRY
Introduction
In this study, a frequently used American media evaluation model will be applied to
a South Korean case study. Therefore, it is necessary to understand South Korean
economic development, the growth and recent changes in the advertising industry,
advertising media, and the present state of South Korean advertising. Also in this chapter,
South Korean media practitioners’ perceptions of media evaluation terms and concepts,
such as effective frequency and the current state of reach/frequency methods are reviewed.
South Korean Economic Development
The Republic of Korea has become the object of growing international interest
because of its rapid rise to world industrial prominence. The South Korean economic
growth has been driven by the export market, which rose from 20.6 percent of GNP in
1972 to 45 percent by 1988. According to Choi (1994):
The composition of the South Korean exports has changed dramatically from
agricultural and simple labor-intensive goods to increasingly more sophisticated
value-added products, including steel, petrochemicals, automobiles, consumer
electronics, computers and large shipping vessels.
42

43
With the dramatic growth of exports, the South Korean per capita GNP has been
increased from $1,012 in 1977 to $4,994 in 1989 (AD DATA, 1995). The total
population of South Korea was 36 million in 1977 and 42 million in 1989. South Korea’s
world trade success has not been simply the result of entrepreneurial talent and the
contribution of high quality, but of low-cost labor as well. South Korean economic
development has been planned, directed, and controlled by aggressive government
intervention and financed by government-controlled banks (Koo, 1993).
The Growth of the Advertising Industry in South Korea
In Korea, the first modern advertisement was seen in the “Han-Sung weekly
newspaper” in 1886; it advertised a German trading house which sold products imported
from Germany (Shin, 1986). However, successive political, economic, and social
disorders deterred the development of a capitalistic economic system. Accordingly,
advertising and industrial activity did not fully develop in relation to other specialized
fields and was generally held in lower esteem than other business. The stages of Korean
economic development and the circumstances of consumer life were so different from
those of advanced countries that consumer awareness of advertising was relatively low
(Yoon, 1980).
After the Korean War ended in the 1950s, the Korean advertising agency business
had to start over because the few Japanese agencies operating in Korea during its
occupation (1910-1945) left Korea after World War II (Shin, 1989). The introduction of
television broadcasting, which was initiated by the American corporation RCA, greatly

44
facilitated the growth of advertising in South Korea (Yoon, 1994). South Korean TV
broadcasting has quantitatively and qualitatively broadened since the introduction.
Along with the influence of the United States on South Korean TV broadcasting,
foreign professionals also initiated the establishment of advertising agencies in the 1960s.
A media representative company, Impact, started its operation as an advertising agency in
1962 (Shin, 1989). The birth of Manbosa Advertising in 1969 led the market entry of
‘COKE’ and ‘Pepsi’ and helped to shape agency recognition and the commission system
(Kim, 1996).
The first Korean Advertising agency, Hap-Dong, appeared in 1967 as a technical
“tie-up” with a Japanese advertising agency, McCann-Erickson Hakuhodo. However, it
was a failure. A successful Korean advertising agency did not appear until the Korean
advertising agency, Cheil Communications, started in 1973. With the success of Cheil
Communications, several advertising agencies such as Oricom and Yon-Hap Advertising
began their business in 1973. The progress of Cheil Communication’s annual billings,
South Korean per capita GNP, and total expenses for advertising in South Korea
increased drastically as shown in Table 6.
The South Korean advertising industry has grown very fast, primarily as a result of
the democratic reforms of 1987, the development of a free press and the effects of the
1988 Summer Olympics held in Seoul. South Korea has a population of about 44 million,
a per capita GNP of US $10,994 (1996), and spent US $6,184 million on advertising in
1995 (AD DATA, 1995).

45
Table 6. The progress of the South Korean advertising industry
Year
Cheil Billings
(US$)
Per Capita GNP
(US S)
Total Expenses For
Advertising
(US $)
1973
1,100,000
361
28,875,000
1975
5,000,000
532
81,250,000
1983
61,000,000
2,014
706,625,000
1994
575,492,500
8,483
5,355,000,000
1995
750,125,000
10,096
6,184,071,250
(US $1 = 800 Won)
Advertising Media in South Korea
In South Korea, newspaper, television, radio, and magazines are the four major
advertising media. Newspaper advertising has grown very fast for the last 10 years and it
covers approximately 56.4 percent of current total advertising expenses in South Korea.
There are about 90 newspapers that can be used as advertising media. The cost for each
1/3 page advertisement with four major national newspapers is about $900 to $2,400 per
advertisement.
TV is the second most widely used medium. About one third of total advertising
expenses (34.3%) are spent on television advertising. Currently, there are three network
television channels that can be used as advertising media in South Korea. Like the United
States, South Korean television advertising is categorized as program advertising and spot
television advertising. Program advertisements are 15, 20, and 30 second in length. The
cost depends on the popularity of each television program and the time when it is on the

46
air. In South Korea, most program advertisements are 15 seconds and the cost range is
$8 to $46 per second. On the other hand, spot television advertisements are on the air
between two programs. Most spot television advertisements are 30 seconds and cost
about $60 to $240 per second.
There are approximately 450 magazines that can be used as advertising media in
South Korea (Park, 1995). Among them, women’s magazines and news magazines for
men are the most widely used advertising media. The estimated cost for a one page
advertisement ranges from $588 to $2,500 in South Korea. The advertising media, total
expenses and costs in South Korea are summarized in Table 7.
Table 7. South Korean advertising media, total expenses, and costs
Newspapers
TV Networks
Magazines
Radio Stations
No. of Advertising
Media
90
3
450
7
Total expenses in
1995 (US$)
2,676,323,000
1,627,854,000
220,170,000
216,734,000
Cost Range
$588 ~ 2,400
for one 1/3
page with
major national
newspapers
Program Ads.
$8-46
per second
Spot TV Ads.
$60 - 240
per second
$540-
2,400
for
one page.
Program Ads.
$3-14
per second
Spot Radio
Ads.
$6-25
per second
(US 1$ = 800 Won)

47
Recent Changes in the South Korean Advertising Industry
Despite its growing market size, the South Korean advertising industry prohibited
foreign investment until the 1980s. It was the United States that estimated the market
potential and began to address market liberalization. During the 1985 American - South
Korean Economic Consultations, liberalization of the South Korean advertising industry
first emerged as a trade issue. In 1986, the American trade representatives became much
more aggressive and then, during the ninth American-South Korean Economic
Consultations in 1988, South Korea announced liberalization of its advertising industry
(Kim, 1990).
Along with the liberalization of the economy and the increased economic power,
South Korean mass media started to be expanded dramatically to satisfy diverse consumer
interests. The number of print media outlets was doubled and the number of weekly
newspapers was increased about five times between 1987 and 1990. Daily newspapers
also increased their editorial and advertising pages (1996, Kim). Broadcasting media
outlets also have increased in number. The new private broadcast network, SBS, opened
in 1991 and emerged as a new media channel
Another potential medium for advertising space and time is cable television. The
South Korean CATV began service in March 1995 with 15 to 20 channels and plans to
increase the channels up to 40 within the next five years. Therefore, the coming CATV
era in the next five years in South Korea will provide much more opportunities for
advertising (Kim, 1996).

48
The opening of the South Korean advertising industry to foreign companies also
brought both the establishment of an audit bureau of circulation and the development of
television rating services that were awaited by the South Korean advertising industry. In
1989, although it has been experiencing difficulties, the Korean Audit Bureau of
Circulation (KABC) was finally established and started to organize circulation audits of
print media. In addition to the KABC, the Korean advertising industry also prompted the
need for a television rating system. Along with the establishment of the Korea Survey
Gallup Polls (KSG) in 1990, three additional media-rating companies began operating
such as Media Service Korea (MSK), Lee’s PR, and Korean Marketing Research (KMR)
(Park, 1995).
The opening of the South Korean advertising industry also prompted advertising
agencies to turn to market research and audience studies. Since the leading South Korean
advertising agency, Cheil Communications, established the first marketing research
institute in 1991, other major large agencies, such as LGAd, Daehong, KeumKang, Korad,
Oricom, and Seoul DMB&B opened their research institutes (Kim, 1996).
Studies on Measuring and Forecasting Advertising Effects in South Korea
Although the studies on measuring and forecasting advertising effects have been
active in the advanced countries, such as the United States and Japan, since the early
1900s, there has been little systematic study of evaluating advertising effects in South
Korea. It is assumed that some South Korean advertising agencies have developed their
own advertising evaluation models that are applicable to the South Korean advertising

49
environment since the late 1980s. However, those models are proprietary and therefore
are not available for access from outside of the agencies (Kim, 1996).
In South Korea, the first study in measuring and forecasting advertising effects was
conducted by Sung in 1989. In his study, “The establishment of a measurable model for
pre-estimation and evaluation of advertising effects in television commercials,” Sung tried
to develop a model of predicting and measuring advertising effects that is applicable to the
South Korean environment. Based on the Japanese JNN Data Bank (JDB) model, as
shown in Figure 3, he suggested that GRPs or exposure to advertisements lead to
awareness that may affect brand purchase through brand preference. Using regression
analysis, he examined the relationship between GRPs and advertising exposure, GRPs and
brand awareness, awareness and preference, preference and purchase, awareness and
purchase, and GRPs and purchase. The results of this study showed that all these six
relationships are statistically significant and different regression equations are suggested.
The independent and dependent variables used in his study and regression equations are
shown in Table 8
However, his study identified problems for future studies Sung assumed that all
the other intermediate factors, such as competitors’ promotional activities and quality of
advertising message that may mediate the relationship between independent and dependent
variables, are the same for all brands studied. His model also may not be applicable to
established brands because he used only new products.
As explained earlier, although some South Korean advertising agencies have been
trying to develop their own models to evaluate and predict advertising effects, most of
them are not published. However, in 1991, LGAd, which is one of the top five advertising

50
agencies in South Korea, developed an advertising effectiveness measuring model (AEM)
and the model was introduced by Lee and Park (1995). In addition to brand awareness,
preference, and purchase, this model also considers advertising message awareness and
advertising preference to predict advertising communication effects (Lee & Park, 1995).
Figure 3. Sung’s model to estimate advertising effects
Table 8. Sung’s independent and dependent variables and regression equations
Independent
Variables
Dependent
Variables
Regression Equations
for Food Products
GRPs
Exposure to Ads.
Y = 100 - 61.207128exp(-0.000125065X)
GRPs
Brand Awareness
Y = 100 - 55.473exp(-0.0000135X)
Brand Awareness
Brand Preference
Y = 2.3 7916exp(0.032628345X)
Brand Preference
Brand Purchase
Y = 1.18511IX + 1.85653
Brand Awareness
Brand Purchase
Y = 2.35262279exp(0.035743144X)
GRPs
Brand Purchase
Y = 100 - 95.2616282exp(-0.000091683X)

51
Although the South Korean advertising industry has been growing very fast, the
evaluation of advertising effects is still in its early stage. Therefore, it is necessary that a
new measurable model of advertising effects that fits to advertising environment of South
Korea should be developed to predict and measure advertising effects. From the two
previous Korean studies, it is also assumed that awareness, preference, and purchase or
purchase intention are the most widely used measures to predict or evaluate advertising
communication effects in South Korea
Media Planning and Evaluation Procedures of South Korean Advertising Practitioners
Chapter II introduced concepts and studies dealing with the aspects of media
planning, such as the kind of communication effects that can be used to predict or measure
advertising effects, how effective reach is typically measured to achieve communication
effects, and the minimum and maximum levels of frequency required for target market
members to be effectively reached.
Studies of the development of media plans, media directors’ perceptions of
effective reach/frequency, and the consideration of weighting and timing factors have been
frequently conducted in the US (Leckenby & Kishi, 1982b; Leckenby & Boyd, 1984;
Kreshel, Lancaster, & Toomey, 1985; Lancaster, Kreshel, & Harris, 1986; Leckenby, &
Kim 1994). However, in spite of the fast growth of the South Korean advertising
industry, no such research has been conducted in South Korea. Therefore, prior to the
application of a frequently used American model to the South Korean market, it may be
important to examine how media directors of South Korean advertising agencies view the

52
current state of reach/frequency estimation methods and which communication effects are
used to predict or measure potential advertising impact.
Purpose of the Study
No information was available about how the South Korean advertising industry
perceives reach/frequency estimation methods and how these concepts are used by the
South Korean advertising agencies. Consequently, this study shows the South Korean
advertising industry’s 1) vehicle/message rating sources, 2) perceptions of the difference
between vehicle and message, 3) definition of reach and effective reach, 4) lower and
upper limit of effective reach, 5) and the use of communication effects, among other
factors.
In this study, the survey results of South Korean media practitioners also will be
compared with previous American studies to inspect similarities and differences between
the two advertising industries. The American studies, which are partially available for
comparison, are Kreshel, Lancaster, and Toomey (1985), Leckenby and Kim (1993), and
Falcheck (1995).
Methods
Currently there are about 120 advertising agencies that can deal with both
broadcasting and print advertising in South Korea. In this study, a mail survey was
conducted of the media directors of these 120 advertising agencies.

53
The questionnaire used in this study was adapted from the one published in 1986
by Lancaster, Kreshel, and Harris. The questions were modified in accordance with the
South Korean situation, such as vehicle rating sources, and translated into Korean by the
author. Both Korean and English versions of the questionnaire are shown in Appendix A.
The questionnaire was mailed to the media director of each agency, if there was any, with
a personalized cover letter, return envelope, and a return postage stamp for international
mail. For the agencies which do not have media directors, a questionnaire was mailed to
senior executives or owners of the agencies.
A follow-up letter, questionnaire, and return envelope were mailed to all non¬
respondents approximately seven weeks after the initial mailing. After two mailings, a
total of 61 responses were received. SPSS PC+ for Windows was used to analyze the
data.
Results
About half (61) of 120 South Korean advertising agencies responded to this study
and 43 of them answered that they develop advertising media plans.
For the titles of the respondents, 36 percent were media directors or vice directors,
3.3 percent were executives, and 32.8 percent were others, not directors or vice-directors
but individuals who work for either media planning or media purchases, and owners of the
agencies. Approximately 28 percent of respondents did not provide their titles.
For the first question regarding sources for television program ratings, 57.4
percent of the South Korean advertising agencies receive the data from Media Service
Korea (MSK). Other sources used in the South Korean advertising industry were Lee’s

54
PR (23%), and Korean Marketing Research (KMR) (26.2%). Since some of the agencies
receive the rating data from more than one source, the sum of percentages for each source
exceeds 100 percent.
For the TV advertising message ratings, 53.5 percent of the South Korean
advertising agencies also use MSK data. Lee’s PR data are used by 8.3 percent of
agencies and 24.6 percent of agencies do not receive advertising message ratings.
This study also shows that the majority (70.5%) of the South Korean advertising
agencies develop media plans for their advertising campaigns. Among the agencies that
develop media plans, the most frequently used systems are non-computerized programs
(39.5%) and in-house computer programs (32.6%) (see Table 9).
Table 9. Systems used to develop advertising media plans (N = 43)
System Used
(%)
In-house computer program
32.6
Foreign computer program
4.7
Non-computerized program
39.5
Others or no answer
23.3
For the next item examining the factors which were generally used to evaluate
advertising media plans, most of the South Korean advertising agencies that develop
media plans answered that they use reach (79.1%) and GRPs (Gross Rating Points)
(72.1%). Other frequently used factors are average frequency (79.1%), CPM (67.4%),

55
and CPRP (72.1%). Two previous American studies (Kreshel, Lancaster, & Toomey,
1985; Leckenby & Kim, 1994) also revealed that reach, GRPs, average frequency, and
CPM are the most widely used factors to evaluate media plans. Table 10 illustrates the
factors used in evaluating media plans in the South Korean advertising industry in
comparison with two American studies.
Table 10. Factors used in evaluating media plans
Factors
US 85
(N = 91)
(%)
US 94
(N = 51)
(%)
Korea 96
(N= 43)
(%)
Reach
90.4
81.0
79.1
Average Frequency
87.2
73.0
79.1
GRPs (Gross Rating Points)
89.4
73.0
72.1
CPRP (Cost-Per-Rating Point)
N/A
N/A
72.1
Effective Reach
86.2
68.3
67.4
CPM (Cost-Per-Thousand)
88.3
77.8
67.4
Gross Impressions
N/A
N/A
2.3
Frequency Distributions
75.5
74.6
N/A
Quintile Distribution
43.6
58.7
N/A
N/A indicates that the factors are not used or no answer.
The evident differences discovered between South Korea and the United States are
the use of CPRP and Frequency Distribution. In spite of little or no use by American
practitioners, the majority (72.1%) of South Korean media practitioners use CPRP to

56
evaluate media plans. On the contrary, frequency distributions are highly used by only
American media practitioners. In other words, the introduction of frequency distributions,
which can provide planners with more pertinent information and a method of determining
the pattern of repetition, can help the South Korean media directors to evaluate their
media plans more effectively.
Definition of Reach and Effective Reach
Although the majority (91.8%) of the South Korean advertising practitioners
answered that the distinction between advertising vehicle and message is important for the
evaluation of advertising campaigns, only 39.1 percent answered that the most
representative definition of ‘reach’ is message exposure. Nearly half of the South Korean
advertising practitioners (53.5%) defined ‘reach’ as vehicle exposure. In other words,
53.5 percent of the South Korean advertising agencies surveyed use vehicle audiences as
a proxy of advertising audiences, while nearly 40 percent of the agencies attempt to
estimate advertising message exposure.
The definition of effective reach applied in the media plan of a particular product
or service can vary across a number of factors, including the media categories and time
frames used. When multiple media categories are evaluated simultaneously, the same
definition of effective reach and time frame must be used for all media categories. This
study found that only half (46.5%) of the media plans developed by the South Korean
advertising agencies include evaluations of schedule reach, frequency, and GRPs for
combinations of two or more media categories.

57
The South Korean advertising practitioners were then asked to note the general
definition of effective reach. The majority of the South Korean advertising practitioners
recognize that the definition of effective reach requires a minimum frequency level. About
93 percent answered that a minimum frequency of two or more should be the lower limit
of effective reach and 74.4 percent answered it requires a frequency of three or more,
while 11.6 percent answered that a frequency of four or more is needed.
For the upper frequency limit, 41.9 percent of the respondents recognize no upper
frequency limit to effective reach, while 37.2 percent set some kind of upper limit, ranging
from 9 to 11. In spite of the different environmental situation, two American studies
(Kreshel, Lancaster, & Toomey, 1985; Falcheck, 1995) also show similar results to this
study Table 11 shows the South Korean results in comparison with the previous
American studies.
Communication Effects Used
The majority (90.3%) of the South Korean advertising practitioners answered that
they use communication effects to evaluate their advertising media plans. The two most
frequently used variables are exposure to advertising and awareness/recognition. Among
those who use communication effects to evaluate media plans, 80 percent answered that
they use advertising exposure. The second most frequently used variable is awareness or
recognition (46.5%). Since there is no distinct difference between awareness and
recognition in the Korean language, these two terms are treated as the same expression.

58
Table 11. Representative definitions of effective reach in comparison with two previous
American studies
Message Impact Required
US 85
(N = 91)
(%)
US 95
(N = 50)
(%)
Korea 96
(N = 43)
(%)
Vehicle Exposure
48.4
70.0
53.5
Message Exposure
31.9
13.3
39.5
Advertising Impact
16.5
13.3
7.0
Others and no answer
3.3
3.3
Number of Exposures Required Regardless of Message Impact Level
Lower Limit
US 85
(Vo)
US 95
(Vo)
Korea 96
<%)
1+
6.6
13.3
4.7
2+
5.5
6.7
18.6
3+
61.5
50.0
62.8
4+
17.6
16.7
11.6
Others and no answer
8.8
3.3
2.3
Upper Limit
US 85
(Vo)
US 95
(%)
Korea 96
(%)
No Upper Limit
58.2
46.7
41.9
9
4.4
6.7
4.7
10
6.6
10.0
20,9
11
2.2
0.0
11.6
Others and no answer
28.6
13.3
21.0

59
Table 12 illustrates the variables used to evaluate advertising communication
effects in the South Korean advertising industry in comparison with a previous American
study (Kreshel, Lancaster, & Toomey, 1985). Though the American study was conducted
about ten years before this study, the two studies show similar results. The only
observable differences are that Korean agencies use attitude toward advertisement or
brand more frequently than American agencies, while American agencies use recall more
than Korean agencies.
Table 12. Variables used to evaluate advertising communication effects
Variable
US 85
(N=94)
(%)
Korea 96
(N=61)
(%)
Variable
VS 85
(N=94)
(%)
Korea 96
(N=61)
(%)
Exposure to
52.1
79.1
Preference
18.1
32.6
advertisement
Awareness/
47.9
46.5
Willingness to
12.8
32.6
Recognition
24.5
purchase
Attitude toward
11.7
44,2
Purchase
31.9
23.3
Ads.
Recall
58.5
37.2
Comprehension
16.0
20.9
Interest
160
34.9
Conviction
8.5
14,0
Attitude toward
18.1
32.6
Others
9.6
4.7
brand
Message Weighting and Reasons of Not Weighting
The media practitioners of the South Korean advertising agencies that develop
media plans and evaluate their plans (about 70 percent of all South Korean agencies) were

60
asked whether they use such information to weight vehicle audience data to estimate
advertising message exposures or other desired communication effects.
Among 43 agencies that develop media plans, 34.9 percent indicated that the data
are used to weight vehicle audiences. More than half (56.3%) of the agencies which
weight audience data use individual judgements to derive weights, while only 18.8%
derive weights from a formula or an established standard.
The respondents who do not weight vehicle audiences gave approximately six
different reasons for not doing so. The most common of these are that media planning
situation for each brand or service is unique and weighting is too difficult to use
accurately. Lack of information about weighting procedures and not enough time are two
other common answers.
Needed Improvements
Almost all respondents (96.7%) answered that their agency’s current procedures
for evaluating advertising media plans are in need of improvement.
The majority (82%) of South Korean advertising practitioners indicated that
recognition of the importance of media planning should be improved. Nearly 80 percent
of them also answered that training of media specialists is the second most important thing
to improve current procedures to develop and evaluate advertising media plans (78.7%).
The South Korean advertising media practitioners showed split responses to the
introduction of foreign media planning procedures, such as American reach and frequency
computer programs, for better media planning. About 34.9 percent answered that the

61
foreign procedures would be helpful for better media planning, while 32 6 percent replied
that the foreign procedures would not be helpful due to the different media environment.
Discussion and Implication for Further Research
In the United States, models for estimating reach and frequency distributions have
made increasing use of the concept of effective reach and frequency among both
practitioners and academicians (Kreshel, Lancaster, & Toomey, 1985; Lancaster, Pelati, &
Cho, 1991; Leckenby, & Kim, 1994). The results of this study clearly demonstrate that
the majority of the South Korean advertising agencies are also aware of the media
evaluation terms such as effective frequency, and more than half of the agencies actually
use reach/frequency estimation methods.
In spite of the different environments, the South Korean media practitioners’
representative definition of effective reach is also close to the American practitioners as
shown in Table 11. The majority of both American and South Korean media practitioners
answered that the most representative definition of effective reach has 3+ as the lower
limit and no upper limit.
In addition to the definition of effective reach, the majority of both American and
South Korean media practitioners also use communication effects to evaluate advertising
campaign effects and media plans. Among the common measures for advertising
communication effects, advertising exposure is the most frequently used measure in both
countries. Awareness is the second most widely used measure in South Korea and third in
the United States

62
With the beginning of CATV in South Korea since 1995, the South Korean
advertising industry started to recognize the importance of media planning and evaluation
of media schedules. According to this study, about 96.7 percent of the South Korean
advertising agencies claim that their current procedures for developing and evaluating
media plans should be improved. The South Korean media directors also indicate that
systematic media evaluation methods that fit the South Korean environment are needed in
the near future.
Although only 34.9 percent of the South Korean media practitioners responded
that the introduction of foreign procedures would be helpful for better media planning and
campaign evaluation, 32.6 percent replied that the procedures would not be helpful The
application of a frequently used American reach and frequency computer program to the
South Korean market will help develop a model that fits the advertising environment of
South Korea.
This study indicates that reach/frequency estimation is a good method to predict or
evaluate advertising communication effects in the South Korean advertising industry
because of the wide acceptance and actual use of reach and frequency as shown in Table
10. Furthermore, since the South Korean media practitioners also define the general
effective frequency level similar to American media directors as shown in Table 11, it is
assumed that in South Korea the effective frequency levels for common communication
measures, such as awareness, preference, and purchase intention, can be decided using the
procedures explained in Chapter II. Based on the optimal exposures for common
communication measures, which are used as bases to estimate the advertising

63
communication effects, the effective frequency level can be decided for each advertising
campaign goal with consideration of marketing, copy, and media factors.
A test market study is a good method for examining these assumptions.
Therefore, Chapter IV, Methodology, will introduce the brand used as a case study, along
with the normative theories, methods, and procedures required to evaluate it.

CHAPTER IV
METHODOLOGY
The primary purpose of this study is to demonstrate how accurately media
exposure distribution models can predict advertising communication effects. In particular,
one goal is to apply such a model to the South Korean market. In this test market study,
advertising campaign communication effects are “predicted” by a variation of the beta
binomial exposure distribution model based on the information available before the
advertising execution Then, the impact of the advertising campaign on its target audience
is measured after a three-month execution. The predicted communication effects will be
compared with the actual target consumer responses to the advertising campaign to probe
the capability of this normative framework.
Generally, as shown in Table 13, regardless of the country or brand being studied,
particular items need to be gathered or defined to complete this type of study. These
items are 1) target audience definition, 2) vehicle data, 3) message data, 4) advertising
expenditures, 5) a media evaluation model, 6) a narrow time frame, 7) tracking study
results, 8) and the media plan used.
Advertising planners who wish to accomplish this type of research should follow
these procedures. After defining the target audience, advertising message and vehicle
ratings and expenditures should be gathered prior to the period of the campaign in order
64

65
to predict the campaign effects. It is also important to evaluate the plans over a narrow
time frame, such as monthly, due to target audience forgetting, competitive advertising,
and other factors that reduce advertising carry-over effects. For example, annual reach-
frequency analyses will grossly overstate likely communication effects because it fails to
account for these factors (Lancaster, Kreshel, & Harris, 1986).
Table 13. Normative framework necessary to predict advertising communication effects
Items
General Meaning
Target Audience
Definition
The desired or intended audience for an advertising campaign.
Usually defined in terms of specific demographic, purchase, or
ownership characteristics.
Vehicle Data
Ratings of particular component of a media class, such as a
particular TV program or magazine, used for an advertising
campaign.
Message Data
Ratings for actual advertisements, not the vehicles.
Advertising
Expenditures
Total media expenses for an advertising campaign.
Media Evaluation
Model
Computer program that embodies normative theories of media
planning, including vehicle and message ratings and duplication
plus corresponding exposure distributions.
Narrow Time
Frame
Due to target audience forgetting, competitive advertising, and
other factors that reduce advertising carry-over effects, a narrow
time frame for analysis, such as monthly, is preferred.
Tracking Study
Results
Actual target audience response to the advertising campaign in
terms of the advertising goals established.
Media Plan Used
Advertising media schedules during the period of the campaign.

66
A media evaluation model such as the beta binomial exposure distribution model
can be used to predict the campaign communication effects. Then, consumer tracking
studies will show actual consumer response to the campaign.
In this chapter, after introducing the brand and advertising campaign used in this
case study, each item in Table 13 will be explained again with reference to the information
available for the case.
Case Study
ASIANA Airline
A South Korean airline company is selected to test the framework of predicting
and measuring advertising campaign effects introduced above. The South Korean
domestic and international air transportation business has grown rapidly (about a 12
percent annual increase in passenger transportation volume) since Korea hosted the 1988
Olympics. Seoul's growing importance in international air traffic, good international
geographic location, congested traffic on the highways, and drastic increase in GNP (see
Table 6 in Chapter III) since 1988 accelerated the growth of the airline business in South
Korea.
Currently, there are two airline companies in South Korea. The Korean airline
(KAL) had been the only air carrier in South Korea until ASIANA started its business in
1990. In 1993, both KAL and ASIANA carried 15,655,000 passengers on their domestic
lines and 11,651,000 on international flights. Cargo services also have grown impressively

67
since the opening of American routes in 1971. The number of employees of ASIANA
airline is about 7,900.
Domestically, KAL has maintained a 70 percent market share since 1990, while
ASIANA has held only 30 percent market share for the last 7 years. Internationally, the
market share of both Korean air carriers was only about 50 percent until 1991 because of
dynamic foreign airline companies, such as Delta Airlines, TWA, United Airlines, British
Airways, and Singapore Airline. Many foreign airlines have been attracted to Seoul’s
growing importance and have opened regular services to and from Korea. As of 1993,
about 26 foreign airlines are operating in Korea. However, the international market share
of both Korean air carriers has been increasing since 1991. While the KAL holds its
international market share at about 47 percent, ASIANA airline has been growing fast for
the last 5 years. The domestic and international market shares for both air carriers are
shown in Table 14.
Current consumer research conducted by SangAm advertising agency shows that
the ASIANA airline company has maintained a good company image on its service
evaluation because of perceived flight attendant service and kindness. However, many
consumers still favor KAL because of safety, tradition, and mileage bonuses. It is
assumed that promotional activities such as advertising and public relations play a major
role in the recent successful growth of ASIANA. Since advertising is an important means
of promotion for ASIANA airline, it is necessary to conduct consumer research to
examine and predict its current advertising effects and to help develop a more efficient
advertising campaign in the future. This case study also provides an opportunity to test
and advance normative advertising theory.

68
Table 14. Domestic and international market shares
Domestic market share
1990
(%)
1991
(%)
1992
(%)
1993
(%)
1994
(%)
1995
(%)
1996
(%)
Korean Airline
70
71
71
70
69
69
69
ASIANA Airline
30
29
29
30
31
31
31
Total
100
100
100
100
100
100
100
International market share
1989
(%)
1990
(%)
1991
(%)
1992
(%)
1993
(%)
1994
<%)
1995
(%)
1996
(%)
Korean
Airline
52.3
45.3
44.8
43.8
47.3
47.4
47.1
46.9
ASIANA
Airline
-
2.0
6.0
8.9
12.1
15.3
17.4
17.8
Total
52.3
47.3
50 8
52.7
59.4
62.7
64.5
64.7
Source: adaptedfrom “SangAm ” advertising agency.
Target Audience and Advertising Campaign
The annual expense for ASIANA advertising was about US $5,920,000 in 1994
and ASIANA is one of the 150 biggest advertisers in South Korea (ranked 101 in 1994 in
‘Kwango Yeonkam 95’).

69
In 1996, ASIANA airline completed its new advertising campaign, “Her name is
ASIANA,” focusing on ASIANA’s better service and kindness. Since the new campaign
was performed by SangAm, one of the major Korean advertising agencies, most of the
information needed for this study, such as consumer survey data, were supplied by
SangAm.
According to the research conducted by SangAm (1996), the primary target
market of this campaign was defined as 20 to 40 year old men who have used any air
carrier at least once. However, there are some difficulties in using this target market
definition. First, previous use of any air carrier makes it difficult to estimate the described
target market size. Second, the geographic area of consumer research conducted by
SangAm was limited to Seoul, the capital city of Korea, where one fourth of South
Korean people live. Consequently, although the primary target audience is defined as 20
to 40 year old men who have used any air carrier at least once, the consumer survey was
conducted with 20 to 40 year old Korean men who live in Seoul. The target market size
used for the consumer survey is about 3 million (2,884,983 in 1995).
ASIANA airline first advertised its new campaign of “Her name is ASIANA” from
February 17 to May 16 in 1996. Two different types of messages were advertised on
television and in magazines. One was called “piano scene” and the other was called
“poetry situation.” Both types of the actual magazine advertisements are in Appendix B.
The summary of the advertising insertions by media category and expenses of the
campaign is shown in Table 15.

70
Table 15. Advertising insertions by media category and expenses of “Her Name is
ASIANA” campaign from February 17 to May 16, 1996
Media
Insertions
Total Expenses
(VSS)
KBS-2TV
96
$414,373
MBC-TV
137
477,035
SBS-TV
105
259,441
Weekly Magazines
33
60,750
Monthly Magazines
9
19,875
Grand Total
380
$1,231,474
Source: adapted from marketing department of SangAm andMSK.
(US $1 =800 Won)
Competitor
The market leader, KAL, which started its business in 1969, is a strong competitor.
The first name of KAL was Korean National Airlines and was managed by the
government. Later the name was changed to KAL and the management was turned over
to private hands. Because of the long tradition and previous monopoly, KAL has
maintained its market share at about 70 percent domestically and 47 percent
internationally.
However, compared with ASIANA’s rapid growth, the market share growth rate
of KAL has been almost zero for the last four years (see Table 14). KAL’s annual
advertising expense was US $11,932,000 in 1994 and KAL has been one of the 50 biggest
advertisers in South Korea (ranked 39 in 1994) (Kwango Yeonkam 95). KAL’s

71
advertising expenses were approximately three times higher than ASIANA during the
period of ASIANA’S first “Her name is ASIANA” campaign. Overall advertising GRPs
and expenses from February through April of 1996 are provided in Table 16.
Table 16. Competitor GRPs and advertising expenses
Media
Month
GRPs
Expenses (US $)
TV
February
969.58
$499,130
March
1158.18
549,560
April
1072.82
524,167
Total
3200.58
1,572,857
Radio
February
272.89
90,395
March
273.54
97,407
April
322.18
93,980
Total
868.61
281,782
Newspaper
February
742.62
889,327
March
236.86
309,484
April
261.58
267,725
Total
1241.06
1,466,536
Magazine
February
89.73
63,625
March
82.10
46,813
April
45.63
35,500
Total
217.46
145,938
Total
February
2074.82
1,542,477
March
1750.68
1,003,264
April
1702.21
921,372
Grand Total
5527.71
$3,467,113
Source: adaptedfrom MSK (Media Service Korea).
(US$1 =800 Won).

72
Media Data
TV Ratings
Television program ratings are collected from Media Service Korea Inc. (MSK)
which is one of the major media research institutes for ratings in South Korea. SangAm,
the advertising agency for ASIANA airline, provides advertising message ratings that are
also collected from MSK. Media Service Korea Inc., founded in 1987, is the only media
research institute in South Korea that uses people-meters to collect its ratings from a total
of 1,200 panel members who live in Seoul. Audience panels were established by MSK and
the panel members were randomly selected based on four critical demographic variables,
such as age, occupation, education, and income. Each panel member’s daily television
watching data are gathered in the MSK central computer from 6 A M to 1 A M and all the
television program and advertising message ratings are reported daily after aggregating
them.
The ratings are also provided by important variables, such as gender and age. It is
possible to gather demographic information with the use of separate remote controls
distributed by MSK Each panel member is assigned to a number and is supposed to input
the assigned number while he/she is watching television. For example, if a panel
household consists of four family members, such as a father, mother, son and daughter,
each family member is assigned to a number from 1 to 4 and is instructed to push the
number button on the remote control. Therefore, in this study, program and advertising
message ratings were gathered for the target audience group. According to MSK, these
ratings are significant with 95 percent confidence levels (p < 0.05). The average sampling

73
error is within ± 0.025. Television program ratings, advertising message ratings, and costs
provided by the ‘ SangAm’ agency are listed in Appendix C.
Television program and advertising message ratings are also collected from Lee’s
PR which is another popular research institute for ratings in South Korea. These are often
used in comparison to MSK. While MSK uses panel members and people-meters, Lee’s
PR conducts face-to-face interviews once a month to gather television ratings from a total
of 1,300 sample members. However, since MSK data have been used by SangAm, MSK
data are used in this study unless otherwise noted. According to the South Korean media
practitioner survey conducted by the author and reported in Chapter III, MSK is also the
most frequently used vehicle and message source in South Korea.
Magazine Ratings
Magazine advertising expenses of $80,625 were 6.5 percent of total media
expenses of $1,231,474 for the ASIANA campaign. Eleven weekly magazines and three
monthly magazines were used for the ASIANA advertising campaign. The circulation and
ratings for these magazines were collected from each magazine company because of the
lack of independent information about magazine ratings in South Korea. A Korean
advertising practitioner called each magazine company and collected the ratings of these
14 magazines. The ratings may be higher than actual ratings because they were gathered
from each magazine publication company. In the future, the magazine ratings should be
gathered from an objective media research institute for better prediction of advertising
campaign effects. Magazine circulation, ratings, and costs are listed in Appendix C.

74
Tracking Data
The tracking data used in this study were gathered from the advertising agency of
ASIANA Airline, SangAm. The main purpose of the survey research was to evaluate the
impact of the new ASIANA advertising campaign, “Her name is ASIANA” and to prepare
the next ASIANA advertising campaign. After the three month advertising campaign from
February 17 to May 16, a post test with a total of 200 samples of the target audience was
conducted in May, 1996.
Sampling
Since the target audience was defined as 20 to 40 year old men who live in Seoul,
Korea, the sample was randomly drawn among the target audience. In this survey
research, “stratified sampling” was used, which separates samples from each of several
subpopulations. The subpopulations were defined by SangAm in advance on the basis of a
critical subject variable, such as age, that may influence scores on the dependent measures.
Rather than relying on random sampling, the target population was divided into
subpopulations on the basis of age to create a total sample by selecting the appropriate
proportion of subjects from each of the subpopulations. For example, if 20 percent of the
target population is between ages 20 and 29, the number of subjects that would represent
20 percent of the total sample was selected from that subpopulation. A total sample of
200 individuals was drawn from three different age groups. Forty percent of the total
sample was selected from the 30s age subpopulation, and another forty percent was drawn
from the 40s age subgroup. The age groups, number of sample individuals in each age
group, and sample characteristics are shown in Table 17.

75
Table 17. Sample characteristics
Number in
Sample
(%)
Age
20-29
40
20.0
30-39
80
40.0
40-49
80
40.0
Total
200
100.0
Education
Up to High School Education
43
18.5
College Education or More
157
81.5
Total
200
100.0
Marital Status
Single
37
21.5
Married
163
78.5
Total
200
100.0
Occupation
Specialist or Freelancer
19
9.5
Management Staff
42
21.0
Office Worker
110
55.0
Self-employed
29
14.5
Total
200
100.0
Source: adaptedfrom the marketing department of SangAm.
Measurement
Since the purpose of this survey was to evaluate the new ASIANA advertising
campaign effects, only a post test was conducted. In this survey, a face-to-face interview
was used to gather target audience information about the advertiser image, advertising
campaign message awareness, evaluation of the advertising message quality, preference,
and willingness to use in comparison with competitors. A total of 18 interviewers, who
were college students or housewives, were temporarily hired by the advertising agency of
ASIANA Airline to conduct the personal interviews for seven days beginning on May 25,
1996. Three questions that were asked of the respondents and which are directly related
to this study are: 1) “Are you aware of the new ASIANA advertisement, ‘Her name is

76
ASIANA’?” 2) “Which airline company do you prefer?” 3) “Which airline are you going
to use for your next trip?” After data were completely collected, 20 percent of randomly
selected responses were verified again before editing, coding, and analyzing the data.
Then, the SPSS PC+ program was used to show frequencies and cross-tabulations. These
tracking data will be used to examine how accurately the normative media evaluation
framework of this study can predict advertising communication effects.
Media Evaluation Model Used in This Study
Beta Binomial Exposure Distribution Model
A variety of computerized media evaluation models have been developed and used
in both the advertising industry and academia over the past thirty years. The availability of
computers in the 1960s led to the development of a number of mathematical media
planning models. The first level of sophistication in the quantitative evaluation of a media
plan involves estimating a media schedule’s reach and frequency. Another related concept
is gross rating points (GRPs), which refer to the total number of exposures generated by
the schedule. In 1961, Agostini developed a simple reach estimation model and in 1964,
Metheringham tried to estimate unduplicated audience measurement (Craig & Ghosh,
1986). They provided the impetus for the development of media models. After that,
media evaluation models have become much more sophisticated, including the beta
binomial models, beta matrix method, heuristic models, and variations or extensions of the
beta binomial models (Rust, 1986).

77
Most recent models deal with the sophisticated calculation of the exposure
distribution by the number of exposures of a given media schedule and they provide all
information relevant to the advertising media schedule The general goal of these models
is to evaluate advertising media schedules by estimating reach, frequency, effective
frequency, effective reach, and exposure distributions.
Exposure distribution models are divided into two groups in terms of estimation
dimensions that include a number of parameters, such as univariate probability distribution
models and multivariate probability distribution models. Univariate probability distribution
models treat all vehicles as one univariate and composite vehicle and then average all
vehicle insertions of a schedule (Ju et al., 1990). On the other hand, multivariate exposure
distribution models provide individual vehicle audience information by preserving as many
estimation parameters as vehicles in the schedule. Though the complexity means that
multivariate models require more data manipulation and computation time, they are
considered to provide more accurate estimation than do univariate models. However,
since true multivariate probability distribution methods that estimate broadcast audiences
have not been developed, most multivariate models are magazine audience estimation
models that are tested in schedules with no more than two insertions in each vehicle
(Leckenby & Kim, 1992).
More than fifteen univariate probability distribution models have been developed
since Agostini’s simple reach estimation model. Some include the beta binomial
distribution (BBD), Markov binomial, Morgenzstem, and Hofmans geometric distribution
(HGD) (Leckenby & Ju, 1990). Among them, one of the most widely used and studied
exposure distribution models is the beta binomial distribution (BBD) model because of the

78
simple, parsimonious, and relatively sound estimation (Ju & Leckenby, 1990; Kreshel,
Lancaster, & Toomey, 1985).
The BBD model, which was first introduced by Metheringham in 1964, is a
statistical procedure for estimating the likelihood of a target audience exposure to a
schedule a given number of times. The binomial distribution estimates the number of
different ways a target audience can be exposed to a schedule from one to N total
insertions. The beta distribution estimates the average probability of exposure to the
schedule from one to N times (Lancaster & Katz, 1988).
The BBD model has been used as a basis of estimating the exposure distribution of
not only univariate but also multivariate models The notable univariate variations of the
beta binomial distribution are BBD-DE, Hofmans BBD, and BBD simulation. The
multivariate analog of the BBD is the Dirichlet distribution.
Although the BBD has been accepted to be fairly accurate across a wide variety of
advertising media circumstances, the use of the BBD to evaluate an entire media category
schedule has some limitations. One of the major problems is that the use of BBD treats all
vehicles and pairs of vehicles in a schedule as a single average vehicle Therefore, using
the BBD alone can mislead the estimation of reach and exposure distributions, particularly
in the case of disparate insertions and/or ratings among the vehicles in a schedule
(Lancaster 1992). It is also known that the BBD model often yields an over-estimation in
reach (Leckenby & Rice, 1985; Leckenby & Kim, 1992), However, the BBD model has
been used as a mathematical basis to predict advertising message exposure distributions of
advertising schedules because of the wide acceptance, proven reliability, and possible
variations over more than two decades (Leckenby & Boyd, 1984; Ju et al., 1990). For

79
example, due to the lack of pair-wise duplication data such as vehicle self-pair and cross¬
pair ratings, many studies use a variation of the BBD model which is the beta binomial
distribution with limited information (BBD-L). The BBD-L model, developed by
Lancaster and Martin (1988), does not require pair-wise duplication data and was shown
to be accurate in predicting pair-wise duplication among vehicles.
Model Selection
Based on the BBD model, a computerized media evaluation program, ADplus, is
used in this study because the program is comprehensive, widely available, and also can
handle vehicle and message reach and frequency.
ADplus, originally developed by Lancaster in 1990, is an easy-to-use program that
helps predict advertising campaign effects and evaluate advertising media plans. The
program is currently distributed worldwide by Telmar Information Service Corp. of New
York (http://www.telmar.com) and is used by thousands of media marketing professionals
worldwide. It is possible to evaluate all major advertising media categories and
subcategories with ADplus. It also helps advertising media planners find optimum
schedules involving multiple media categories.
As input data, ADplus needs advertising media vehicle ratings, target market size,
advertising cost, and number of insertions. As shown in Figure 4, ADplus results provide
media planners with frequency distributions, reach, effective reach, gross rating points
(GRPs), average frequency, gross impressions (Gis), cost-per-thousand (CPM), and cost-
per-rating point (CPP). ADplus also provides two types of frequency distributions,
including vehicle and message frequency distributions. For the magazine example in

80
Figure 4, the message effective reach (3+) is 0.1 percent. Therefore, if it is assumed that
consumers are aware of the new advertising message after three exposures to the
magazine advertisement (effective frequency), only 0.1 percent of the target audience is
estimated to have been exposed at this level.
Although the typical vehicle and message effective reach of 3+ is shown in the
ADplus results, media planners also can examine other definitions of effective frequency
and reach that may be more fitting with their situation including, for instance, effective
frequency and reach of 4+ or 5 through 10.
ADplus can overcome the problems of the BBD model which treats all vehicles
and pairs of vehicles in a schedule as a single average vehicle. With the estimation of
vehicle self-pair reach, ADplus can estimate separate vehicle exposure distributions. A
matrix of individual vehicle exposure distributions is then used to estimate multiple vehicle
exposure distributions in conjunction with vehicle cross-pair reach data. This procedure is
referred to as a beta binomial matrix exposure distribution (BBMD) model (Lancaster,
1993). The technical procedures used to implement the BBMD model are described in
detail by Lancaster (1993) and important dimensions of this model are presented in
Appendix D.
The ADplus program also helps media planners find an optimum combination of
media vehicles and media categories within desired budgets and recommends media
budgets for desired communication goals.

81
ADplus(TM) RESULTS: MAGAZINES
By: Hyunsoo Park
Frequency (f)
Distributions
University of Florida
For: AS I ANA Airline
Vehicle
Message
Typical monthly schedule
f
% f
% f+
% f
% f+
Target: 2,884,983
—
—
—
—
—
Korean men 20-49 years
0
65.1
100.0
82.6
old who live in Seoul
1
28.4
34.9
15.9
ill .4s
2
5.8
6.5
1.4
v
Message/vehicle = 45.0%
3
0.7
0.8
0.1
0T1
4
0.1
0.1
0.0
0.0
5
0.0
0.0
0.0
0.0
6
0.0
0.0
0.0
0.0
7
0.0
0.0
0.0
0.0
8
0.0
0.0
0.0
0.0
9
0.0
0.0
0.0
0.0
10+
0.0
0.0
0.0
0.0
Summary Evaluation
Reach (1+)
34.9%
izat.
Effective reach (3+)
0.8%
Gross rating points (GRPs)
42.2
l57o
Average frequency (f)
1.2
1.1
Gross impressions (000s)
1,
266.0
569.7
Cost-per-thousand (CPM)
21.23
47.17
Cost-per-rating point (CPP)
637
1,
415
Vehicle List Rating Ad Cost
CPM-MSG
Ads
Total Cost
Mix
Sisa Journal 6.60
2,250
25.25
1
2,250
8.4%
Han Kyurei 21 4.40
2,000
33.67
1
2,000
7.4
Jugan Chosun 5.40
2,250
30.86
1
2,250
8.4
News Plus 2.80
1,875
49.60
1
1,875
7.0
Jugan Hankuk 1.60
1,500
69.44
1
1,500
5.6
News Week 2.80
2,250
59.52
1
2,250
8.4
News people 2.00
1,500
55.56
1
1,500
5.6
News Maker 1.70
1,500
65.36
1
1,500
5.6
Jugan MaeKyun 1.60
1,750
81.02
1
1,750
6.5
Economist 1.00
1,750
129.63
1
1,750
6.5
HanKyung Busi 0.50
1,625
240.74
1
1, 625
6.0
Shin Dong-Ah 5.30
1,875
26.21
1
1,875
7.0
WalKan Chosun 5.30
1,875
26.21
1
1, 875
7.0
Win 1.20
2,875
177.47
1
2,875
10.7
Totals:
47.17
14
26,875
100.0%
Figure 4. ADplus Results

82
Model Modification
As shown in Appendix C, the media evaluation model used in this study needs
regression equations for the estimation of R2t and RtJ. R2l is reach as a percentage of the
target audience of one typical insertion in vehicle i (self-pair or two-use reach), while
R2i is cross-pair ratings in vehicle / and j. The estimated regression equations of R2l and
Rtj for the South Korean target audience may be different from US equations. In 1992,
Korean Gallup reported the regression equations for the estimation of R2i (n = 528) and
R:J (n = 560,292) with South Korean target audience groups (see table 18).
Table 18. R2I and Rtj estimation equations for network television
US Equations (for overall audience)
Estimation
R2t = -0.0065 + 1.795/?,-0.331 (R,)2
R.t
Estimation
Rtl = -0.0015 + 1.0095 (R:+ RJ)-\.5RI If
Korean Equations (for 20 - 49 year old men)
R2,
Estimation
R2i= 0.000525+ 1.848325 7?, — 1.432(f?,)2
R,
Estimation
Rtj = 0.00005 + 1.003 (R, + R,)- 1.87755/?, R;
• American source: adaptedfrom Kent Lancaster (1993).
• South Korean source: adapted from Korean Gallup (1992).
Since the Korean Gallup reported the estimated regression equations with gender
and age variables, in this study, the equations for only the target audience group of
ASIANA campaign are estimated. Then, the media evaluation model will be modified
with the regression equations calculated from South Korean target audience which is 20 -

83
49 year old men. The estimated American and South Korean regression equations are
shown in Table 18
Error Estimation for the Media Evaluation Model
The BBD model has been widely accepted in both academia and industry because
of its simplicity, aggregating power, and good empirical fits. This approach has been
improved by the availability of household-level exposure records obtained from people-
meters and the increasing computation power of computers However, this approach is
also assumed to exhibit error in its estimations It is necessary to estimate how much error
the model is likely to produce.
Tests of the performance of media evaluation models have been undertaken in the
US These studies were conducted to compare the performance of the various media
evaluation models Recently, studies tried to compare the accuracy of media evaluation
models with different error definitions. Frequently used error definitions are average error
in reach (AER), average error in levels of exposure (ALE), and average percentage error
in exposure distribution (APE). The error terms are defined as the difference between
observed (tabulated) and estimated values in a schedule. For example, AER is the sum of
the absolute differences between observed and estimated reach in each schedule divided by
the total number of schedules. Sometimes, each error term produced different results that
made it difficult to decide which model’s performance is superior to another. However,
AER, APE, and ALE have been frequently used in these studies.
Leckenby et al. (1990) compared four different models’ performance in terms of
error estimation with about 500 magazine schedules tabulated from SMRB (Simmons

84
Market Research Bureau). This study showed that the BBD produces the smallest
percentage of error for reach estimation among the models used in the study. The study
also revealed that the larger the schedule size is, the more accurate the estimations made
with the models.
The beta binomial distribution (BBD) is an univariate probability distribution
model which treats all vehicles as one univariate and ‘composite’ vehicles and then
averages all vehicle insertions of a schedule. Therefore, the model often produces over¬
estimations in reach. Because of the problem with univariate models, the development of
multivariate models has been attempted. However, because of the excessive
computational difficulties involved in using all of the vehicle numbers together, complete
multivariate models have not been developed. Ideally, multivariate media evaluation
models should be developed because they may allow the most realistic view of complex
media evaluation processes without simplifying or averaging assumptions which are
currently used with univariate models (Leckenby et al., 1993).
An alternative is “sequential aggregation.” Although sequential aggregation is not
a true multivariate method, studies showed that the models using sequential aggregation
are better for estimating reach (»+) than those of univariate models. With sequential
aggregation, two-vehicle bivariate distributions are collapsed to form a new composite
vehicle which is then combined with a third vehicle and this procedure is continued until
all vehicles are aggregated into the distribution of the schedule. The BBMD model used
in this study uses sequential aggregation.
Leckenby et al. (1993) tried to test the performance of several media evaluation
models including model using sequential aggregation. In this study, about 480 magazine

85
schedules were tabulated from SMRB The average error in reach (AER) with BBD
model was about 6.4 percent, while the AER with the model using sequential aggregation
based on the BBD model [Morgenszten Sequential Aggregation Distribution model
(MS AD)] was 3.1 percent. The error estimation of the model used in this study (BBMD)
is similar to MSAD because both models are based on the BBD and use sequential
aggregation.
According to Leckenby et al. (1993), the model which uses sequential aggregation
is also the most accurate in predicting each level of the exposure distribution. The average
total error percentage in the exposure distribution (APE) was about 15 percent with
MSAD, while that of the BBD was 33 percent. Therefore, it can be assumed that about 3
percent error is likely to be made for reach estimation and about 15 percent error is
expected for the total exposure distribution for models which use sequential aggregation
based on the BBD. According to a selected example of observed and estimated exposure
distributions with their large schedule sample in the study, the error estimation for reach
(3+) was about 1.9 percent which is smaller than the error estimation of reach (1+) of 3
percent.
However, these error estimations also have limits. First, all of the studies
mentioned above used vehicle data instead of advertising message data. Since the model
used in this study will predict reach/frequency estimation with advertising message data,
the error estimation with message data would be smaller than that of vehicle data.
Another important point to be considered is that all the studies mentioned above used only
magazine advertisements. No study has been conducted for the error estimation of media
evaluation models for TV advertising. Since the ASIANA campaign mostly used

86
television advertising (about 93.5 percent of total campaign expenses), the error
estimation for television advertising is suggested for a future study. Additionally, this type
of study should be conducted with the actual BBMD model for better error estimation and
it is also recommended that the error estimation studies should be conducted with South
Korean data because the error estimation with US data may be different from South
Korean error
It also should be considered that all of the observed (tabulated) data of previous
studies are based on the sample estimates. Although the possible error can be very small
with proper sampling methods, all samples must have sampling errors. The studies
mentioned above seem to have ignore the possible sampling errors for observed
(tabulated) data to examine error estimation of media evaluation models. The application
of sampling errors should be considered for better error estimation of media evaluation
models.
The model used in this study will be examined to show how accurately the model
can predict the reach/frequency estimation. If a critical amount of error is involved with
the model estimation of reach (n+) in this study, the study results will not demonstrate
significant (accurate) predictions. Therefore, the error estimation may not be included to
test hypotheses in this study. Whatever amount of error occurs from the model used in
this study, the error should be a part of the model performance.
In addition to the error estimation of the model used in this study, additional errors
are likely to be produced because of the unknown factors in South Korea, such as carry¬
over rate and magazine ratings. However, the normative framework combined with a

87
reasonable media evaluation model in this study will suggest ways to predict and evaluate
advertising communication effects in South Korea.
Suggestions for Improving Campaign Efficiency and Budget Levels
In this study, the ASIANA advertising campaign will be critically analyzed to
identify potential for improving campaign efficiency and to suggest appropriate monthly
budget levels. The vehicles used in the campaign with relatively high inefficient CPM-
MSG1 figures will be replaced with those with more favorable CPM-MSG figures. The
estimated effective reach 3+ of the revised plan will then be compared to that of the
original plan. The gap, if there is any, of the estimated effective reach 3+ between
relatively high versus favorable CPM-MSG figures will show how the campaign efficiency
can be increased.
The model used in this study also will suggest optimum budget levels for the
monthly schedules. The possible vehicle insertions within each monthly schedule will be
optimized and the estimated message reach 3+ will be provided for various monthly
budget levels, such as $100,000, $200,000, $300,000, and $400,000. If there is an
approximate saturation point among the various budget levels, the amount would be the
optimum level of monthly spending in terms of message reach 3+. The optimum monthly
budget level will be suggested for similar ASIANA campaigns.
1 CPM-MSG: Cost to achieve eveiy thousand message impressions using a particular vehicle. It is a
useful measure of cost efficiency.

88
Sophisticated versus Naive Comparison
Armed with reach/frequency evaluation models, advertising communication effects
also can be estimated with a sophisticated approach or a number of typical nai ve
approaches. The simple typical procedures are naive at best and mostly misleading. One
of the most important issues related to the naive approach is time frame. Many advertising
media practitioners are likely to pool quarterly or yearly media options and develop global
reach and frequency estimates for media plans executed (Lancaster, Kreshel, and Harris,
1986) The main reason would be a lack of expertise or a simple desire to impress
advertisers with favorable numbers. If tracking studies show lower than expected results,
many often conclude that it must be a creative problem because media has delivered big
numbers. For an example of annual figures, it is assumed that consumers exposed to an
advertisement on January are likely remember it on December. If 40 percent of target
audience reach 3+ is expected with a yearly figure, the tracking study results will show a
significantly lower percentage. This is mainly due to consumer forgetting and competitive
advertising which are constantly conspiring to diminish advertising effects.
The use of vehicle ratings, instead of message ratings, also can cause considerable
problems for accurate estimation of advertising communication effects because those who
are exposed to an advertising vehicle may not be exposed to the actual message.
Therefore, in this study, a sophisticated approach will be employed with a monthly time
frame, carry-over rate, magazine message/vehicle ratio, and message data for television
advertisements. The estimated results using a sophisticated approach will be compared
with those with the naive procedures.

89
Summary of the Normative Framework
Table 19 summarizes the items which should be gathered or defined to complete
the normative evaluation of this case study.
Table 19. Summary of the normative framework applied to the ASIANA case study
Items
ASIANA Case Study
Target Audience
Definition
20-49 year old men who live in Seoul, Korea (2,884,983).
Vehicle Data
Ratings of 338 TV Programs and 14 magazines, used for the
advertising campaign, gathered from MSK and Lee’s PR
Message Data
TV advertising message ratings are gathered from MSK. For
magazine advertising, the message/vehicle ratio is estimated
based on US averages available from Roper Starch Worldwide.
Advertising
Expenditures
Total of (US) $1,231,474 for three month advertising campaign.
Media Evaluation
Model
A variation of beta binomial exposure distribution model,
ADplus, is used. The important technical procedures are
described in Appendix C.
Narrow Time
Frame
Within the three-month campaign period, each monthly
campaign schedule will be analyzed separately and assigned to a
flowchart with an appropriate month-to-month advertising carry¬
over rate.
Tracking Study
Results
Since the campaign was new, only a post-test was used to
measure consumer awareness of the advertising message,
preference, and willingness to use.
Media Plan Used
The ASIANA Advertising media plan is obtained from the
advertising agency, SangAm.

90
Research Hypotheses
Using data available before the campaign was launched, a variation of the BBD
model (ADplus) will be used to forecast the impact of the South Korean advertising plan.
Then, the predicted advertising campaign communication effects, such as advertising
message awareness, preference, and willingness to use will be compared with the actual
target consumer responses to the advertising campaign which were obtained from tracking
studies conducted by SangAm. Since there are only two South Korean air carriers, most
South Korean people are already aware of the two brands. Additionally, the purpose of
this study is to predict and measure the effects of the new campaign, “Her name is
ASIANA.” Therefore, in this study, advertising message awareness is used to measure
advertising campaign communication effects instead of brand awareness.
As explained in Chapter II, the optimal exposure frequency level for each
communication effect is estimated In this study, the base (3+) is used for target audience
awareness of the ASIANA campaign message because the survey of South Korean media
practitioners reported in Chapter III and previous studies support the use of 3+ as a base
for advertising message awareness. The marketing, copy, and media factors are also
considered to decide the effective frequency level for the message awareness based on the
suggestions of Ostrow (1982), and Sissors and Bumba (1996) explained in Tables 4 and 5.
Based on the estimated message effective frequency for the ASIANA campaign message
awareness, which is explained in Table 20, the first research hypothesis of this study is as
follows:

91
Hypothesis 1. The estimated message effective reach of the ASIANA advertising
campaign, based on the estimated message effective frequency of 3+, is not
different from the actual target audience awareness of the message.
Table 20. Suggested points and modified effective frequency levels for ASIANA Airline
campaign effects
Suggested Points for Marketing, Copy, and Media Factors
for ASIANA Airline Campaign
Marketing
Single kind of image
-.2
Established brand
-.2
New message
-.2
Low market share
+.2
Smaller and less well-known
+.1
Media
-.1
Low brand loyalty
+.2
Low advertising clutter
-.2
Product used occasionally
+.2
Continuous advertising
-.1
Highly competitive market
+.2
A few advertising media
High opportunities for media repetition
-.2
Copy
Simple copy
-.1
Additional Items
- 1
More unique copy than
Good advertiser image
+ 2
competition
-.1
Total service quality (relatively poor)
- 1
New copy campaign
+.2
Advertising message quality (high)
+.2
Image type copy
+.2
More competitor promotional activities
Large advertising units
-.2
Sum of Points =(-0.1)
Communication Ooal
Effective Modified Effective
Frequency Base Frequency Level
Awareness of ASIANA Campaign
3+ 3.0-0.1 =2.9 or 3+
Preference of ASIANA Airline
b+ (b > 3) or b - 0.1 = at least 3.9 or
at least 4+ 4+
Willingness to Use ASIANA Airline
c+ (c > 4) or
at least 5+
c - 0.1 = at least 4.9 or
5+

92
The effective frequencies for target audience preference of ASIANA airline and
willingness to use also can be estimated with the same procedures However, it may not
be possible to forecast target audience preference and willingness to use because of many
confounding variables. For example, the predicted consumer preference and willingness to
use can not be explained sufficiently with the advertising campaign effects because of the
preconception that existed before the campaign began. Furthermore, it is assumed that
more marketing effects, such as service quality, are involved with the target audience
preference and willingness to use. Therefore, in this study, only the minimum levels of
effective frequency are estimated for preference and willingness to use. It is assumed that
the actual effective frequencies for target audience preference and willingness to use are
higher or at least the same as the estimated minimum effective frequency levels.
As shown in Figure 4, as the effective frequency level increases, the estimated
message reach decreases because the estimated message reach «+ represents the percent
of target audience exposed to the advertising campaign message at least n times. For the
example frequency distributions in Figure 4, the percent (17.4%) of the target audience
exposed to the advertising message at least once is higher than the percent (1.5%) of the
target audience exposed to the message at least twice. Therefore, the estimated message
reach based on the minimum effective frequency levels for preference and willingness to
use will likely show higher percentages of the target audience exposed to the campaign
message than actual target audience message reach needed to achieve preference and
willingness to use.

93
With the procedures explained in Chapter II, Table 20 also shows the minimum
frequency level of each communication effect. Consequently, the second and third
research hypotheses are as follows.
Hypothesis 2. The estimated message reach, based on effective frequency for
predicted target audience preference of ASIANA Airline of at least 4+, will be
higher than or equal to the actual target audience preference of ASIANA Airline.
Hypothesis 3. The estimated message reach, based on effective frequency for
predicted target audience willingness to use ASIANA Airline of at least 5+, will be
higher than or equal to the actual target audience willingness to use ASIANA
Airline.
Testing Procedures
After the essential data are provided by MSK and SangAm, the effective frequency
levels for three communication effects are determined with the consideration of marketing,
copy, and media factors. Then, ADplus will be used to estimate the probability of target
audience members being exposed to advertising messages in the schedule any given
number of times within a narrow time frame, such as monthly.
The ADplus program provides exposure or frequency distributions based on the
BBD model which was shown to be an accurate estimation tool compared to tabulated
exposure distributions (Kreshel, Lancaster, & Toomey, 1985). When using ADplus, the
separate media category evaluations using message scores are weighted to generate

94
frequency distributions for both television and magazine categories combined. The
weights make the separate media categories comparable for their ability to achieve
common communication goals (e g., advertising message awareness) (Lancaster &
Helander, 1987).
The only inputs that may be different from the ADplus defaults would be the
message/vehicle ratios for television and magazine advertisements. Since there are
fundamental differences between Korean and American advertising media environments,
the Korean message/vehicle ratio would be different from that of the United States In
this study, actual television advertising message ratings were collected from South Korea.
Consequently, the message/vehicle ratio for television advertising is 100 percent, because
actual message ratings will be used as input data. However, for magazine advertising,
since the advertising message ratings are not available in South Korea, the average
message/vehicle ratio in the United States, according to Roper Starch Worldwide, for one-
page full color magazine advertisements, 45 percent, will be used.
The campaign effects also will be predicted by the program over a narrow time
frame, such as monthly Each monthly advertising media schedule will be assigned to a
flowchart and then an appropriate carry-over rate will be applied to estimate the three-
month advertising campaign effects. Since there is little information about advertising
carry-over rates in South Korea, a rate for an airline service is estimated by the author
with previous American and South Korean studies and will be applied to the flowchart
According to Lancaster (1993), a typical monthly advertising carry-over or
retention rate might be somewhere around 18 percent, which means that just under one
fifth of previous monthly advertising influence continues to exist among target audiences

95
in subsequent months. Recently, Leone (1995) summarized 10 previous studies of
econometric sales-advertising analysis and concluded that the mean monthly coefficient of
the lagged dependent variable is 0.44. The studies reviewed by Leone were Bass and
Clarke (1972), Montogomery and Silk (1972), Palda (1964), and seven studies of Samuels
(1970 - 1971). However, since these studies were conducted about 25 years ago, current
average carry-over coefficients might be different from 0.44. For example, according to
AD $ Summary (1971 and 1996), the number of advertising media in 1995 has almost
doubled from that of 1970. Therefore, since there was not as much advertising clutter in
previous decades as exists today, the coefficient of 0.44 is probably too high in today’s
market. Furthermore, according to the Statistical Abstract of the United States (1971 and
1996), the per capita advertising in 1995 ($690) has increased about seven times from that
of 1970 ($97) in the United States. The Statistical Abstract of the United States (1996)
also shows that average consumer prices have increased about 3.5 times from those of
1970. Therefore, accounting for inflation, per capita advertising has almost doubled in
real terms (7 + 3.5 = 2). The real difference in per capita advertising between 1970 and
1995 suggests that average monthly carry-over rate also must be divided by 2 (0.44 / 2 =
0.22). This is based on the assumption that consumers now have twice as many
advertisements to contend with as they did in the 1960s and 1970s when the monthly
carry-over studies were conducted It is unlikely that consumer memory has kept pace.
The carry-over rate will also vary by the size of the advertising exposures, creative
factors, and other marketing situations Products which have short purchase cycles may
have higher carry-over rates than those of long purchase cycles. The frequently used
products, such as household cleanser and toilet soap, also may have higher carry-over

96
rates than products or services that are not frequently used. Since the products used in the
10 previous studies were toilet soap, household cleanser, washing up liquid, and scouring
powder, it is assumed that the carry-over rate for an airline service is at least 20 percent
less than those frequently used products. Therefore, another reduction of 20 percent from
the rate of 0.22 results in an 18 percent of carry-over rate [ 0.22 - (0.2 x 0.22) = 0.176 ].
On the other hand, there are a few studies to estimate carry-over rates in the South
Korean market. Ryu (1993) analyzed foreign studies of advertising carry-over rates and
concluded that, generally, monthly advertising carry-over rates are between 0.30 and 0.64.
He also suggested that carry-over rates vary by product or service categories and the
carry-over rates in the South Korean market may be similar to the results of foreign
studies. Kang (1995) examined the carry-over rates for confectionery brands in South
Korea and concluded that the mean monthly carry-over rates are 0.393 for the brands less
than 2 years old, 0.618 for those between 2 and 3 years old, and 0.805 for brands more
than 3 years old. Though the ASIANA Airline service is more than 3 years old, the rate of
0.393 may be appropriate for the present study because the advertising campaign was new
and this study is to predict the new campaign communication effects. As explained earlier,
since the brands used in Kang’s study were also frequently purchased confectionery
brands, such as candies, chewing gums, and chocolates, the carry-over rate of 0.393 might
be too high for an airline service. Therefore, a 20 percent reduction from the rate of
confectionery brands was estimated by the author, resulting in a 32 percent carry-over rate
[ 0.393 - (0.2 x 0.393) = 0.3144 ].
Both 18 and 32 percent carry-over rates seem to be reasonable for this study.
However, since this study is to estimate advertising campaign effects in the South Korean

97
market, the rate calculated from the Korean studies will be initially applied in this study.
The rate of 18 percent also may be used as a calibrating option which will be explained
later in this Chapter. Consequently, it is assumed that 18 to 32 percent of those who are
exposed to the ASIANA advertising campaign message can recall it after a month without
further exposures in the South Korean market Both 18 and 32 percent carry-over rates
also indicate a moderate 90 percent duration interval for advertising effects, such as 2.5
and 3.2 months2 Practically speaking, it is also reasonable that, without aid, people can
not remember an advertising campaign message to which they were exposed more than
three months ago
Finally, the predicted, weighted, combined frequency distribution will be compared
with the consumer tracking study results. At this step, the estimated percent of the target
audience exposed to the advertising campaign message at the pre-determined frequency
level such as 3+ will be matched to the consumer tracking study percent of the target
audience who exhibit measured communication levels. For example, the estimated percent
of the target audience exposed three or more times to the campaign message may match
the percent that can correctly recall the salient ASIANA advertising message content.
Statistical Procedures
The predicted advertising communication effects will be statistically tested with the
tracking data. For the predicted percent of the target audience awareness of the
2 The 90 percent duration interval for advertising effects is calculated as follows:
Assuming that effective reach n+ is 0.5 and the carry-over rate is 0.32, the second month loss is 0.34 [ 0.5
- (0.5 x 0.32) = 0.34 ], third month loss is 0.4488 [ 0.5 - (0.5 x 0.32 x 0.32) = 0.4488 ], and fourth
month loss is 0.4836 [ 0.5 - (0.5 x 0.32 x 0.32 x 0.32) = 0.4836 ]. Therefore, 90 percent of the
advertising effects (0.9 x 0.5 = 0.45) will be lost after about 3 .2 months.

98
advertising message, a large-sample confidence interval for sample probability will be used
(Agresti and Finlay, 1986). The formula is as follows:
C ./. = ii ±
Where:
C. I = Confidence interval,
71 = Percent of the sample indicating awareness of the message,
-a/2 = Two tailed confidence coefficient for 95% confidence
level,
n = Sample size of 200 males 20 - 40 years of age who live in
Seoul, Korea.
Source: adaptedfrom Agresti and Finlay (1986).
If the message effective reach 3+ is within the confidence interval calculated with 95
percent confidence level, it can be said that the predicted awareness of the advertising
message is not different from the actual consumer awareness of the campaign message.
Since only the minimum levels of effective frequency are used to estimate
consumer preference and willingness to use, a Z-test will be used to determine whether the
actual target audience preference and willingness to use are at least the same or less than
the predicted percents (Agresti and Finlay, 1986). The formula is as follows:

99
/V
Where:
Z = Test statistic,
A
ft = Percent of the sample indicating preference or willingness to use,
ft$ = Minimum predicted percent of consumer preference or
willingness to use,
CJ
ft
Standard error of sample consumer preference or
willingness to use, where a. = )x0(\ - tt0) / n .
Source: adapted from Agresti and Finlay (1986).
If the calculated Z-scores are big enough to accept the research hypotheses ( Ha :
n > n 0 ) at 90 - 95 percent confidence levels, the second and third research hypotheses
are statistically supported. In other words, the framework will support the second and
third hypotheses if the estimated reach 4+ and 5+ respectively are higher or equal to the
actual consumer preference and willingness to use.
Calibrating Forecasting Procedures
Since this study lacks some necessary information, such as the message/vehicle
ratio for magazine advertising and advertising carry-over rate for an airline service in
South Korea, the normative framework may show a gap between what is predicted and
what is found in the tracking study. In this case, the effective frequency levels (n+) used,

100
message/vehicle ratio for magazine advertising, and the monthly carry-over rate can be
adjusted to obtain a closer match between the ADplus prediction and tracking study.
Ideally, the message/vehicle ratio and monthly carry-over rate for an airline service
can be acquired from another consumer tracking study in South Korea. However, since
no such information is available in South Korea, backward adjustment can solve the
problem for better prediction in the future. For example, if the predicted consumer
awareness of the ASIANA advertising campaign message is 10 percent more than the
tracking data, the monthly carry-over rate of 32 percent, which is used initially in this
study, might be too high, or the message/vehicle ratio of 45 percent for magazine
advertisement might also be too high. Adjustment also can be made for the effective
frequency level of 3+. Perhaps, the adjustment of effective frequency level (n+) has the
greatest effect, while the message/vehicle ratio and carry-over rate have second order and
least effects.
For the example ADplus results in Figure 4, which is the monthly magazine
advertising schedule for the ASIANA campaign, if the message/vehicle ratio is decreased
to 35 percent from 45 percent, the estimated target audience message reach 2+ is
decreased from 1.5 percent to 0.9 percent. When the carry-over rate is adjusted from 32
percent to 18 percent, the estimated message reach 3+ on April is decreased from 0.4
percent to 0.2 percent. On the other hand, the estimated message reach 3+ is increased
from 0.4 percent to 0.6 percent when the carry-over rate is changed from 32 percent to 46
percent.
The adjustment of estimated effective frequency (n+) produces a more drastic
change in estimated reach than the adjustment of carry-over rate or message/vehicle ratio.

101
For the example ADplus results in Figure 4, a one unit change in message frequency level
from 2+ to 1+ shows a dramatic increase in estimated message reach from 1.5 percent to
17.4 percent.
In addition to these calibrating procedures, there are more options to be
considered. First, in this study, magazine vehicle ratings are gathered from each magazine
publication company because no independent information is available in South Korea.
These ratings may be higher than actual ratings because of the intent of publishers in
selling their publications to advertisers. However, since the magazine advertising is only a
small part of the ASIANA campaign (6.5 percent of total media expenses), the possible
errors caused by these ratings may not be critical. Second, in spite of the random
sampling and the use of neutral interviewers, who are not related to ASIANA Airline,
there is a possibility that the data also might be gathered in favor of ASIANA Airline
because the interviewers were hired by the advertising agency of ASIANA Airline.
However, generally, the tracking study was well designed and executed without any
critical problems. The only confounding variable concerned is the ‘instrumentation’ which
may be possible when the measuring instrument is a human observer. For better
prediction in the future, magazine ratings should be gathered from an objective media
research institute. It is also recommended that the interviewers should be trained to
exclude a possible confounding variable such as ‘instrumentation.’
With the calibrating options explained above, the possible options in this study
may be the adjustment of effective frequency n+, carry-over rate or message/vehicle ratio
or some combination of these options. It also may be appropriate to lower the
message/vehicle ratio for magazine advertisements because the magazine ratings are

102
probably biased upward. Although, generally, the adjustment of carry-over rate has the
least effects, it may be an important adjustment in this study because it is unknown in the
South Korean market.
These calibrations of forecasting procedures ultimately lead media planners to
better predictions for future advertising campaigns. Table 21 shows the possible
adjustment options and combinations, expected effects, and approximate decisions in this
study.
Table 21. Adjustment options, expected effects, and decisions in this study
Options
Expected Effects
Possible Combinations
1. Effective frequency level
Most effects
1, 2, or 3 alone.
2. Message/vehicle ratio
Second most effects
Combine 1 & 2, 1 & 3, and 2 & 3.
3. Carry-over rate
Least effects
Combine 1, 2, & 3.
Likely Decisions in this Study
1. The calibration of effective frequency level (n+) may not be appropriate because
strong effects are expected. In this case, message/vehicle ratio and carry-over rate will
be considered for calibrating adjustment procedures. Message/vehicle ratio alone,
carry-over rate alone, or balance of these two adjustments may be more appropriate.
2. The message/vehicle ratio for magazine advertisements may be lowered because the
magazine ratings might be biased upward.
3 The adjustment of carry-over rate may be most interesting to calibrate the procedures
in this study because it is largely unknown in South Korea.
Once appropriate frequency levels, message/vehicle ratios, and carry-over rates are
recommended, they can be used as benchmarks to evaluate the likely effective frequency
and advertising effects for future schedules because they can forecast consumer tracking

103
study results. It is also possible for advertising planners to estimate what is expected from
their advertising budget and the gaps between the budget and likely communication effects
for future advertising campaigns (Lancaster & Helander, 1987).
In Chapter V, the predicted campaign effects will be explained and tested in
comparison with the consumer tracking data. These results will suggest what should be
considered to improve effective reach and advertising communication effects for ASIANA
campaigns in the future. Possible improvements in the normative framework, theory,
methods and procedures also will be implied by the results of the case study.

CHAPTER V
RESULTS
Introduction
The main purpose of this study is to demonstrate how to predict advertising
communication effects using a variation of beta binomial exposure distribution model
Previous studies and literature on forecasting or measuring advertising communication
effects were reviewed in Chapter II. The test market in this study is South Korea because
of the lack of systematic study of the relation between media planning and advertising
effects in that country. A South Korean advertising campaign, “Her name is AS1ANA”
was selected as a case study It was a new three-month campaign and the primary target
audience for the campaign was 20 - 49 years old Korean men who live in Seoul.
Survey research was conducted prior to the model application and the results were
provided in Chapter III. The results showed South Korean media practitioners’ awareness
of current states of reach/frequency evaluation methods and what kind of communication
effects are used to measure or evaluate advertising campaign effects in South Korea.
A normative framework necessary to predict advertising communication effects
also was developed in this study and applied to the AS1ANA case study. The items
required to implement the normative framework are target audience definition, vehicle
104

105
data, message data, advertising expenditures, media evaluation model, narrow time frame,
tracking study results, and the media plan details. The summary of the normative
framework applied to the ASIANA case study is shown in Table 19 of Chapter IV.
Based on the survey results, previous studies plus marketing, copy, and media
factors, in this study, an effective exposure frequency level was estimated to predict target
audience awareness of a new ASIANA campaign “Her name is ASIANA.” A hypothesis
test will show how accurately the model used in this study can predict the target audience
awareness of the new campaign message in comparison with the tracking study results.
In addition to the campaign message awareness, this study also examines target
audience preference and willingness to use. However, due to many confounding variables
contributing to preference and willingness to use, such as target audience preconception of
ASIANA Airline, only minimum levels of effective exposure frequency were estimated
with the same procedure described above. Then, the minimum effective frequency levels
will be analyzed in comparison with the actual tracking study results. The hypotheses will
be tested based on the normative framework and the methods described in the previous
chapters
A calibration procedure will then be used, if necessary, to improve the model
predictions. The test market results will be compared with alternative media evaluation
model predictions of desired communication effects.
The ASIANA campaign also will be critically analyzed and suggestions will be
provided to increase effectiveness and efficiency for future ASIANA campaigns. Finally,
typical naive estimations with vehicle data or with long time frames, such as quarterly, will

106
be compared with the results of the sophisticated method employed in this study. The
comparison will show how the potential difference is large and misleading.
Test Market Results
For the predictions of the media evaluation model used in this study, target
audience vehicle and message ratings for television advertisements were collected from the
MSK (Media Service Korea). Magazine vehicle ratings were obtained from each
magazine publication company and the average US message/vehicle ratio of 45 percent
was initially applied to estimate target audience magazine message ratings. The
advertising expenditure and media plan used for the campaign were also gathered from the
advertising agency of ASIANA Airline, SangAm.
Within the three-month campaign period, each monthly campaign schedule was
analyzed separately and assigned to a flowchart with an appropriate month-to-month
advertising carry-over rate. The monthly carry-over rates were estimated from both
American and Korean studies. The estimated monthly carry-over rate from a Korean
study was 32 percent, while it was 18 percent for various US studies. Along with the
message/vehicle ratio for magazine advertisements, the carry-over rate will be used as
subsequent calibrating options because they are largely unknown in South Korea.
Media Evaluation Model Predictions
As explained in Chapter IV, the media evaluation model initially uses a 45 percent
message/vehicle ratio for magazine advertisements and a 32 percent monthly carry-over

107
rate. Initially, the media evaluation model used in this study predicted that about 60
percent of the target audience is likely to be exposed to the ASIANA campaign message
at least three times with the given television and magazine schedules combined. The
model also predicted that about 91 percent of the target audience is likely to be exposed to
the campaign message at least once. It was estimated that consumer target members must
be exposed to the new ASIANA campaign message at least three times before they are
aware of the new campaign message. Therefore, the model prediction indicates that the
given television and magazine advertising schedules combined would lead 60.4 percent of
the target audience to be aware of the new campaign message.
Figure 5 shows the ADplus predictions with the flowchart of the three-month
ASIANA advertising campaign for reach 1+, 3+, 4+, and 5+. In Figure 5, the column
headed “FEB” stands for the first monthly campaign, February 17 to March 16. “MAR”
represents the second monthly period from March 17 to April 16, while “APR” represents
the third month from April 17 to May 16. Figure 5 also shows that there were 338
insertions for television advertisements and 42 insertions for magazine advertisements for
the entire three-month campaign. The total cost for television advertising was
$1,150,900, while it was $80,600 for magazine advertising. Monthly expenditures and
insertions are also shown in Figure 5.
Reach 3+ indicates that the estimated percentage of the target audience exposed to
the campaign message at least three times is 60.4 percent, while it is 45.8 percent for 4+
and 34.7 percent for 5+. The Gross Rating Points (GRPs) row summarizes the monthly
sum of all message rating points delivered by the monthly campaign schedule. The
Frequency row shows the mean number of times target audience members expected to be

108
reached by the monthly campaign message are exposed to it. It is 3.9 for the first month,
5.2 for second month, and 4.5 for the third month. For example, the 95.6 percent of the
target audience reached by the advertising message in February are exposed, on average.
3.9 times.
ADplus (TM) FLOWCHART for AS I ANA Campaign
FEB. 17 THRU MAY 16 1996
Target audience: 2,884,983 Korean men 20-49 years old who live in Seoul
Prepared by: Hyunsoo Park, University of Florida
Media
Ads
Cost (000)
FEB
MAR
APR
NETWORK TV
338
1,150.9
125
128
85
MAGAZINES
42
80.6
14
14
14
TOTALS
380
1,231.5
139
142
99
COVERAGE
FEB
MAR
APR
Message:
Reach 1+
95.6
96.8
Reach 3+
68.7
77.8
/60.4
Reach 4+
49.7
63.8
( 45.8
Reach 5+
32.4
49.9
V 34.7
GRPs
372.3
501.2
411.8
Frequency
3.9
5.2
4.5
MONTHLY COST (000s)
NETWORK TV
470.7
398.3
282.0
MAGAZINES
26.9
26.9
26.9
TOTALS
497.5
425.1
308.8
The advertising carry-over fixed function rate is 32.0%.
The message/vehicle ratio for magazine advertisements is 45%.
Figure 5. Media evaluation model prediction
Detailed ADplus results are shown in Appendices E - L. Appendices E - G include
the monthly television schedule for each of the three months in the campaign, while

109
Appendix H shows the monthly magazine schedule, which is constant during the three
months of the campaign. Appendices I - K show each monthly television and magazine
advertising schedule combined. Finally, Appendix L shows the flowchart of the three-
month campaign with a 45 percent message/vehicle ratio for magazine advertisements and
a 32 percent monthly carry-over rate. All vehicle and message information is also shown
in Appendices E - L.
For the communication effects of preference and willingness to use, it was
estimated that the minimum effective frequency levels are 4+ and 5+. Figure 5 also shows
the model predictions with these effective frequency levels indicating that 45.8 percent and
34.7 percent of target audience members may be exposed to the ASIANA campaign at
least four and five times respectively. However, as explained in Chapter IV, these
percentages are estimated with the minimum effective frequencies. Therefore, the actual
target audience exposure frequencies needed to achieve these communication effects may
be higher than the minimum effective frequencies. Consequently, it is assumed that the
estimated reach 4+ and 5+ are higher than, or equal to, the actual target audience
advertising exposures needed to achieve the desired communication effects, such as
preference and willingness to use.
Tracking Study Results
The tracking study conducted by the marketing department of the advertising
agency of ASIANA Airline shows that 54.5 percent of the consumer target audience was
aware of the new ASIANA campaign message after the three-month advertising
campaign. For preference and willingness to use, the percentages are 34.5 and 29.5

110
respectively. A majority (66.1%) of the target audience members responded that the
advertising message quality of the new ASIANA campaign was superior and also
answered that they understood the campaign as a presentation of ASIANA comfort and
good service. Table 22 shows the tracking study results for the three communication
effects in comparison with the media evaluation model predictions.
Table 22. Consumer tracking study results and model predictions
Communication Effects
Model
Predictions
Tracking
Study Results
Difference
(Predicted -
Actual)
Awareness of the new
campaign message
60.4%
54.5%
5.9%
Preference
45.8%*
34.5%
11.3%
Willingness to use
34.7%*
29.5%
5.2%
* Indicates that the prediction uses only the minimum level of effective frequency.
However, the model predictions for preference and willingness use employ only
minimum levels of effective frequency (»+). The actual target audience preference and
willingness to use are expected to be lower than the predictions. Therefore, it may not be
appropriate to directly compare tracking study results with model predictions for
preference and willingness to use. If the tracking study results are higher than or at least
equal to the model predictions, this study can support the second and third research
hypotheses.

Ill
Testing Hypotheses
Awareness of the Campaign Message
The first hypothesis in this study is that the estimated message effective reach of
the AS1ANA advertising campaign, based on the estimated message effective frequency of
3+, is not different from the actual target audience awareness of the message. As shown
in Table 22, the predicted percent of target audience awareness of the new campaign
message is higher than the tracking study results. The difference is about 5.9 percent. A
large-sample confidence interval for sample probability shows that the confidence interval
for consumer awareness of the campaign message at the 95 percent confidence level is
between 47.6 percent and 61.4 percent'. The predicted message awareness of 60.4
percent is within the confidence interval. Therefore, the statistical test supports the
proposition that the predicted consumer awareness of the campaign message is not
significantly different from the actual consumer awareness of the campaign message.
Although this difference is not statistically significant, procedures will be described
later in this chapter to narrow the gap between predicted and actual target audience
awareness.
Preference
The second research hypothesis in this study is that the estimated message reach,
based on effective frequency for predicted target audience preference of AS1ANA Airline
of at least 4+, will be higher than or equal to the actual target audience preference of

112
ASIANA Airline. The survey of South Korean media practitioners reported in Chapter
III and previous studies support the use of an effective frequency of 3+ for advertising
message awareness. Therefore, needed exposure to achieve preference would be higher
than 3+ or at least 4+. Because of the difficulty in accurately measuring target audience
preference as a function of the campaign effects, only a minimum level of effective
frequency of 4+ was used to estimate preference.
The media evaluation model prediction with effective frequency of 4+ is 45.8
percent and the corresponding tracking study results for target audience preference is 34.5
percent. The Z test shows that the model prediction is significantly higher than the
tracking study results ( Z = 3.41, p< 0.01). It is assumed that, generally, more advertising
exposures are needed to achieve target audience preference than to achieve awareness of
the advertising message. Needed target audience exposure to achieve preference is higher
than four for the ASIANA Airline campaign. Apparently, more frequency is needed to
achieve preference than the minimum level, which was originally supposed. However, it is
difficult to accurately estimate needed target audience exposures to an advertising
message to achieve preference because of many confounding variables, such as other
marketing efforts and consumer preconceptions of ASIANA Airline.
Willingness to Use
The third research hypothesis is that the estimated message reach, based on
effective frequency of at least 5+ is required to achieve target audience willingness to use
ASIANA Airline, and that it will be higher than, or equal to, the actual target audience
willingness to use ASIANA Airline.

113
The media evaluation model predictions with an effective frequency of 5+ is 34.7
percent and the tracking study results show that 29.5 percent of the target audience has
willingness to use ASIANA Airline. The Z test shows that 34.7 percent is not significantly
higher than 29.5 percent (Z = 1.61, p>0.1). However, since the tracking study result is
not higher than the model prediction, the third research hypothesis also can be statistically
supported.
However, as mentioned earlier, needed target audience advertising exposure to
achieve willingness to use is also difficult to estimate because it may be the result of the
overall marketing effort of ASIANA Airline since 1989. As sales are difficult to measure
in terms of advertising effects, estimates of advertising’s contribution to willingness to use
also should hold constant many other influential factors.
It was assumed that more consumer exposures are needed to achieve willingness
to use than to achieve preference or awareness. This hypothesis test shows that the
needed target audience advertising exposure for willingness to use is higher than, or at
least equal to, the needed exposures for preference. The statistical test also provides
evidence that the needed exposure level for willingness to use is significantly higher than
that for advertising message awareness (p<0.05).
Calibrating Procedures
As explained in Chapter IV, this study lacks some necessary information, such as
accurate magazine ratings, a message/vehicle ratio for magazine advertising, and the
advertising carry-over rate for an airline service in South Korea. The gap between what is
predicted and what is found in the tracking study may be influenced by the lack of this

114
information. Both the magazine message/vehicle ratio and the carry-over rate, which were
initially used in this study, may be higher than actual values in South Korea. Therefore,
backward adjustments can be used for better prediction.
Three calibrating options explained in Chapter TV are the effective frequency levels
used («+), the message/vehicle ratio for magazine advertising, and the monthly carry-over
rate. To obtain a closer match between the model prediction and tracking study results,
adjustments of message/vehicle ratio for magazine advertisements and of the carry-over
rate may be appropriate because they are largely unknown in South Korea.
First, it may be appropriate to lower the message/vehicle ratio for magazine
advertisements. By lowering the message/vehicle ratio for magazine advertisements, the
message ratings for magazine advertisements will be lowered and likely will be closer to
the actual ratings because it is assumed that the vehicle ratings are biased upward.
Second, the carry-over rate of 32 percent may be too high for an airline service.
Since the carry-over rate estimation with US studies produced a monthly rate of 18
percent, the rates between 18 and 32 percent would be appropriate for better prediction.
Finally, the combination of these two options also can be applied. Table 23 shows
possible options and the resulting changes in predictions.
The adjustments of the magazine message/vehicle ratio and the carry-over rate
seem to be appropriate for better prediction in this study. It is estimated that the
reasonable message/vehicle ratio for magazine advertisements in South Korea is about 35
percent. It is approximately a 20 percent decrease from that initially used based on the US
average of 45 percent. A 20 percent decrease in the magazine message/vehicle ratio
would be appropriate because that is probably the extent to which magazine ratings are

115
biased upward. Otherwise, the message quality is high and probably effective because of
the attractive female models in the advertisements targeted to adult men 20 - 49 years of
age. Furthermore, in spite of the drastic decrease in the magazine message/vehicle ratio
from 45 percent to 5 percent in Table 23, which is unreasonably low, the predicted reach
3+ for the entire campaign is still higher than the tracking study results. Therefore, the
magazine message/vehicle ratio can not be used alone to improve the results. The rest of
the calibration should be explained by the carry-over rate. The combination of these two
options would be the most reasonable approach to calibration in this study.
As shown in Table 23, with a 21.7 percent monthly carry-over rate, the prediction
can be exactly matched with the tracking study results without adjusting the magazine
message/vehicle ratio from the initial use of 45 percent. Consequently, there are two
many degrees of freedom for the calibrating options. However, since the magazine ratings
are likely to be biased upward, the magazine message/vehicle ratio also should be used as
a calibrating option. As explained earlier, a 20 percent decrease from the initial use of 45
percent results in a 35 percent message/vehicle ratio, which seems to be reasonable. The
best prediction using the combination of these two options is produced with a 23.4
percent carry-over rate when a 35 percent magazine message/vehicle ratio is used. With
these reasonable adjustments, the media evaluation model prediction is exactly matched
with the tracking study results for advertising message awareness.

116
Table 23. Adjustment options and changes in predictions
Option 1. Message/vehicle ratio for magazine advertisements only.
Alternate Magazine
Message/Vehicle Ratios
Overall Campaign Predictions (%)
(Reach 3+) (Reach 4+) (Reach 5+)
45%
59.8
45.3
34.2
40
59.4
45.0
33.9
35
59.1
44.6
33.6
30
58.7
44.3
33.3
25
58.3
43.9
33.0
20
57.9
43.6
32.7
15
57.5
43.2
32.4
10
57.1
42.8
31.8
5
56.7
42.5
31.8
Option 2. Carry-over rate only.
Alternate (
Rates
"any-over
Overall Campaign Predictions (%)
(Reach 3+) (Reach 4+) (Reach 5+)
32.0%
60.4
45.8
34.7
28.0
57.7
42.7
31.3
24.0
55.6
40.1
28.5
21.7
54.5
38.6
27.0
20.0
53.6
37.5
25.8
16.0
51.6
35.0
23.2
Option 3.
Combination of message/vehicle ratio and carry-over rate.
Joint Alternatives
Overall Campaign Predictions (%)
Message/
Carry-
Vehicle
over
(Reach 3+)
(Reach 4+)
(Reach 5+)
Ratio
Rate
35.0%
32.0%
59.1
44.6
33.6
35.0
28.0
56.9
41.9
30.7
35.0
24.0
54.8
39.3
28.0
35.0
23.4
54.5
38.9
27.5
35.0
20.0
52.7
36.7
25.2
35.0
16.0
50.7
34.2
22.6

117
The media evaluation model predictions with calibrating options are shown in
Figure 6. From the predictions with calibration, it is estimated that 91.8 percent of the
target audience is likely to be exposed to the campaign message at least once and 54.5
percent are likely to be exposed to the campaign message at least three times. The
calibrating procedures would lead ADplus to predict results exactly matching the tracking
study outcome for target audience awareness of the new campaign message.
ADplus (TM) FLOWCHART for ASIANA Campaign
FEB. 17 THRU MAY 16 1996
Target audience: 2,884,983 Korean men 20-49 years old who live in Seoul
Prepared by: Hyunsoo Park, University of Florida
Media Ads Cost(000) FEB MAR APR
NETWORK TV 338 1,150.9 125 128 85
MAGAZINES 42 80.6 14 14 14
TOTALS 380 1,231.5 139 142 99
COVERAGE FEB MAR APR
Message:
Reach 1+
Reach 3+
Reach 4+
Reach 5+
95.2
67.5
48.3
31.2
GRPs
Frequency
364.9 460.1 351.6
3.8 4.8 3.9
MONTHLY COST (000s)
NETWORK TV
MAGAZINES
470.7 398.3 282.0
26.9 26.9 26.9
TOTALS
497.5 425.1 308.8
The advertising carry-over fixed function rate is 23.4%.
The message/vehicle ratio for magazine advertisements is 35%.
Figure 6. Media evaluation model prediction with calibration

118
The second and third research hypotheses are also statistically supported after
calibration because the tracking study results are not significantly higher than the model
predictions. As shown in Table 24, although the gaps between the tracking study results
and the model predictions with minimum levels of effective frequency 4+ and 5+ are
decreased after calibration, the tracking study results for preference and willingness to use
are still not significantly higher than the predictions. In other words, more target audience
exposure to advertising message is needed to achieve target audience willingness to use
than to achieve preference and awareness of the campaign message. It is also assumed
that more target audience exposure is needed to achieve preference than to achieve
awareness of the campaign message.
Table 24. Consumer tracking study results and model predictions with calibration
Communication Effects
Calibrated
Model
Predictions
Tracking
Study Results
Difference
(Predicted -
Actual)
Awareness of the new
campaign message
54.5%
54.5%
0.0%
Preference
38.9%*
34.5%
4.4%
Willingness to use
27.5%*
29.5%
-2.0%
* Indicates that the prediction uses only the minimum level of effective frequency.
The consumer tracking study results and model predictions after calibration are
shown in Table 24. For the awareness of the new campaign message, the model
prediction is exactly matched with the tracking study results. For preference and

119
willingness to use, the gap between predictions and tracking study results are dramatically
decreased after calibration. It is assumed that needed target audience exposure levels to
achieve target audience preference and willingness to use are higher than 4+ and 5+.
However, since there are only two South Korean airline companies in the market, the
tracking study results showed relatively higher target audience preference and willingness
to use which can not be sufficiently explained by the new campaign effects. For better
predictions of preference and willingness to use, it is recommended to use pre-test results
to predict and measure changes in target audience preference and willingness to use before
and after the ASIANA campaign.
To summarize these results, the estimated frequency distribution for the three-
month campaign with calibration is shown in Figure 7 in comparison to the tracking study
results and to the estimates without calibration.
Figure 7. Estimated frequency distributions for the three-month ASIANA campaign

120
It is estimated that about 54.5 percent of the target audience is likely to be exposed
to the campaign message at least three times and about 39 percent and 27.5 percent of
target audience is likely to be exposed to the campaign message at least four times and
five times, respectively.
Critical Analysis of the ASIANA Advertising Campaign
The total expenditure of the new three-month ASIANA advertising campaign “Her
name is ASIANA” was $1,231,474. Most of the expenses (93.5%) were for network
television advertisements and only 6.5 percent was used for magazine advertisements.
Total network television and magazine advertising costs are $1,150,849 and $80,625
respectively. Vehicle allocations for network television and magazine advertising are as
shown in Figures 8 and 9.
1749SP |
|0.5%
Jayeon |
B 0.8%
News Today |
01.5%
1929SP |
â–¡
o
*
2249SP |
032.5%
Jae 3 Noon |
003.2%
Pankwan 1 |
00*05 4%
Mini-series |
1259SP I
1 :’16 0%
21:00 News |
mmam7.3%
127 Others I
20.0% 30.0% 40.0% 50.0%
Percent of Television Budget ($1,150,849)
Figure 8. Budget allocation among television programs

121
After calibration, the average CPM-MSG for network television advertising is
about $41, while it is about $61 for magazine advertising. As shown in Figure 8, network
television advertising used a variety of vehicles which are mostly less than 0.5 percent of
total television expenditures. A total of 137 television programs or spot television
advertisements were used for network television advertising. For magazine advertising, a
total of 14 weekly and monthly magazines were used once a month for the three-month
campaign.
News Maker
News People
Jugan Hankook
HanKyung Business
Economist
Jugan Maekyung
Walgan Chosun
1 Shin Dong-Ah
Newsplus
Hankerei 21
Newsweek
Jugan Chosun
Sisa Journal
Win
5.0% 10.0% 15.0% 20.0% 25.0% 30.0%
Percent of Magazine Budget ($80,625)
Figure 9. Budget allocation among magazines
The four most frequently used television vehicles are “21:00 News,” “spot
television advertising at 12:59,” “Mini-series,” and “Pankwan Pochung-chun part one.”
However, the CPM-MSG figures for these programs are mostly higher than the average

122
CPM-MSG ($41). Only the CPM-MSG for “spot television advertising at 12:59” is below
the average ($34), while other programs’ CPM-MSG levels are higher than $50. Among
the 137 television programs, the best CPM-MSG figures are shown in Table 25. About
seven of the 10 best vehicles in terms of CPM-MSG are spot television advertisements,
although the spot television advertisements were less than 25 percent of the total network
television advertising. It is estimated that, generally, spot television advertisements are
better than program advertisements in terms of CPM-MSG for the ASIAN A campaign.
Table 25. CPM-MSG figures for the most efficient 11 television programs (vehicles)
Program (Vehicle)
Ads
Cost ($)
Per Insertion
Message
Ratings
CPM-MSG
m
Comedy Punch (Comedy)
1
$213
8.60
$0.83
0109 Spot Advertisement
1
206
2.70
2.54
0629 Spot Advertisement
2
155
1.55
3.33
0729 Spot Advertisement
4
318
2.73
3.88
0159 Spot Advertisement
1
206
1.00
6.87
0749 Spot Advertisement
8
766
3.40
7.51
Manam (Drama)
2
730
2.75
8.85
1334 Spot Advertisement
1
1,096
3.50
10.44
Mangang (Drama)
1
730
2.30
10.58
0559 Spot Advertisement
1
1039
3.00
11.54
Insang part three (Drama)
1
2,524
7.20
11.69
($1 = 800 Won)

123
In the US, according to SMRB, within-program commercials are considered to be
better in brand/message recall than between-program commercials. However, according
to MSK, between-program television commercials are better than within-program
commercials in terms of advertising message ratings in South Korea. The main reason is
that there is no real within program commercial in South Korea. Unlike the US television
commercials, the South Korean within-program television commercials can not be
televised in the middle of the programs. The within-program commercials can be shown
only two ways in South Korea, such as after program greetings and before ending
captions. Therefore, there is no major difference between these two types of
advertisements in South Korea. However, there appears to be no good explanation as to
why the between-program commercials are more favorable than within-program
commercials in terms of CPM-MSG figures. South Korean television audiences seem to
escape within-program advertisements more than between-program advertisements with
“zapping” because they can anticipate when the typical within program advertisements are
shown, such as after program greetings and before ending captions.
As shown in Table 25, the “Comedy Punch” seems to be the best vehicle for
television advertising in terms of CPM-MSG. However, the CPM-MSG is unreasonably
low. It might be an unusual result because the vehicle was only used once for the three-
month campaign. The average CPM-MSG for network television advertisements ($41)
also seems to be better than that for magazine advertisements ($65).
The magazine ratings, costs, and CPM-MSG figures are shown in Table 26.
Among the 14 magazines used for the three month ASIANA campaign, the CPM-MSG

124
figures are relatively high for several magazines, such as Han Kyung Business, Win,
Economist, and Jugan Mae kyung.
Table 26. CPM-MSG figures for all magazine advertisements
Magazine
Ads
Cost ($)
Per Insertion
Vehicle
Ratings
CPM-MSG
m
Sisa Journal
3
$2,250
6.6
$32.47
Shin Dong-Ah
3
1,875
5.3
33.69
Walkan Chosun
3
1,875
5.3
33.69
Jugan Chosun
3
2,250
5.4
39.68
Han Kyurei 21
3
2,000
4.4
43.29
News Plus
3
1,875
2.8
63.78
News People
3
1,500
2.0
71.43
Newsweek
3
2,250
2.8
76.53
News Maker
3
1,500
1.7
84.03
JuganHankuk
3
1,500
1.6
89.29
Jugan Maekyung
3
1,750
1.6
104.17
Economist
3
1,750
1.0
166.67
Win
3
2,875
1.2
228.17
Hankyung Business
3
1,625
0.5
309.52
($1 =800 Won)

125
Suggestions for Future ASIANA Campaigns
To improve the campaign efficiency, the vehicles that have unreasonably high
CPM-MSG figures will be replaced with those that have more favorable CPM-MSG
figures in the campaign schedules. Then, the increased efficiency will be suggested for
future campaigns. It is also important to examine the estimated percentage change of
reach 1+ or 3+ as the monthly budget changes. Appropriate monthly budget levels will be
suggested after reviewing the three monthly schedules.
Selecting Appropriate Vehicles
As explained earlier, some of the advertising vehicles have unreasonably high
CPM-MSG figures. Therefore, adjustment of vehicle selection also can increase the
desired advertising communication effects with the same or lower expenditure. According
to the media model evaluation, there are about 12 television vehicles plus four magazines
which have relatively high CPM-MSG levels. The list of these vehicles is shown in Table
27. The special programs, which are not likely to be televised again within a year, are
excluded from the list.
If the vehicles in Table 27 are replaced with the vehicles in Table 28, which
includes those with relatively favorable CPM-MSG levels, the estimated campaign
message reach 3+ would be increased from 54.5 percent to approximately 57.2 percent.
Although the total expense with the modified schedule is reduced about $1,800, the
estimated reach 3+ is increased about 2.7 percent. Since the cost-per-effective reach2
2 $1,231,474 _ ^22^% Where, $1,231,474 is die total advertising campaign budget and 54.5 percent is
54.5%
the effective reach 3+.

126
point is about $22,596, a 2.7 percent increase is worth $61,009. The estimated message
reach 1+ also would be increased from 90.1 percent to 91.0 percent with the replacement.
Table 27. Relatively high CPM-MSG vehicle list in the ASIANA campaign
Television Vehicles
Ads
CPM-
MSG
Television Vehicles
Ads
CPM-
MSG
Sae-sang Yeut-bo-gi
2
$403
Ah-dul-eun Hae-kyul-sa
i
$231
1159SP
1
268
Bo-do-teuk-jip
i
174
SOS Hae-yang Gu-jo-dae
1
265
Bak-sang Yae-suel-sang
i
166
1459 SP
1
263
Jo-kwang-jo
2
156
Magazine Vehicles
Ails
CPM-
MSG
Magazine Vehicles
Ails
CPM-
MSG
Hankyung Business
3
$310
Economist
3
$167
Win
3
228
Jugan Maekyung
3
104
Table 28. Relatively low CPM-MSG vehicles added for schedule modification
Television Vehicles
Ails
CPM-
MSG
Television Vehicles
Ails
CPM-
MSG
2249 SP
1
$34
2054 SP
2
$17
Bam-kwa Eum-ak Sai
1
25
14:35 News
1
23
In-sang part 1
1
34
0729 SP
2
13
In-sang part 3
1
12
2029 SP
2
36
Magazine Vehicles
Ails
CPM-
MSG
Magazine Vehicles
Ails
CPM-
MSG
Sisa Journal
3
$32
Jugan Chosun
3
$40
Hankyurei 21
3
43
News Plus
3
64

127
The significance of the apparently minor change also can be underscored based on
a description of the vehicles listed in Table 28. According to the model evaluations of the
three-month schedules, early morning and late night spot television advertising, sports
related programs, and newsmagazines are more effective for the ASIANA campaign For
television advertising, since the primary target audience of the ASIANA campaign is 20 to
49 year old men in Seoul, Korea, they are likely to be more effectively exposed to the
programs between 8:00 p.m. and 12:00 p.m. Prime time dramas are also good vehicles for
the ASIANA campaign. These are the characteristics of the television programs shown in
Table 28.
For magazine advertising, general newsmagazines, such as Sisa Journal, Jugan
Chosun, and Han Kyurei 21 are better than specialized magazines, such as Economist,
Win, and Hankyung Business because they have greater coverage of the ASIANA target
audience.
Appropriate Budget for Monthly Schedule
Optimum budget levels for each monthly schedule will be suggested after
determining the maximum possible vehicle insertions within each month. The maximum
insertions for monthly and weekly magazines are set at one and four times respectively
Although the monthly maximum insertions for a daily television program can be thirty
times, the maximum insertions is defined to be ten times per program within a month to
insure that a broad range of programs will be part of the optimum schedule. For the
weekly television programs, the maximum insertions are also set at four times. After
setting the maximum monthly insertions for each vehicle, the ADplus optimization model

128
was used to determine the best schedules in terms of message reach 3+ for various budget
levels as shown in Figure 10, 11, and 12.
Figure 10. Reach percentage changes with budget increases for February vehicles,
television and magazines combined
Figure 11. Reach percentage changes with budget increases for March vehicles,
television and magazines combined

129
Cost ($ 000)
-Reach 1 +
-Reach 3+
Figure 12. Reach percentage changes with budget increases for April vehicles,
television and magazines combined
As shown in Figures 10, 11, and 12, the “saturation” budget level for each monthly
schedule is approximately $400,000. It is at this budget level that increases in reach 1+
become virtually flat in response to budget increases, and significant increases in reach 3+
become relatively difficult to achieve. Therefore, it is recommended for future ASIANA
campaigns with similar targets and media to set monthly budgets at approximately
$400,000. Beyond this budget level further expenditures are not cost-effective, and
expenditures below this level have yet to exploit potential
The estimated monthly message reach (3+) for the actual ASIANA campaign is
shown in Figure 13. Since the third monthly advertising expense was about 37 percent
less than the first and second monthly expenses, the estimated reach (3+) is decreased
during the third month. Although portions of the first and second monthly advertising

130
effects are assumed to be carried over into the third monthly campaign effects, the expense
reduction for the third monthly advertising campaign appears to be severe particularly in
relation to the “saturation” budget level described earlier.
-With Carry-over rate
-Without Carry-over rate
35.0% 4
30.0% 4
1st Month
($497,500)
2nd Month
($425,100)
3rd Month
($308,800)
Figure 13. Estimated reach (3+) for each monthly advertising schedule
Since this campaign was new, it might be necessary for the first monthly campaign
to dispose of more expenditure than the average of the three-month campaign
expenditures. However, considering a monthly carry-over rate of 23.4 percent, the end-
of-campaign message reach (3+) would be greater if the monthly expenses are reallocated
according to the “saturation” level identified earlier, as shown in Table 29.

131
Table 29. Suggested monthly budget levels and estimated message reach 3+
First
Month
Second
Month
Third
Month
Total
Monthly Budget Levels Used in
Actual Campaign
$497,500
$425,100
$308,800
$1,231,400
Estimated Message Reach 3+
67.5%
74.5%
54.5%
196.5 %
Cost-Per-Effective Reach Point
$7,370
$5,745
$5,666
$6,267
Suggested Monthly “Saturation”
Budget Levels
$439,712
$399,998
$389,796
$1,229,506
Estimated Message Reach 3+
65.5%
74.0%
66.6%
206.1%
Cost-Per-Effective Reach Point
$6,713
$5,405
$5,853
$5,966
($1 = 800 Won)
The estimated message reach 3+ at the end of the third month could be improved
to 66.6 percent from 54.5 percent without increasing the total three-month campaign
budget The reach 3+ gap of 12.1 percent that occurs between actual and suggested
monthly budget levels is gained with the same vehicles used in the actual ASIANA
campaign. Optimization, which selects the most efficient vehicles, is not involved. The
total points of estimated message reach 3+ are also improved by 9.6 from 196.5 to 206.1
Using the original plan’s cost-per-effective reach point as a benchmark, the 9.6 additional
points are worth $60,163 (9.6 x $6,267 = $60,613).

132
Sophisticated versus Naive Comparison
As explained in Chapter IV, this study employed a sophisticated method to
estimate cumulative advertising communication effects for the three month AS1ANA
campaign Advertising campaign effects also can be analyzed by a number of relatively
simple procedures which are naive at best and most probably misleading (Lancaster,
1993). In this study, the typical naive estimations will be compared with the results of the
sophisticated method employed in this study and the comparison will show how the
potential difference is large and misleading.
Pooled Approach
An important issue is the time frame that should be used in reach-frequency
analysis. In many cases, the selected time frame may be too broad to estimate appropriate
advertising communication effects. Armed with reach-frequency models, many planners
often employ quarterly or annual evaluations. However, with these figures, consumer
forgetting effects and advertising clutter are essentially ignored. The comparison is shown
in Figure 14.
As shown in Figure 14, the estimated message reach 3+ with the naive pooled
approach is almost 100 percent on the third month when the tracking study was
conducted. The other two lines use a monthly time frame with or without the 23.4
percent carry-over rate. The second line with the carry-over rate based on a monthly time
frame produced the estimated message reach 3+ exactly matched with the tracking study
results. The gap between the estimated message reach 3+ for the pooled estimation and
monthly estimation with carry-over rate is about 45 percent.

133
30.0% 4— i f
1 st Month 2nd Month 3rd Month
($497,500) ($425,100) ($308,800)
-Pooled Insertions
-With Carry-over rate
-Without Carry-over rate
Figure 14. Estimated message reach (3+), naive versus sophisticated approaches
Vehicle versus Message
This study also uses actual message ratings for television advertisements and
message/vehicle ratios for magazine advertisements. Since there is an obvious difference
between advertising vehicle and message exposure, a significant amount of over¬
estimation is also expected if the vehicle data are used. Figure 15 shows the differences
between vehicle data and message data, for the estimated reach 1+, 3+, 4+ , and 5+
For the estimation of reach n+ with vehicle data, the monthly carry-over rate of
23.4 percent was also applied. The estimated reach 3+ with vehicle data is about 70
percent, for example, while the message reach 3+ is 54.5 percent.

134
0.0% 1 i
Reach 1 + Reach 3+ Reach 4+ Reach 5+
-Vehicle Data With Carry-over Rate
-Message Data With Carry-over Rate
-Message Data Without Carry-over Rate
Figure 15. Estimated frequency distributions with vehicle and message data
In this Chapter, the estimated campaign effects were compared and tested with the
tracking study results. Possible improvements in the normative framework and procedures
were explained with backward adjustments. The ASIANA campaign was also analyzed
for possible improvements and appropriate vehicle selections were suggested for future
ASIANA campaigns. In this study, the sophisticated approach which uses a carry-over
rate based on a monthly time frame and actual message ratings or message/vehicle ratios
were used for accurate prediction. A comparison shows that the typical naive approaches
that use inappropriately long time frames or vehicle data may be inadequate for accurate
prediction
Future ASIANA campaign effects may be better predicted with the improved
framework and the media evaluation model used in this study. Chapter VI will present a
summary of key findings, conclusions, implications, limitations, and suggestions for future
study.

CHAPTER VI
SUMMARY, CONCLUSIONS, AND IMPLICATIONS
Summary and Conclusions
Contrary to the skepticism at the possibility of predicting advertising
communication effects, this study argued that advertising communication effects can be
predicted by taking into account the target audience, media schedules, vehicle/message
information, and other influential marketing factors, such as advertising message quality,
market share, and competitor promotional activities. This can be done with confidence by
analyzing the target audience, costs, vehicle and message data using a media exposure
distribution model and normative framework
Due to the lack of scientific development of media evaluation models in South
Korea, a South Korean brand, ASIANA Airline, and its advertising campaign were
selected as a case study. Although the model used in this study was developed in the US
and the South Korean advertising environment is different from the US, the principles and
procedures in this study demonstrated that the predicted advertising communication
effects and the target audience awareness of the new campaign message were close to the
consumer tracking study results. It could be done by comparing the model predictions
135

136
with the tracking study results. The ideal procedures for predicting and evaluating
advertising communication effects are summarized in Figure 16.
Decision Factors
such as Marketing,
Copy, Media, and
Competitor factors
Calibrating Options
1. Effective Frequency n+
2. Message/Vehicle Ratio
3. Carry-over rate
Figure 16. Ideal procedures for predicting and evaluating advertising effects using media
exposure distribution models
Prior to the application of the media exposure distribution model used in this
study, survey research was conducted to examine what kind of advertising communication

137
effects are used in the South Korean advertising industry. The survey research results
showed that South Korean advertising practitioners were aware of the current state of
media evaluation concepts and that over 90 percent of them actually used advertising
communication effects for the evaluation of their advertising schedules. However, most of
them also answered that their current procedures are in need of improvement. The survey
results also supported the use of reach/ffequency estimation to predict or evaluate
advertising communication effects in the South Korean advertising industry because of the
wide acceptance and actual use of these terms.
For better prediction in the South Korean market, the exposure distribution model
was modified with estimated self- and cross-pair regression equations based on data
collected from the Korean Gallup. The original US model produces slightly lower
estimations for reach n+, compared to the modified model’s estimation. This seems to be
reasonable because there are many more advertising media options and vehicles in the US.
The South Korean self- and cross-pair ratings would be higher than those of the US
because on the average there are fewer chances to be exposed to more than one vehicle in
an advertising schedule in South Korea, hence greater reach for the average vehicle or
combination of vehicles.
In this study, target audience advertising message ratings, instead of vehicle
ratings, were used for television advertisements and a message/vehicle ratio was also
employed for magazine advertisements. Since there is a clear difference between
advertising vehicles and messages, the use of actual advertising message ratings can
increase the confidence of prediction. The tracking study results also allowed the
calibration of media planning procedures to increase the confidence of future campaign

138
predictions. The calibrating option used in this study was the combination of monthly
carry-over rate and magazine message/vehicle ratio because they are largely unknown in
South Korea.
The study results showed the model’s accuracy in predicting target audience
awareness of the campaign message. This study also revealed that there may be a
hierarchy in consumer processing of an advertising message. The needed minimum target
audience exposure to achieve preference was larger than the needed exposure for
awareness. The needed minimum message exposure for willingness to use was also
significantly higher than the target audience exposure for awareness and preference.
Although many more factors should be considered for accurately predicting or evaluating
advertising effects in terms ofpreference and willingness to use, this study still provided
meaningful results for the estimation of advertising effects on brand preference and
willingness to use
A critical analysis of the three-month ASIANA campaign was presented in Chapter
V, suggesting appropriate monthly budgets and vehicle selections for future campaigns.
The model estimation for an appropriate monthly budget was around $400,000. CPM-
MSG figures were also provided for the vehicles used in the ASIANA campaign.
This study employed a relatively sophisticated approach, which uses actual
message ratings, a narrow time frame, and a monthly carry-over rate, to estimate desired
advertising communication effects. The potential gaps between the sophisticated and
typical naive approaches were presented to show how the differences could be large and
misleading.

139
Although the South Korean advertising industry has grown into one of the world’s
13 largest markets, current media planning and evaluation procedures are in the early
stages and need to be improved. Therefore, in this study, the normative framework,
evaluation methods, including many influential factors, and calibrating procedures will
provide useful suggestions for the South Korean advertising industry.
Implications
There are many benefits to be derived from the principles and procedures used in
this study. First, this study examined how South Korean advertising practitioners perceive
media evaluation terms, such as reach and frequency, and their current procedures for
media planning and evaluation. This exploratory diagnosis can help the South Korean
advertising industry develop a good media evaluation method which fits the South Korean
advertising environment.
Second, the management of AS1ANA Airline can now search for optimum
advertising budgets and schedules with some confidence that their evaluation will be
related to desired communication impact. Appropriate frequency levels, the
message/vehicle ratio for magazine advertisements, and the carry-over rate are
recommended based on key characteristics of the case study results. They can be used as
benchmarks to evaluate the likely advertising effects and effective frequencies for future
ASIANA advertising campaign schedules because they can forecast consumer tracking
study results. It also is possible for advertising planners to understand what is expected
from their advertising budget and the differences between the budget and desired
communication effects for future advertising campaigns.

140
The normative framework and procedures also can be used to predict different
communication effects, such as preference and willingness to use. Although, the lack of a
pretest does not allow this study to do that completely, these effects can be predicted if a
consumer pretest is used. Management also can use these results to determine whether to
put additional advertising expenditures into media or into research designed to improve
the quality of advertising messages.
Because of the enormous expenditures on advertising, advertisers and advertising
planners need to identify the results to be obtained from advertising investments. This
type of study also can be done with other brands in addition to ASIANA Airline.
Ultimately, the normative framework used in this study will provide a useful point of
departure for predicting and evaluating advertising communication effects regardless of
the brand or country being studied.
Limitations and Suggestions for Future Study
Although the problems, hypotheses, data collection, and assumptions are clearly
specified, there are limitations in this study due largely to the lack of information in South
Korea. First, information related to magazine advertisements is not available in South
Korea. Therefore, in this study, a South Korean advertising practitioner called each of the
magazine publication companies and collected the ratings and circulation. It is expected
that the ratings may be biased upward. However, the total magazine advertising
expenditures were only 6.5 percent of the total campaign expense. Also, a calibrating
procedure, such as lowering the message/vehicle ratio, may reduce the possible error. The
magazine ratings and message/vehicle ratio still should be gathered from an objective

141
media research institute. Estimating current magazine ratings and message/vehicle ratios
may be a useful future study in South Korea.
The estimation of advertising carry-over rate based on assumptions also may
introduce error in the prediction of advertising communication effects. For better
predictions, research should be conducted to examine advertising carry-over rates in South
Korea since the advertising carry-over rate is largely unknown. It also can be an
interesting issue for future study, focusing on advertising carry-over rates for various
product categories and time intervals.
Since this study is targeted to the 20 - 49 year old South Korean men who live in
Seoul, Korea, the study results may not be applicable to other cities in Korea or other
demographic groups such as women generally and men over 50 years old. Studies can be
done in other important cities in South Korea if a national approach is appropriate
Another limitation is the lack of pretest results. Since the ASIANA campaign used
in this study was a new campaign, the pretest of consumer awareness of the new campaign
message was not necessary. However, for better prediction of the campaign effects for
preference and willingness to use, it is important to conduct a pretest. With pretest
results, the actual change in consumer preference and willingness to use, instead of their
absolute levels, can be examined using the media exposure distribution model predictions.
Another future study subject is the error estimation of media evaluation model
predictions versus tabulated schedules based on syndicated research samples. Since no
study has yet been conducted for the error estimation of media evaluation models for
television advertising, the error estimation with television advertising schedules may be an
interesting topic for future study. Also, the error estimation of the BBMD model used in

142
this study can be conducted with South Korean advertising media data because the error
estimation with US data may be different from that of South Korea.
For future studies, the desired advertising communication effects also can be
predicted and evaluated for other advertising campaigns in South Korea or in the US.
More studies will help improve and generalize the normative frameworks and procedures
for accurately forecasting advertising communication effects. Clearly there is a great deal
of useful research yet to be done. Hopefully, this study will stimulate other academics and
practitioners to pursue this line of inquiry.

APPENDIX A
SURVEY QUESTIONNAIRE
IN
BOTH ENGLISH AND KOREAN

ENGLISH VERSION

145
In spite of the recent progress of the Korean advertising industry, both the industry and
academic world still do not fully understand the conditions and future prospects of media planning
in Korea. Therefore, the advertising departments of Chung-Ang University and University of
Florida in the United States are conducting joint research to understand how Korean advertising
agencies develop media plans and how they evaluate them. This research is conducted by Kent M.
Lancaster (Professor, University of Florida), who is an internationally known expert in the media
planning field, Jung-Sik Cho (Assistant Professor, Chung-Ang University), and Hyunsoo Park
(Ph.D. Student, University of Florida).
This research is totally confidential and your identity will not be disclosed in any way. Your
answers to the following questions will provide valuable data for the development of the Korean
advertising industry.
If you have any further questions, please contact Dr. Jung-Sik Cho (Department of Advertising,
Chung-Ang University, TEL: (02) 820-5508, FAX (02) 825-5893)
Please return your questionnaire using the enclosed pi e-paid return envelope, or you can
fax your Questionnaire to Dr. Jung-Sik Clio, (FAX 02-825-5893)
* The following questions are about the sources you use to get program ratings and
advertising ratings.
1. Please note the sources used to get vehicle (program or publication) ratings and note how
satisfied you are with the data provided by the source.
Sources Media Ratings Degree of Satisfaction
TV Radio Publication Not satisfied Neutral Satisfied
-2-1012
a. MSK
(Media Service Korea)
b. Lee’s PR
c. KMR
e. Others (Please specify)

146
2. Please note the sources used to get advertising message ratings (not program ratings) and
note how satisfied you are with the data provided by the source.
Sources Media Ratings Degree of Satisfaction
TV Radio Publication Not satisfied Neutral Satisfied
-2-1012
a. MSK
(Media Service Korea)
b. Lee’s PR
c. KMR
e. Others (Please specify)
3. How important do you feel is the distinction between advertising (message) ratings and
program or publication (vehicle) ratings?
Vehicle/message distinction is Vehicle/message distinction is
not important at all -2 -1 0 1 2 very important
* The following questions are about how you develop and evaluate media plans.
4. Does your agency currently develop and evaluate advertising media plans ?
a. Yes
b. No
* If your agency does not develop or evaluate media plans, what is the main reason?
( )
* Please Go to Question 16 on page 5

147
5.Which of the following systems does your advertising agency generally use to evaluate
advertising media plans ? (Please circle all that apply.)
a. In-house computer programs
b. Foreign computer programs (Please specify country, company, and product name
c. Non-computerized procedures
d. Others (Please specify)
6.In evaluating advertising media plans, does your agency evaluate a schedule for
combinations of two or more media categories (e.g., magazines and television together)
a Yes
b.No
7. Please indicate which of the following factors are generally used by your agency when
evaluating advertising media plans (Please circle all that apply)
a. Reach
b. Effective Reach
c. GRP’s
d. Gross Impressions
e. Average Frequency
f. CPM
g. CPRP
f. Others (Please specify)
8. If your agency uses reach when evaluating advertising media plans, note the most
representative definition of reach.
a. Media vehicle exposure (e.g, saw publication or program)
b. Advertising exposure (e.g., saw advertisement in publication or program)
c. Advertising impact (e.g. recall of product message)
9.Typically, how many times must audience members be exposed or affected to be counted
as being “effectively reached?”
Lower Frequency Limit (check one) Upper Frequency Limit icheck one)
-1+
-2+
-3
—4+
—Other (specify).
— No Upper Limit
- 9
-10
-11
—Other (specify)

148
10. What kind of communication effects, if any, are used when evaluating advertising
media plans? (Please circle all that apply.)
a. Not used
b. Advertising exposure
c. Recognition
d. Recall
e. Comprehension
f. Interest
g. Conviction
h. Preference
i. Attitude toward advertisement
j. Attitude toward brand
k. Intentions
l. Willingness to buy
m. Purchase
n. Others (Please specify) .
11. How important do you feel estimates of communication effects are in the overall
evaluation of advertising media plans ?
Communication Effects are Communication Effects are
Not important at all -2-1012 Very important
12. Does your agency typically weight vehicle audience data, GRP’s, or schedule exposure
distributions to account for advertising communication effects (i.e., exposure, recall, etc.)?
a.Yes, Approximately percent of all of our agency’s media plans are weighted
b.No, our agency does not typically use a weighting procedure.
* Please Skin to Question 15 oil next page
13.How are the weights derived ? (Please check all that apply.)
a. Standard set of weights which are established by your agency
b. Judgmental weights that are largely subjective, based upon a review of
communication effects data (e.g., exposure, recall, attitude change), and
vary for each brand and product
c. Formula weights that are based upon review of communication effects using
some standard formula or procedure
d. Others (Please specify).

149
14. For the media categories listed below, please: (1) note whether schedules are weighted
and, if yes, (2) indicate the range of the weights typically used.
Media Category Weights Used? Typical Overall Weight
Yes No Not Sure (Please Specify Range
e.g„ 20%-50%....)
a. Newspapers
b. Network TV:
Morning time
Prime Time
c. Spot TV:
Morning time
Prime Time
d. Cable TV
e. Magazines
f. Network Radio
g. Spot Radio
15. In those situations when your agency does not use weights in evaluating advertising media
plans, please indicate the major reasons. (Please check all that apply.)
a. Lack of data to substantiate assumptions
b. Too much time spent justifying weights
c. Difficult to be accurate
d. Clients don’t require such sophistication
e. Each media planning situation is unique
f. Overkill in the manipulation of numbers
g. Too judgmental
h. Others (Please specify)
16. Do you feel that vour agency’s current procedures for evaluating advertising media
plans are in need of improvement ?
a. Yes
b. No
17. Please indicate which of the following factors are in need of improvement for the
development of the media planning field ? (Please circle all that apply.)
a. No improvement needed
b. The improvement of media buying procedures for newspaper and magazine ads
c. The improvement of media buying procedures for TV and Radio ads
d. Keeping the balance of demand and supply for TV and Radio ads
e. The training of media planning specialists
f. The advertisers’ recognition of the importance of media planning
g. Others (please specify) __

150
18. Here are four solutions to keep the balance of demand and supply for TV and Radio
advertising. Please insert your ranking from I to 4. The most important solution would be
No. 1 and the least important solution would be No. 4.
A drastic improvement of advertising rates
The actualization of advertising rates
The enlargement of advertising time
The establishment of new broadcasting company
19. Do you think that the beginning of CATV (cable TV) can be helpful to keep the balance of
demand and supply for TV advertising ?
CATV is CATV is
not helpful -2 -1 0 1 2 very helpful
in balancing in balancing
supply and demand supply and demand
20. Do you think that the introduction of foreign media planning procedures such as American
reach and frequency computer programs would be helpful for better media planning?
a. Yes
b. No
21. Further comments.
22.What is your title ?
Thank you very much.
Please return your questionnaire using the enclosed nre-naid return envelope or
you can fax your questionnaire to Dr. Jung-Sik Cho (FAX 02-825-5893)
If you would like a copy of the survey results, please complete the following:
Name :
Address:

KOREAN VERSION

152
g'á 4 "ti 4 4? 4 U "1 4 University of Florida 43 4 4 4 4 4 31 4 4 ti '1'- 4 ll 4.
nL4 ¿1 '!) 4.4 4U 44 -‘ó1’ --11 -°| 444 r| -V,- o] ( University of Florida *2 3 ri o]] á]
°1 ‘ti31 "1131 315!(Media Planning)22 .frgg 2*4*4 *3) t>* 42 3144*®! "1131 31
Í) 431®! 443) 3131 4** *MI ®)4*44. "1131 3151 *43)4 43)422 44 *44
University of Florida®! Kent M. Lancaster 2*-* lll * 44. University of Florida®! 4
4* (Ph.D. student) *4 444* * g*2) *4* mg 7)|g ^Jtd| ^ a)-e-o|| m4 4
* 24-i- S3|2 "11*1131512! -am®! 422 4444 4 44* 3144-*3| 2 244- 42 51
*44. 4B14442 f-44 44431 444 *4244, 4 4444 4-44 *4422 4
4 31^4 44 4231 4-1-44 a* 44-, £4 444 4-4312 4*44 444431* 44
44 a* 4-1- 44 2444.
44 42 44-1- 431 444 44* *4244 4*4 444 *4* 44 444.
444 4* 4444 44 * 4*31 431 4*4 44 51244 *4 3142 *2
*244 244 24 4244 4444. (4*1 (02)820-5508, Fax (02)825-5893).
*4*4 4-Ü44 4* *** **4*4, 44* **4 **3| ^4 **31 *442.
44 Fax 42 (825 5893)2 23) *4* **44.
♦ 4** 444 31*44 43| 4*4^4 42 *4 4*31 *4 4* 444.
1. TV 44* (TV Program Ratings), 44* 314* (Radio Program Ratings), 4*4
444 *4** 42 *22 44 42 4*44? 2* 444 42-g-t- 4-1- 4*
43) 42 42 4442 4*44? (3)44* *31 v2 2431 *3142.)
42* *4* * mm* ** *#2
TV 442 4* *4 444 2*
4** 44* *4* -2 -1 0
D.MSK
(444 44— 244)
2)4¿ 44
(Lee's PR)
3).KMR
(44 4314 444)
4).*4 22 514 a4.
5). 44
(44 4 *3| 44 *442)

*
i
tí
tí
i
&
r|r
*
tí
jí
tí
ja
tí
«a
tí
<&
_\J,
41
tí
Ki
tí
*l>
SL
tí
[tí
S
H»
bü
tí
tí
i
A
tí
tí
i
s
i
p
tí
k
EÍ
s
r£
tí
tí
tí
#
s
s
tg
tí
tí
A
o(X
¡»
tí
«W _
«a
to
i
H»
tí
I»
tí
tí
w
CTbfB'i 0 ll<>Tft ^fes-g.#*)
¿■fthfH» K-b^-s-4 ifr =rk k-ftib ft Ib K^-k-¿
Th Hr fr)* ge hr ir-gc -ft (4^-t ‘íffBWb fcfx ‘ft ft -ftftft Is tr IK ft ft Trft
(Tkfti k-fc I»* ft #¡6 ftft)
g»k (S
tí
O
l®
w
tí.
Jí,
tí
tí
tí
-Hi
a
tí
»&
w
?!
s
S
tí
i
jí.
_ to
5 ji
ü- iv
3 Ja.
S «»,
Je,
ir
IV
m
je
tí
|*i
B|ni
4
tí
r|r
m
a
tí
ja
r|r
XJ
a
tí <
4»
tí sg
tí tí
tí H>
J| *
tí *
tí tí
<
ft
k
tí
tí
Jk
tí
HH
nb
tí
41
tí
Jo
te
I tí
M tí
tí
tí
1= tí
0 o|fl ft
tí
to J|4(
¿\á h^Io íT{s^{a -alo Ih tí; -§-[y -t-i^-l* -ft-12 lo ft "5
¿■l^h^ío ¿Tío la-¿r üTiTi^ k IT'S: ■^•'3‘Tb tolT^r ‘-¿2 (n{o lo fp» IT"2? jé ’<

154
5. 7)44 4S))7Hr *2 d]| *j) /]]4(media planning)?) s§7}® 7] 4 cf® ® 44
'Ü-’g* *5. 4-|-*BM4-? (*H^S)ir 3H S-Jf- j£7) o(| ^*7)2.)
1). 44 44-g- 4* 7}4| S£3i
2). 3*4 23.24 ( = 3.24 $ ^4- 44 **7)2.)
3). 3®4 44 <8->g
4). ?M (*4* 'S'á -f* £.444.)
( )
6. 44 444 «7>* *4 444 447}* * 7fl 0)^4 44* *4 7}-§-s}2 sl®44?
(°fl: TV4- *7l 4-31® *4 44*}2 Í7})
1). “II
2). 44 i
7. 444 447}* 42 44 4144 44* 4*11 4* ■? 44 .a¿-S-fr 4* 44)44?
(44-4® 4241 a.® 5.44 ^-84A.)
1). 2>£-|- (Reach)
2). X4 24* (Effective Reach)
3). Grp
4). Gross Impressions
5). 'll3-4' (Average Frequency)
6). CPM (4 44 424)
7) CPRP
8). 44. ( )
8. 44 #2 44 444 Í4-Í- 44 £.4» (Reach) * 4-8-444 444 447}44®
2 44* 4H4 44 4444?
1). 42 4141-11 44 2® (4: 427)- *44® *714- ££3^1- Jit" 4 )
2). *2 7}44 44 2# (4: *7]4- as.a.44 *44® 42* * 4-f)
3). 4-24 44 (Advertising effects) (4: *2 47)7)4 7)4 2* i)*,!)
9. *24 ÜJ&5fcS:(Effective Reach)® 471* 4, 42 *24 '?!*:rltreciuency)4
¿47144 4*«>4-2 *71- 4*44? (4 ¿4 444 £* V &714 ^*7]a.)
1). 4 ¿ (Lower Limit) 2.) 4 4 (Upper Limit)
—
1
4
4*
(+1)
— '4“ * 4 #4-. (No upper limit)
2
4
4*
(+2)
--44 9 4
3
4
4*
(+3)
—44 104
4
4
4*
(+4)
—44 114
-7M
(
)
-m ( )

155
10. *51 4)44 ;'o 7}f °] 4, L1 fr Ó1 °l 4 *51 3.4-(Cominunication Effects)^!-
4-8- 44 42 4f44? (4*84f 4*4 H.-f 44 «I ftiH-i-)
1). 4444 44 $4.
2). *5141 44 44 (advertising exposure)
3). *5l4 ©] 4 (recognition)
4). 4:2.2] i]* (recall)
5). *514 4 til (comprehension)
6). *51 til til 4 44 (interest)
7). 4512)4- *fti) ti]4 44 (conviction)
8). 4 l2] ¡- **til ti]4 44 (preference)
9). 451 til ti]4 4-i- (attitude toward advertisement)
10). 451S|4 4# (brand)ti] til 4 *8-f- (attitude toward brand)
11). 4512)4 4s f4| til 4 * til 2] 4 (willingness to buy)
12). 4512)4 tiff 4 ^ti) (purchase)
13). 7)4 ( )
11. 451 tiltil 442) $7}ir s(, 43 24(Communication effects)2) 44 “1 444
4-S.445! 44 4444?
451 2* 444 451 S.4 444
44 4444 $4. -2-1012 til-?- 44 44-
4?)
12. 444 tiH84 til44 444 $4-2.34-4, "Util 444 áMt 44 •8-44 44451
4f44? 4 4-4 44, 7ltil44 *51 3.4 (*5iti)4 i)4 -f- ) til 44
4 4til4 451 4$ 4—41 44 f-4 *8*8-4- 4-8-451 4#44?
(4-8- #4 x 44 4|#4 *5*1- 44 tiltil til4til4, 7)444 *51 3.44 44,
44 3.47} 7] til 4 4 44 444 7}f>4-& 4-t *4*ti -§-)
l). 4, 44 ( ) % 4£4 4til 444 7>^4«- *8*8-1- 4-1-451 4f44.
2). 445-, -44 tiH>4ti!4t 7V-S-4* tv *8*8-1- 4*44 4^44.
* (48 4444 44 15 ")°5 7HI45-)
13. 44 *8*84 44 7>*4-g- ^ji 58*3*44? (4*844 4*4 £f 114 4*844)
1). 444 *2444 ti) 4*8* *8*8
2). 4-5* 2.4- ct-fr, 4$-, 4£ 44-.. f)4 4S-4 4444, tilf-g- **44
*4-22 *4*4 tiff-til 44 *4 444.
3). *5i 244 4 £4 -e-444, 4* *8*8 44 444 4-8-
4). 7)4 (4til4 4*8 44 444.)
(
)

) ftk -a
*rlk kft ft Hi Itefrlk Itelte tsürft '(9
kft fcft-g-ft fck [te Ha (S
CT>
Ja. CO te T-
4 oj l-i 4
4 °i> F[0 Á
otl 4
ft afe
-|> •a
4l 0$.
tutu ^
p4j A
oí afe
a ft
IV
xa ja
*¿
>4 TT-
>lo
A
4
A
a 4
S r'r
JÜ -|o
,2. A
-*â–  4
A A
4 Ja
W 0*1
ofe ft
A i
A afe
afe
4
ja.
i\t
A
ft
afe
ft
+
1*
A
afe
HX
ni
A
ft
r+t
4
efe
4t
Ja
S
i
4
0|o
Ja
a
efe
Ji
Pft
»ln
A ^
A
ft
A
ni Ji
A 4.
A Ja
•o
afe
ofe
V
si,
<41
ft
A
¿i
J»
«l|o
A
A
4
r|r
ja
rU
Ja
A
4
m|n
COSClínjitOMH
i 4 4
4 4 4
— 4 4
4 >fr
4 4
4 Ja
4 4
4 4
s, &
H« F[o
afe afe
4 4
4 ft
4 -ftt
B -I-1
afe [J
TU
* «
Ai
4 S
ta *
«¡lo
ft
ft 4
4 4
4 °ÍX
oi, 4
4 «1«
4 4
4 4
4
4
flr
afe
4
4
¡s.
r|°
OÜ
ft
ja
4
4
sa
i»
lo
ft
4
IV
[o
efe
4
4
1»
A
4
ft
C-fthftft k®' 4--S ÜoTBt Ats<e) ¿4hB>
ü-á- ^ftlo IT ftia« kft Kft-S-4 ft&ft *k4rk kltekífelte kkk 4ft 'SI
iv 4 4 -at
A ft A 4 -fe
ft Ja
JU, ft
4 Jü
[o 4
ft
a|i 4 4
4 4
H
<
IV 4 efe
A A 4 ft ift
Ini jS, 4 ja
4 4 H
H ^ ♦ H
< <
4
4
k
#
p-
4 *tk
4 4
*4
sa
o[o
4
oj»
ja
ft
ta
«fr
4
4
ft «Di
ÃœL
tN3
i *
¿i fd£
- Ja
4
Iteftlte Itelte [s^-ft % ft ftft ftftft üMs-f-k ftft(z) "rHi-i ftftk¥ A
^■ft «rfe-g-k -t-ftft l-kí-ft Itei-lkv 4|t(T) kite4- kite tsH» ti

■¥
iu)a
w|o
H»
£
in
SL
&
«0
£
£
A
-M
r.i
df
n|ni
»?.
Hi
Hi
I»
ci
r+t
r|o
0$
dH
HU
0
«In
£
si
£
-w
si
_s,
K>
W
£
«U
£
r|r
£
Hu
2£
do do
£
5
«V
*
4fc
£
£
O
Hu
i£
£
£
nifE
s,
-r-
J£
£
rfi
*
ft
T
s
|U(S
f1°
£
A
H*
r—>
£
r$i
£
1^
1*»
£
4J
£
A
m oin r|r
id.,.
01
g £ d«
w j$ it|«
^ o|fl
Hi £
*£
(a X‘
-e -+>
JH o|H
W £
^£
Hi '
H>
wi
Hu
Tt
R
»|«
I®
Hi
#
r£
»|o
da
£
£
â– HX
£
_x,
Hu
Hu
£
°H>
£
£
i£
£
si
£
£
HH
bB
¿■^ h [o fc'a ?o-
-sMH-ir H*h&4« &bro& lo*fc M* *4? k?lk Ikklk IH* ‘12
to i;
£ £
£
H>
it
£
£
|w
Hu
U
r&
5
Ht
si
H*
£
£
A
£
o
£
Hi
£
£
£
si
£
£
•o
Hi
£
H
via r£
»(o
£
s*
£
*2
s
?s
£
ia
1
»|o
to
da
£
1
£
-HI
da
•f
£
si =
£
£
ja
1»
Hu
U
to
<&
2.
A £
a.
£ *
4J
o dá
is,
£ £
3,
' £
Ü
°t*>
5 «I*
f <
si
£p
£ >
- b
£ <
£ da
£ ®|n
nr of
is, £
H*
1*1
-ft.
H*
ah
H>
<4>
Hi
H>
«tp
mi*
iHJ
oh
ft
-2,
ir
oh
Hi
£
£
£
£
Hi
ojsj oSfL o^L oijL
0^ hJ Hi Hi
■^ £ H> |s>
of. £ op all
t| ^ J2, -H
•S £ A PH
£ mi £
rh. jé, -
jh
r°
rh
'WS5TF&T írfcMzfr <£fcM?^£
-b>h^Ii. ± tr« log.* * tó art* #IF|H tsfei* ^ tsTT-M’* 81

APPENDIX B
TWO TYPES OF MAGAZINE ADVERTISEMENTS

159
HER NAME IS ASIANA (Piano Scene)
“I like to play piano because music makes us beautiful. Although I can not write a great
song, 1 can play it and let you hear it. I'd like to keep the beauty of music in my hands. ”
She who has a beautiful hand.
Her name is ASIANA.

160
HER NAME IS ASIANA (Poetry Scene!
“I like to write a poem because poetry makes us beautiful. Although / can not write a
great poem, I can let you hear it. I’d like to keep the beauty ofpoetry in my smile. "
She who has a beautiful smile,
Her name is ASIANA.

APPENDIX C
TV PROGRAM RATINGS,
ADVERTISING MESSAGE RATINGS,
MAGAZINE RATINGS, AND COSTS
FOR THE ASIANA ADVERTISING CAMPAIGN
IN BOTH ENGLISH AND KOREAN

162
ASIANA Advertising Campaign Target Audience Vehicle and Message
Ratings and Costs in English
Network Television
Media
Program
Vehicle
Ratings
Message
Ratings
No. of
Ads.
Cost US $
(Per Ad)
KBS-2TV
0959SP (Spot Ad)
1.95
1.95
2
2,113
1159SP
0.3
0.3
1
2,414
1259SP
3.5
3.5
1
4,736
1339SP
2.7
2.7
1
6,314
1354SP
2.1
2.1
1
6,314
13:00 News
2.0
1.7
1
4,523
1749SP
2.9
2.9
2
1,811
2054SP
6.23
6.23
6
3,103
2249SP
4.57
4.57
6
4,736
SOS Haeyang Gujodae
0.6
0.5
1
3,971
Dawon Bok
PI
0.9
1.2
1
4,523
Dawon Bok
P2
3.1
1.2
1
4,523
Mok-yok-tang-zip. . .
3.6
2.2
1
3,158
Mini Series
6.7
4.33
10
6,570
Bamkwa Eumak Sai
P2
5.9
4.15
2
3,101
Bob Roberts
2.7
3.5
1
3,956
'Sulnal' ahchim
PI
0.4
0.8
1
1,309
'Sulnal' ahchim
P2
2.2
1.0
1
1,894
Keum
P2
4.9
2.7
1
7,988
Keum
P3
5.1
1.7
1
4,961
Saekye-ro Kaneun...
P2
5.5
1.3
1
3,128
Saegi Masul
4.0
1.75
2
4,523
Star wa gohyang
PI
1.7
1.8
1
5,138
Sigan Morebat
PI
5.8
2.6
1
4,545
Symya Chujuck
PI
4.0
2.5
7
3.101
Wuri gayojae
PI
1.1
1.1
1
4,879
Jae 3 noon
4.5
2.99
8
4,545
Quiz, Norae-reul Cha..
5.0
2.2
1
4,523
Quiz Show, Myung Jang.
5.4
6.7
1
4,523
Towering
PI
5.1
4.8
1
1,856
Towering
P2
3.8
3.3
1
1,463
Teuksun Younghwa
11.0
2.4
1
4,110
Pankwan Pochungchun'
PI
5.5
3.3
10
6,244
Pankwan
P3
7.4
6.0
2
3,021
Pankwan
P4
1.8
1.5
1
4,110
Pankwan
P5
5.8
2.6
1
4,110
Sports Jeungkesuk
6.3
3.47
3
6,053
Aachim Dalinda
P2
1.8
1.22
7
1,564
TV Insang
4.9
3.65
2
3,394
Jo, KwnagJo
2.6
1.4
2
6,570
Total
96
*414.373

163
Media
Program
Vehicle
Ratings
Message
Ratings
No.
of
Ads.
Cost $
(Per Ad)
MBC-TV
0559SP
3.0
3.0
1
1,039
0729SP
2.85
2.85
2
1,123
0749SP
3.4
3.4
8
766
1019SP
4.1
4.1
1
6,314
1239SP
4.1
4.1
1
6,314
1259SP
2.31
2.31
6
2,348
12:55 News
5.9
4.4
1
4,691
1459SP
0.6
0.6
1
4,736
1929SP
0.95
0.95
8
2,889
21:00 News Desk
6.7
4.89
11
7,684
24:00 News
1.4
0.6
1
3,161
MBC Grand Prix
1.9
2.1
1
4,459
MBC News Today
1.7
1.8
1
1,376
Kaseum-ye Dot-Neun Kal
4.2
2.7
1
2,336
Kajok
PI
1.5
1.9
1
5,865
Keudae An-ae Blue
3.3
3.7
1
2,599
Geurimdosa Hong Kildong
5.2
1.9
1
4,328
News Today
P2
2.7
1.97
21
1,335
News Teukbo
1.5
0.5
1
2,535
Rocky V
6.9
5.9
1
8,306
Myungtae
P2
7.0
4.8
1
3,668
'Back-Sang' Yesul Sang P1
1.4
1.2
1
5,951
Bulhang-han Hangbok
2.0
2.3
1
5,239
BBongbat Nagunae
4.0
1.8
1
2,288
Saehae Bok Mani Ba
8.3
4.2
1
5,269
'Seucho' Paewang
PI
5.0
2.9
1
5,764
'Seucho' Paewang
P2
5.8
6.1
1
3,656
Saesang Yutbogi
4.6
0.5
2
6,056
Sports Highlight
5.1
4.0
2
6,056
Encore Best keukjang
1.7
2.3
1
4,264
Ongojip Juen
0.4
0.6
1
1,406
Wujung Mudae
4.5
3.28
10
4,264
Worldcup Kaneun Kil
P3
1.7
1.2
1
1,470
Worldcup Kaneun Kil
P6
2.3
1.8
1
1,628
Worldcup Highlight
5.3
3.89
8
6,056
Jaban Kodeung (Re. )
6.4
4.93
3
2,508
Jayeon-yeseu Baeunda
1.9
1.66
10
896
Junglebook
9.9
4.4
1
4,748
Chernoville Sypnyuen
too
5.5
3.8
1
5,051
Chung-Sun
P10
0.4
0.4
1
2,948
Pll
2.4
3.0
1
2,948
P12
5.4
3.7
1
1,658
P13
8.9
5.3
1
2,190
PI
8.6
6.3
1
7,238
P7
6.8
6.8
1
2,948
P9
1.0
1.0
1
2,948
Chunhyang ah ssi
PI
4.1
2.8
1
5,370
Chunhyang ah ssi
P2
6.4
4.4
1
5,119
Teukjip News Today
PI
2.5
3.5
1
1,534
Teukjip News Today
P2
3.0
3.2
1
1,856
Teukjip Stage of Friendship
2.7
2.5
1
4,264
Hwangya 7
7.8
4.0
1
3,851
19:00 News Line
1.3
1.5
3
4,804
Sakwa GGot Hyanggi
5.07
4.1
3
6,458
Total
137
$479.035

164
Media
Program
Vehicle
Ratings
Message
Ratings
No.
of
Ads.
Cost $
(Per Ad)
SBS-TV
0109SP
2.7
2.7
1
206
0159SP
1.0
1.0
1
206
0629SP
1.55
1.55
2
155
0729SP
2.73
2.73
6
318
12:00 News
3.9
4.5
1
3,968
1334SP
3.5
3.5
1
1,096
1349SP
1.1
1.1
1
1,096
1404SP
1.1
1.1
1
1,096
14:35 News
3.6
6.6
1
4,571
1959SP
1.78
1.78
10
1,881
2029SP
1.74
1.74
11
1,881
20:00 SBS News
3.0
2.08
13
5,644
96 Chung-Sun
0.5
0.7
1
2,809
Keumsoon' GutseYeora
1.4
1.2
1
4,571
Kurium-yen YiYuka
2.7
2.4
1
2,363
News Teukbo PI
2.1
1.7
1
1,421
ManKang (Re. )
9.3
2.3
1
730
Manam (Re. )
3.1
2.4
2
730
Myungsa ChoChung
0.3
2.3
1
2,130
Bodo Teukjip
1.8
1.0
1
5,220
Bom-Chunue Jae 0-si-nae
3.8
0.5
1
1,346
Baduk choikangjun P2
0.9
0.4
1
2,430
Spark Man
0.8
0.5
1
4,155
Sin-to-bul-yi
5.4
5.2
1
3,968
Aadul-eun Haekyulsa
2.0
0.6
1
4,155
Worldcup Kihang
2.2
2.7
2
3,653
Sorim Sa
7.1
4.4
1
3,604
Insang PI
3.1
2.5
1
2,528
Insang P3
6.8
7.2
1
2,524
Junkuk-eul Dalinda
3.12
3.01
22
1,331
Jungmyun-euro Seungboo
5.5
0.7
1
3,604
Chung-Sun Kukmin P10
0.5
0.3
1
2,861
P6
2.9
1.5
1
2,861
P7
1.2
1.0
1
2,861
P8
0.4
0.5
1
2,861
P9
0.3
0.0
1
2,861
Comedy Punch Punch (Re. )
7.1
8.6
1
213
Tomato Dae Sodong
3.0
2.8
1
3,604
Teukjip Kipuen Wuri Toyo-yil
6.9
6.4
1
5,340
Teukjip Kim, Jongseo
1.4
1.0
1
2,738
TV Kayo 20 Nuen
1.2
1.8
1
4,658
Teukjip Joeun Chingu-deul
9.3
4.4
1
3,604
Pokso High School
2.3
1.2
3
4,013
Total
105
$259.441
* Ratings in this table indicate mean vehicle and message ratings during the
three-month campaign period. Monthly mean vehicle and message ratings used
in this study may be different from the ratings described above.

165
Magazines
Media
Name
Vehicle
Ratings
Circulation
No.
of
Ads.
Cost $
(Per Ad)
Weekly
Shisa Journal
6.6
200,000
3
2,250
Magazines
HanKyere 21
4.4
178,000
3
2,000
Weekly Chosun
5.4
180, 000
3
2,250
News Plus
2.8
180,000
3
1,875
Weekly Hanguk
1.6
80,000
3
1,500
Newsweek
2.8
185,000
3
2,250
News People
2.0
155,000
3
1,500
News Maker
1.7
100,000
3
1,500
Weekly Maekyung
1.6
130,000
3
1,750
Economist
1.0
110,000
3
1,750
Hankyung Business
0.5
90,000
3
1,625
Monthly
Shin Dong-Ah
5.3
230,000
3
1,875
Magazines
Monthly Chosun
5.3
180,000
3
1,875
Win
1.2
150,000
3
2,875
Total
42
$80,625

166
aM«W ^ 3H °J °1) 4-i-€ ülcI'íü «{
*¿*1 ^f-
•*] ^ -i:?)-
vfla^la
SST13
xias
4ft
isi aa
1.95
1.95
2
2,113
0.3
0.3
1
2,414
3.5
3.5
1
4,736
2.7
2.7
1
6,314
2.1
2.1
1
6,314
2.0
1.7
1
4,523
2.9
2.9
2
1,811
6.23
6.23
6
3,103
4.57
4.57
6
4,736
0.6
0.5
1
3,971
0.9
1.2
1
4,523
3.1
1.2
1
4,523
3.6
2.2
1
3,158
6.7
4.33
10
6,570
5.9
4.15
2
3,101
2.7
3.5
1
3,956
0.4
0.8
1
1,309
2.2
1.0
1
1,894
4.9
2.7
1
7,988
5.1
1.7
1
4,961
5.5
1.3
1
3,128
4.0
1.75
2
4,523
1.7
1.8
1
5,138
5.8
2.6
1
4,545
4.0
2.5
7
3,101
1.1
1.1
1
4,879
4.5
2.99
8
4,545
5.0
2.2
1
4,523
5.4
6.7
1
4,523
5.1
4.8
1
1,856
3.8
3.3
1
1,463
11.0
2.4
1
4,110
5.5
3.3
10
6,244
7.4
6.0
2
3,021
1.8
1.5
1
4,110
5.8
2.6
1
4,110
6.3
3.47
3
6,053
1.8
1.22
7
1,564
4.9
3.65
2
3,394
2.6
1.4
2
6,570
96
$414,37
â–¡ |C|0|
!sa|
KBS-2TV
0959SP
1159SP
1259SP
1339SP
1354SP
13:00 *fr-
1749SP
2054SP
2249SP
SOS *B°f
nía “
, —:0H
¥« &ohaf ig
=SEfxj y-jsjs (xn) 2_r
“ L A B|S
-af M
2\ _
â–  sm
tí 0
40| 2Â¥
ffeJMci i*
2-
3Â¥
M
±B
Al?
íüotel
f
f
E
E
7
°| □
-£
i 2“
- 3Sf
■°| seiM iV
H ~7' *
’?Í2X|
32) -
2 2E|» gOjEf
s ¿t Cs-9 sena
a« i¥
- as 2Â¥
«}Ü a*i
gg ssa i¥
3Â¥
4Â¥
5Â¥
—ü-
oms asni 2Â¥
tv °rái ui¥7i
29£

167
ssag
AI9#
»a
AIS#
isi as
3.0
3.0
1
1,039
2.85
2.85
2
1,123
3.4
3.4
8
766
4.1
4.1
1
6,314
4.1
4.1
1
6,314
2.31
2.31
6
2,348
5.9
4.4
1
4,691
0.6
0.6
1
4,736
0.95
0.95
8
2,889
6.7
4.89
11
7,684
1.4
0.6
1
3,161
1.9
2.1
1
4,459
1.7
1.8
1
1,376
4.2
2.7
1
2,336
1.5
1.9
1
5,865
3.3
3.7
1
2,599
5.2
1.9
1
4,328
2.7
1.97
21
1,335
1.5
0.5
1
2,535
6.9
5.9
1
8,306
7.0
4.8
1
3,668
1.4
1.2
1
5,951
2.0
2.3
1
5,239
4.0
1.8
1
2,288
8.3
4.2
1
5,269
5.0
2.9
1
5,764
5.8
6.1
1
3,656
4.6
0.5
2
6,056
5.1
4.0
2
6,056
1.7
2.3
1
4,264
0.4
0.6
1
1,406
4.5
3.28
10
4,264
1.7
1.2
1
1,470
2.3
1.8
1
1,628
5.3
3.89
8
6,056
6.4
4.93
3
2,508
1.9
1.66
10
896
9.9
4.4
1
4,748
5.5
3.8
1
5,051
0.4
0.4
1
2,948
2.4
3.0
1
2,948
5.4
3.7
1
1,658
8.9
5.3
1
2,190
8.6
6.3
1
7,238
6.8
6.8
1
2,948
1.0
1.0
1
2,948
4.1
2.8
1
5,370
6.4
4.4
1
5,119
2.5
3.5
1
1,534
3.0
3.2
1
1,856
2.7
2.5
1
4,264
7.8
4.0
1
3,851
1.3
1.5
3
4,804
5.07
4.1
3
6,458
137
$479.03
n|c|01
0559SP
0729SP
0749SP
1019SP
1239SP
1259SP
12:55 Iff—
1459SP
1929SP
21:00 ^ tJiH
24:00 ¥—
MBC ZlSHEI
MBC Iff— =q|o|
7fs0|| #S«
l¥
aq°f°] sj
ais.^ 4ag
^f± ¥d|o| 2¥
' ' “a
9 E||
hija}
S8S
üs. ti
%H
AHi
(Rocky) V
ofe CHS a|SS i¥
Raí
¥ Sf0| gf°A||2
Â¥ Bfu
1
" 2“
■MIS 8JS7|’r
rESS Sl0|Bf0|S
gfas
as S
Â¥Q1
2i 7|Â¥ 1J 3Â¥
* 6 *7*
SEB o|-o afo 5.
ago) (i)
xigo)|Ai unge
Si¥
istJJ a f log 1Â¥
Stí 7fa sé 10¥
' n¥
12Â¥
13 ¥
i¥
7Â¥
9Â¥
eif oi-AAi ss sui i¥
" O t=d
Sa if— ¥c(|o| i¥
" 2Â¥
fa ¥S£| ¥01
sfof°i 7g
19:00 ¥— Bfg
A|2| * If71
971

v* felr tT-fc ¡fí ¡k £;§£ fit filhfc Ib&W ¡jpfi- ‘(n ¡o-§-fer[y
£& tó(firH^S) iQ-ir&l¿ [ofelfe -$'-§-&Ir TT-fk ¡n fiklTlfir //o#/o//3 -j?t *
¿¿
SBS-TV
â–¡
n
3.
JtoJliJliJliJlii Hn HJ 0f«íiKr°ro °n4°°r!-1 >=£n|i WOHrsnMr UW § g g £ £ £ C
I»!*!*!*!» ° s rirüJUoaarfl 1 n iu|n Hü-H^LS: m >ocqm |> §©
smxzi'*1 ° *« “ *Z**
orlOOEoms-arSi ' * » *r°HllttrIHU-^°i—na iln.airwSSHr®-5'^ f.r
&** K jaw * siafer !>
iuw-U3J|iuHo|nrSL riS— — oi> 1; rjyg |>
—
(O0|d(O >H U3 CD -J CD O 0Í p
torfiniH wHHUHt-H ra. v
o
to
Hi
0109SP
0159SP
0629SP
0729SP
12:00 *ff±
1334SP
IN
Hu
U
Dffi
tOlD^h*CD(*)OOOI-*tOOUlWOlW->]tOtOU100Wt-‘OC*)^tOtOh‘OWt-‘l-‘WHh‘WWS3l-kl-*tO
CJC0t0^y3Oh-»C0i^t0VD0UJII-‘C»h-v»-»t0O4^00v£)00 00 C0l-‘Wt-*-0tf^CJlO-0'0CTil-‘t-‘U1V0-0CDO'0
to £»• CD CO (J1
>IH
a»
*â– 
Pi^PHOltOOOOOMMOOWOtOi^tOOUlOOOPtOtONJPWI-'OtOHHOlMI-'CO^tOPPtO
tO^OOOiP»'OOCDOU10ü1í-*f-» CO ^ CO CO (J1
02 u
mow
I-*
-“POltOPP
>02
**
hi rf* co to cn co tototototocoH‘totococo^co^tot--aito p» to *». to cn t-* i-» ** p* p* p* co
oa\a\'0(ocDCococococococj>co(ji(jicriOM-‘ví3h-‘^>.cotoi-i--j-o^coaioocricoco(jiooovcicoi-ktoto
■ P>OCnCO£*OP‘G>CyiCDCDOCOtOtOOCJlUlCDUlCO.r*tOCOCOCOtOC7i-00**C»(X>-OVOvOCOa>P‘CJIOO
CO.f*OOOOOif*COP‘P‘P‘P‘P‘.^P‘.f*CO.F»'COCn P*
JOH
121
Q£
_cr
0(0

169
â–¡|c|o)
«Í3¥¥
5¡¥
is| his
AIA1- *|U
6.6
200,000
3
2,250
3*1
&7|Ej| 21
4.4
178,000
3
2,000
5.4
180, 000
3
2,250
Sf± S&
2.8
180,000
3
1,875
2£?y «13
1.6
80,000
3
1,500
tt— ¥|a
2.8
185,000
3
2,250
TT^ U S
2.0
155,000
3
1,500
m|o
5H
1.7
100,000
3
1,500
Dll?
1.6
130, 000
3
1,750
oiat â–¡
W E=
1.0
110,000
3
1,750
eg u|x
M—
0.5
90,000
3
1,625
sm
tí sof
5.3
230,000
3
1,875
§*i
§>J 71 ^Aj
5.3
180,000
3
1,875
a
1.2
150,000
3
2,875
42
$80,625

APPENDIX D
GENERAL FORM OF THE BBMD MODEL

171
General Form of the BBMD Model
Individual Vehicle Exposure Distribution
vi. fi=o n (*>+fi )/(“■+ +fi) w
Í-0
v,o Where:
f,
n.
a¡
b,
R,
probability of exposure f . times to vehicle i,
number of exposures to vehicle /, where
f, = 0 to n, total insertions,
total insertions in vehicle /,
vehicle i exposure parameter
)/[2fl,
vehicle i non-exposure parameter = a (\ - R,)/r, ,
reach as a percentage of the target audience of
one typical insertion in vehicle / (one time
rating),
reach as a percentage of the target audience of
two typical insertions in vehicle i (self-pair or two-use
reach),
total combinations of n, vehicle insertions taken
/, at a time.
Source: adapted from Kent M. Lancaster, “ADplus: For Multi-media Advertising Planning” (1993).

172
Multiple Vehicle Exposure Distribution
>.fi=o
-1 +d
y j.fj=o
v N
Vr E Wujd
/i+/j=°
v F* 0 = ^p=o+1
Where v(Ji and/, are as defined above, and where:
W
(5)
(6)
v U probability of exposure/, times to one or more
vehicles in the media category schedule,
VF - probability of exposure F times to the vehicles
in the plan,
F = number of exposures where F=0 to N total
insertions in the schedule,
N = total insertions in the schedule = £ „ , where
1 = 1
n, is as defined above and where m is the number
of vehicles in the plan,
d = average relation to chance of cross-pair reach
and duplication:
I E +*,*;)/((2ÍR, +RJ)-R,RJ -*,)]

173
Individual Vehicle Message Exposure Distribution
The procedures for estimating individual vehicle message exposure distributions
are the same as those for vehicles as described in equations one through six above,
with the following substitutions.
R = pRl
(7)
R2l (8)
subject to the following constraint:
Vi,/=0,2íií500
where the variables on the right hand side of the equations pertain to vehicles and
those on the left pertain to messages, and where:
p = the average probability of exposure to an
advertising message in a schedule, given exposure
to a vehicle.

174
Combined Media Category
Vehicle or Message Exposure Distributions
cF = t vFvPj
F, * Fj = 0
Where:
V, F and N are as defined above, and where:
C = probability of exposure F times to the combined
vehicle or message exposure distribution of a
pair of separate and/or combined exposure
distributions / and j
Self- and Cross-pair Estimation Equations for Network TV in the US
R2 = -0.0065 + 1.795R,- 0.331 (R,)2
Rt] - -0.0015 + 1.0095 (« + Rj) - 1.5R, R,
Source: adaptedfrom Kent M. Lancaster, “ADplus: For Multi-Media Advertising
Planning" (1993).

175
Self- and Cross-pair Estimation Equations for Network TV in South Korea
R1: Estimation Equation
Target Audience
Estimation Equation
Total (Individuals)
RJ( =0.00039+ 1 82212 (R,) - 1.17271 (r,)2
Male, 20 - 34
R2i =0.00056+ 1 82834 (r,) - 1.57613(/Í,)2
Male, 35-49
R2i =0.00049+ 1.86831 (/?,) - 1.28784(i?,)2
Male, 20 - 49
R2 = 0.000525 + 1 848325 («,) - 1.432 (t?,)2
Male, Total
R2, =0.00075 + 1.80324 (r>) - 1.17273(t?,)2
R:J Estimation Equation
Target Audience
Estimation Equation
Total (Individuals)
RtJ =0.0000+ 1.0022 (fl, +fij) - 1.7925(7?, xR,)~ 0.0011
Male, 20-34
Rt¡ =0.0001 + 1.0027 («, + /?,) - 2.0973 (/^, x7?;)-0.0008
Male, 35-49
Ri] =0.0000+ 1.0033 (/?, +7?y) - 1.6578(/?, xTiJ-O.OOlO
Male, 20 - 49
Rv = 0.00005 + 1.003 (r, + 7?,)- 1.87755(t?, xT?,)- 0.0009
Male, Total
Rl¡ =0.0001 + 1.0027 (t?, + /?,) - 1.8857(t?, x 7?,)-0.0014
Source: adaptedfrom ‘Korean Gallup' (¡992).

APPENDIX E
ADPLUS RESULTS FOR NETWORK TELEVISION IN FEBRUARY

177
ADplus(TM) RESULTS: NETWORK TV
Hyunsoo Park
University of Florida
FEB. 17 ~ MAR. 16, 1996
Frequency (f) Distributions
Vehicle
Message
Target: 2,884,983 f
Korean men 20-49 years
old who live in Seoul. 0
1
Message/vehicle = 100.0% 2
3
4
5
6
7
8
9
10+
20+
Summary Evaluation
Reach 1+ (%)
Reach 1+ (000s)
Reach 3+ (%)
Reach 3+ (000s)
Gross rating points (GRPs)
Gross impressions (000s)
Average frequency (f)
Cost-per-thousand (CPM)
Cost-per-rating point (CPP)
Cost-per-net reach point (CPRP)
Cost-per-response (CPR)
Vehicle List
Rating
Ad Cost
0959SP KBS
1.95
2,113
1259SP
3.50
4,736
1339SP
2.70
6,314
13:00 News
1.70
4,523
2249SP
4.57
4,736
SOS Haeyang
0.50
3, 971
Dawon Bok 1
1.20
4,523
Dawon Bok 2
1.20
4,523
Mini Series
4.33
6,570
Bamkwa 2
4.15
3,101
Bob Roberts
3.50
3, 956
Sulnal 1
0.80
1, 309
Sulnal 2
1.00
1, 894
Kum 2
2.70
7,988
Kum 3
1.70
4, 961
Saekye 2
1.30
3,128
Sagi Masul
1.75
4,523
Star wa Go
1.80
5,138
Sigan 1
2.60
4,545
Simya
2.50
3,101
% f
% f+
% f
% f+
5.5
100.0
5.5
100.0
11.6
94.5
11.6
94.5
18.0
82.9
18.0
82.9
19.5
64.9
19.5
64.9
16.7
45.4
16.7
45.4
12.1
28.7
12.1
28.7
7.7
16.6
7.7
16.6
4.4
8.9
4.4
8.9
2.3
4.5
2.3
4.5
1.2
2.1
1.2
2.1
1.0
1.0
1.0
1.0
0.0
0.0
0.0
0.0
94.5%
94.5%
2,
725.1
2,
725.1
64.9%
64.9%
1,
872.2
1,
872.2
350.1
350.1
10,
101.5
10,
101.5
3.7
3.7
46.59
46.59
1,
344
1,
344
4,
983
4,
983
0.17
0.17
CPM-MSG
Ads
Total Cost
Mix %
37.56
2
4,226
0.9
46.90
1
4,736
1.0
81.06
1
6,314
1.3
92.22
1
4,523
1.0
35.92
4
18,944
4.0
275.29
1
3, 971
0.8
130.65
1
4,523
1.0
130.65
1
4,523
1.0
52.59
4
26,280
5.6
25.90
2
6,202
1.3
39.18
1
3, 956
0.8
56.72
1
1, 309
0.3
65.65
1
1, 894
0.4
102.55
1
7, 988
1.7
101.15
1
4,961
1.1
83.40
1
3,128
0.7
89.59
2
9, 046
1.9
98.94
1
5,138
1.1
60.59
1
4,545
1.0
43.00
2
6,202
1.3

178
Jae 3 Noon
2.99
4,545
Queeze Norae
2.20
4,523
Queeze Myung
6.70
4,523
Pankwan 1
3.30
6,244
Pankwan 4
1.50
4,110
Pakwan 5
2.60
4, 110
MBC Grand
2.10
4,459
21:00 News De
4.89
7,684
Rocky V
5.90
8,306
Kajok 1
1.90
5, 865
TEUK WOOJUNG
2.50
4,264
BLUE
3.70
2,599
ONGOJIP
0.60
1,406
MBC NEWS TODA
1.80
1,376
1019SP
4.10
6, 314
CHUNHANG 1
2.80
5, 370
CHUNHANG 2
4.40
5,119
SEOCHO 1
2.90
5,764
SEOCHO 2
6.10
3,656
HWANG YA 7
4.00
3, 851
HONGKILDONG
1.90
4,328
12:55 NEWS
4.40
4,691
BEST
2.30
4,264
1459SP
0.60
4,736
JUNGLEBOOK
4.40
4,748
SEHAE BOK
4.20
5,269
NEWS TODAY 2
1.97
1,335
0749SP
3.40
766
WORLDCUP 3
1.20
1,470
WORLDCUP 6
1.80
1,628
SAESANG
0.50
6, 056
WOOJUNG
3.28
4,264
1239SP
4.10
6,314
24:00 NEWS
0.60
3,161
1929SP
0.95
2,889
WORLDCUP HIGH
3.89
6, 056
JAYEON
1.66
896
TEUK GI SBS
6.40
5,340
TEUK KIM
1.00
2,738
TEUK SANG
1.80
4,658
TEUK JO
4.40
3,604
BOM
0.50
1,346
SHINTO 2
5.20
3, 968
12:00 NEWS
4.50
3, 968
SORIMSA
4.40
3, 604
INSANG 1
2.50
2,528
INSANG 3
7.20
2,524
1404SP
1.10
1,096
KUT
1.20
4,571
GURIUM
2.40
2,363
NEWS TUEKBO 1
1.70
1, 421
JUNGUK
3.01
1,331
1334SP
3.50
1, 096
20:00 SBS NEW
2.08
5, 644
0109SP
2.70
206
1959SP
1.78
1, 881
2029SP
1.74
1, 881
BODO
1.00
5,220
Totals:
.69
1
4,545
1.0
.26
1
4,523
1.0
.40
1
4,523
1.0
.59
4
24,976
5.3
.97
1
4,110
0.9
.79
1
4,110
0.9
. 60
1
4,459
0.9
.47
5
38,420
8.2
. 80
1
8,306
1.8
.00
1
5, 865
1.2
.12
1
4,264
0.9
.35
1
2,599
0.6
.23
1
1, 406
0.3
.50
1
1,376
0.3
.38
1
6,314
1.3
.48
1
5,370
1.1
.33
1
5,119
1.1
.89
1
5,764
1.2
.77
1
3,656
0.8
.37
1
3,851
0.8
. 96
1
4,328
0.9
.95
1
4,691
1.0
.26
1
4,264
0.9
.60
1
4,736
1.0
. 40
1
4,748
1.0
.48
1
5,269
1.1
.49
10
13,350
2.8
. 81
4
3, 064
0.7
.46
1
1, 470
0.3
.35
1
1, 628
0.3
.83
2
12,112
2.6
. 06
3
12,792
2.7
.38
1
6, 314
1.3
.61
1
3,161
0.7
. 41
2
5,778
1.2
.96
2
12,112
2.6
.71
1
896
0.2
. 92
1
5, 340
1.1
.91
1
2,738
0.6
.70
1
4,658
1.0
.39
1
3,604
0.8
.31
1
1, 346
0.3
.45
1
3, 968
0.8
.56
1
3, 968
0.8
.39
1
3,604
0.8
. 05
1
2,528
0.5
.15
1
2,524
0.5
.54
1
1, 096
0.2
.03
1
4,571
1.0
.13
1
2,363
0.5
.97
1
1, 421
0.3
.33
6
7,986
1.7
.85
1
1, 096
0.2
.05
4
22,576
4.8
.64
1
206
0.0
.63
4
7,524
1.6
.47
3
5, 643
1.2
.94
1
5,220
1.1
.59
125
470,658
100.0
52
71
23
65
94
54
7 3
54
48
107
59
24
81
2 6
53
66
40
68
2 0
33
78
36
64
273
37
43
23
7
42
31
419
4 5
53
182
105
53
18
28
94
89
28
93
26
30
20
35
12
34
132
34
28
15
10
94
2
36
37
180
46.

APPENDIX F
ADPLUS RESULTS FOR NETWORK TELEVISION IN MARCH

ADplus(TM) RESULTS: NETWORK TV
Hyunsoo Park Frequency (f) Distributions
University of Florida
MAR. 17 ~ APR. 16, 1996
Vehicle
Message
Target: 2,884,983
f
% f
% f+
% f
% f+
Korean men 20-49 years
—
—
—
—
—
old who live in Seoul.
0
5.8
100.0
5.8
100.0
1
11.5
94.2
11.5
94.2
Message/vehicle = 100.0%
2
17.3
82.7
17.3
82.7
3
18.7
65.4
18.7
65.4
4
16.2
46.7
16.2
46.7
5
12.1
30.6
12.1
30.6
6
8.0
18.5
8.0
18.5
7
4.9
10.5
4.9
10.5
8
2.7
5.7
2.7
5.7
9
1.5
2.9
1.5
2.9
10+
1.4
1.4
1.4
1.4
20+
0.0
o
o
0.0
0.0
Summary Evaluation
Reach 1+ (%)
94.2%
94.2%
Reach 1+ (000s)
2,718.5
2,
718.5
Reach 3+ (%)
65.4%
65.4%
Reach 3+ (000s)
1,887.1
1,
887.1
Gross rating
points (GRPs)
359.9
359.9
Gross impressions (000s)
10,384.5
10,
384.5
Average frequency (f)
3.8
3.8
Cost-per-thousand (CPM)
38.35
38.35
Cost-per-rating
point (CPP)
1,106
1,
106
Cost-per-net
reach point
(CPRP)
4,226
4,
226
Cost-per-response (CPR)
0.15
0.15
Vehicle List
Rating Ad Cost
CPM-MSG Ads
Total Cost
Mix %
JAE 3 NOON
2.99
4,545
52.69
5
22,725
5.7
2249SP
4.57
4,736
35.92
2
9,472
2.4
MINI SERIES
4.33
6,570
52.59
4
26,280
6.6
SHIMYA
2.50
3,101
43.00
3
9, 303
2.3
PANKWAN 1
3.30
6,244
65.59
4
24,976
6.3
1749SP
2.90
1, 811
21.65
2
3, 622
0.9
PAKWAN 3
6.00
3,021
17.45
1
3,021
0.8
1354SP
2.10
6,314
104.22
1
6, 314
1.6
WURI
1.10
4,879
153.74
1
4,879
1.2
2054SP
6.23
3,103
17.26
2
6,206
1.6
1159SP
0.30
2,414
278.92
1
2,414
0.6
TEUKSUN
2.40
4,110
59.36
1
4,110
1.0
TOWERING 1
4.80
1, 856
13.40
1
1,856
0.5
TOWERING 2
3.30
1,463
15.37
1
1,463
0.4
WUJUNG MBC
3.28
4,264
45.06
5
21,320
5.4
JAYEON
1.66
896
18.71
9
8,064
2.0
0749SP
3.40
766
7.81
4
3, 064
0.8
WORLDCUP
3.89
6, 056
53.96
4
24,224
6.1
JABAN
4.93
2,508
17.63
3
7,524
1.9
21:00 NEWS DE
4.89
7, 684
54.47
4
30,736
7.7

181
1929SP
0.95
2, 889
BAKSANG
1.20
5, 951
KASEUM
2.70
2,336
BULHANG
2.30
5,239
1259SP
2.31
2,348
BBONGBAT
1.80
2,288
NEWS TODAY 2
1.97
1,335
TEUK NEWSTODA
3.50
1,534
TEUK NEWSTODA
3.20
1, 856
MYUNGTAE 2
4.80
3, 668
CHONGSUN 1
6.30
7,238
CHONG 7
6.80
2, 948
9
1.00
2,948
10
0.40
2, 948
11
3.00
2,948
12
3.70
1,658
13
5.30
2,190
0559SP
3.00
1,039
NEWS TEUKBO
0.50
2,535
JUNKUK
3.01
1,331
20:00 SBS NEW
2.08
5, 644
1959SP
1.78
1, 881
2029SP
1.74
1, 881
WORLDCUP KI
2.70
3,653
SPARKMAN
0.50
4,155
96 CHONG
0.70
2, 809
CHUNG 6
1.50
2,861
7
1.00
2, 861
8
0.50
2,861
CHUNG 9
0.01
2,861 9
CHUNG 10
0.30
2,861
COMEDY
8.60
213
TOMATO
2.80
3,604
MANAM
2.75
730
1349SP
1.10
1,096
0729SP
2.73
318
MANGANG
2.30
730
MYUNGSA
2.30
2,130
BADUK
0.40
2,430
JUNGMYUN
0.70
3,604
14:35 NEWS
6.60
4,571
AHDUL
0.60
4,155
0159SP
1.00
206
Totals:
.41
3
8,667
2.2
.90
1
5, 951
1.5
. 99
1
2,336
0.6
.95
1
5,239
1.3
.23
2
4,696
1.2
.06
1
2,288
0.6
.49
3
4,005
1.0
.19
1
1,534
0.4
.10
1
1, 856
0.5
.49
1
3, 668
0.9
.82
1
7,238
1.8
.03
1
2, 948
0.7
.18
1
2, 948
0.7
.46
1
2, 948
0.7
.06
1
2, 948
0.7
.53
1
1, 658
0.4
.32
1
2,190
0.5
.00
1
1,039
0.3
.74
1
2,535
0.6
.33
10
13,310
3.3
.05
5
28,220
7.1
.63
4
7,524
1.9
.47
4
7,524
1.9
.90
2
7,306
1.8
.04
1
4,155
1.0
.09
1
2,809
0.7
. 11
1
2,861
0.7
.17
1
2, 861
0.7
.34
1
2,861
0.7
.87
1
2, 861
0.7
.56
1
2, 861
0.7
.86
1
213
0.1
.62
1
3,604
0.9
.20
2
1,460
0.4
.54
1
1,096
0.3
.04
2
636
0.2
.00
1
730
0.2
.10
1
2,130
0.5
.57
1
2,430
0.6
.46
1
3, 604
0.9
. 01
1
4,571
1.1
.04
1
4,155
1.0
.14
1
206
0.1
.35
128
398,253
100.0
105
171
29
78
35
44
2 3
15
20
26
39
15
102
255
34
15
14
12
175
15
94
36
37
46
288
139
66
99
198
916
330
0
44
9
34
4
11
32
210
178
24
240
7
38

APPENDIX G
ADPLUS RESULTS FOR NETWORK TELEVISION IN APRIL

183
ADplus(TM) RESULTS: NETWORK TV
Hyunsoo Park
University of Florida
APR. 17 ~ MAY 16, 1996
Frequency (f) Distributions
Vehicle
Message
Target: 2,884,983 f
Korean men 20-49 years
old who live in Seoul. 0
1
Message/vehicle = 100.0% 2
3
4
5
6
7
8
9
10+
20+
Summary Evaluation
Reach 1+ (%)
Reach 1+ (000s)
Reach 3+ (%)
Reach 3+ (000s)
Gross rating points (GRPs)
Gross impressions (000s)
Average frequency (f)
Cost-per-thousand (CPM)
Cost-per-rating point (CPP)
Cost-per-net reach point (CPRP)
Cost-per-response (CPR)
Vehicle List
Rating
Ad Cost
MINI SERIES
4.33
6,570
SHIMYA
2.50
3,101
PANKWAN 1
3.30
6,244
PANKWANG 3
6.00
3, 021
2054SP
6.23
3,103
JAE 3 NOON
2.99
4,545
MOKYOKTANG
2.20
3,158
AHCHIM 2
1.22
1,564
SPORTS
3.47
6, 053
TV INSANG
3.65
3, 394
JOKWANGJO?
1.40
6,570
NEWSTODAY 2 M
1.97
1,335
WORLDCUP HIGH
3.89
6,056
21:00 NEWS DE
4.89
7, 684
1259SP
2.31
2,348
WUJUNG
3.28
4,264
1929SP
0.95
2,889
CHERUNOVILLE
3.80
5,051
SPORTSHIGHLIG
4.00
6,056
19:00 NEWS LI
1.50
4,804
% f
% f+
% f
% f+
14.8
100.0
14.8
100.0
23.0
85.2
23.0
85.2
23.0
62.3
23.0
62.3
17.3
39.3
17.3
39.3
10.8
22.0
10.8
22.0
5.9
11.2
5.9
11.2
3.0
5.3
3.0
5.3
1.4
2.3
1.4
2.3
0.6
1.0
0.6
1.0
0.2
0.4
0.2
0.4
0.1
0.1
0.1
0.1
0.0
0.0
0.0
0.0
85.2%
85.2%
2,
458.5
2,
458.5
39.3%
39.3%
1,
133.7
1,
133.7
229.2
229.2
6,
612.4
6,
612.4
2.7
2.7
42.64
42.64
1,
230
1,
230
3,
309
3,
309
0.11
0.11
CPM-MSG
Ads
Total Cost
Mix %
52.59
2
13,140
4.7
43.00
2
6,202
2.2
65.59
2
12,488
4.4
17.45
1
3, 021
1.1
17.26
4
12,412
4.4
52.69
2
9,090
3.2
49.76
1
3,158
1.1
44.44
7
10,948
3.9
60.46
3
18,159
6.4
32.23
2
6,788
2.4
162.66
2
13,140
4.7
23.49
8
10,680
3.8
53.96
2
12,112
4.3
54.47
2
15,368
5.5
35.23
4
9, 392
3.3
45.06
2
8,528
3.0
105.41
3
8, 667
3.1
46.07
1
5, 051
1.8
52.48
2
12,112
4.3
111.01
3
14,412
5.1

184
SAKWA
4.10
6,458
54.60
3
19,374
6.9
0729SP
2.85
1,123
13.66
2
2,246
0.8
1959SP SBS
1.78
1,881
36.63
2
3,762
1.3
2029SP
1.74
1,881
37.47
4
7, 524
2.7
0729SP SBS
2.73
318
4.04
4
1,272
0.5
JUNKUK
3.01
1,331
15.33
6
7, 986
2.8
20:00 SBS NEW
2.08
5, 644
94.05
4
22,576
8.0
POKSO
1.20
4,013
115.92
3
12,039
4.3
0629SP
1.55
155
3.47
2
310
0.1
Totals:
42.64
85
281,957
100.0

APPENDIX H
ADPLUS RESULTS FOR MAGAZINE ADVERTISEMENTS
(February - April: Schedules are identical)

186
ADplus(TM) RESULTS: MAGAZINES
Hyunsoo Park
University of Florida
ONE MONTH, 1996
Target: 2,884,983 f
Korean men 20-49 years
old who live in Seoul. 0
1
Message/vehicle =35.0% 2
3
4
5
6
7
8
9
10+
Summary Evaluation
Reach 1+ (%)
Reach 1+ (000s)
Reach 3+ (%)
Reach 3+ (000s)
Gross rating points (GRPs)
Gross impressions (000s)
Average frequency (f)
Cost-per-thousand (CPM)
Cost-per-rating point (CPP)
Cost-per-net reach point (CPRP)
Cost-per-response (CPR)
Vehicle List
Rating
Ad Cost
Sisa Journal
6.60
2,250
Han Kyurei 21
4.40
2,000
Jugan Chosun
5.40
2,250
News Plus
2.80
1,875
Jugan Hankuk
1.60
1,500
News Week
2.80
2,250
News people
2.00
1,500
News Maker
1.70
1, 500
Jugan MaeKyun
1.60
1,750
Economist
1.00
1,750
HanKyung Busi
0.50
1, 625
Shin Dong-Ah
5.30
1, 875
WalKan Chosun
5.30
1, 875
Win
1.20
2,875
Totals:
Frequency (f) Distributions
Vehicle Message
% f
% f+
% f
% f+
65.1
100.0
86.2
100.0
28.4
34.9
12.9
13.8
5.8
6.5
0.9
0.9
0.7
0.8
0.0
0.0
0.1
0.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
34.9%
13.8%
1,
006.1
398.4
0.8%
0.0%
21.7
1.1
42.2
14.8
1,
217.5
426.1
1.2
1.1
22.07
63.07
637
1,
820
771
1,
946
0.03
0.07
CPM-MSG
Ads
Total Cost
Mix %
33.76
1
2,250
8.4
45.02
1
2,000
7.4
41.26
1
2,250
8.4
66.32
1
1, 875
7.0
92.85
1
1,500
5.6
79.58
1
2,250
8.4
74.28
1
1,500
5.6
87.38
1
1,500
5.6
108.32
1
1,750
6.5
173.31
1
1,750
6.5
321.86
1
1, 625
6.0
35.04
1
1, 875
7.0
35.04
1
1, 875
7.0
237.27
1
2,875
10.7
63.07
14
26,875
100.0

APPENDIX I
ADPLUS RESULTS FOR NETWORK TELEVISION AND MAGAZINES
COMBINED IN FEBRUARY

188
ADplus(TM) RESULTS: NETWORK TV, MAGAZINES
Hyunsoo Park
University of Florida
FEB. 17 ~ MAR. 16, 1996
Target: 2,884,983 f
Korean men 20-49 years
old who live in Seoul. 0
1
Message/vehicle =93.0% 2
3
4
5
6
7
8
9
10+
20+
Summary Evaluation
Reach 1+ (%)
Reach 1+ (000s)
Reach 3+ (%)
Reach 3+ (000s)
Gross rating points (GRPs)
Gross impressions (000s)
Average frequency (f)
Cost-per-thousand (CPM)
Cost-per-rating point (CPP)
Cost-per-net reach point (CPRP)
Cost-per-response (CPR)
Vehicle List
1 NETWORK TV
Rating
Ad Cost
Totals:
0959SP KBS
1.95
2,113
1259SP
3.50
4,736
1339SP
2.70
6,314
13:00 News
1.70
4,523
2249SP
4.57
4,736
SOS Haeyang
0.50
3,971
Dawon Bok 1
1.20
4,523
Dawon Bok 2
1.20
4,523
Mini Series
4.33
6,570
Bamkwa 2
4.15
3,101
Bob Roberts
3.50
3, 956
Sulnal 1
0.80
1,309
Sulnal 2
1.00
1, 894
Kum 2
2.70
7, 988
Kum 3
1.70
4,961
Saekye 2
1.30
3,128
Sagi Masul
1.75
4,523
Star wa Go
1.80
5,138
Frequency (f) Distributions
Vehicle Message
% f
% f+
% f
% f+
3.6
100.0
4.8
100.0
9.1
96.4
10.7
95.2
15.3
87.3
17.0
84.5
18.5
72.0
19.2
67.5
17.5
53.5
17.1
48.3
13.9
35.9
12.7
31.2
9.5
22.1
8.3
18.4
5.9
12.5
4.9
10.1
3.3
6.6
2.7
5.2
1.7
3.3
1.3
2.5
1.6
1.6
1.2
1.2
0.0
0.0
0.0
0.0
96.4%
95.2%
2,
891.7
2,
856.7
72.0%
67.5%
2,
158.8
2,
024.6
392.3
364.9
11,
770.2
10,
947.3
4.1
3.8
42.27
45.45
1,
268
1,
363
5,
162
5,
225
0.17
0.17
CPM-MSG
Ads
Total Cost
Mix %
44.81
125
470,658
94.6
36.12
2
4,226
0.8
45.10
1
4,736
1.0
77.95
1
6,314
1.3
88.69
1
4,523
0.9
34.54
4
18,944
3.8
264.73
1
3, 971
0.8
125.64
1
4,523
0.9
125.64
1
4,523
0.9
50.58
4
26,280
5.3
24.91
2
6,202
1.2
37.68
1
3, 956
0.8
54.54
1
1, 309
0.3
63.13
1
1,894
0.4
98.62
1
7,988
1.6
97.27
1
4,961
1.0
80.21
1
3,128
0.6
86.15
2
9,046
1.8
95.15
1
5,138
1.0

189
Sigan 1
Simya
Jae 3 Noon
Queeze Norae
Queeze Myung
Pankwan 1
Pankwan 4
Pakwan 5
MBC Grand
21:00 News De
Rocky V
Kajok 1
TEUK WOOJUNG
BLUE
ONGOJIP
MBC NEWS TODA
1019SP
CHUNHANG 1
CHUNHANG 2
SEOCHO 1
SEOCHO 2
HWANG YA 7
HONGKILDONG
12:55 NEWS
BEST
1459SP
JUNGLEBOOK
SEHAE BOK
NEWS TODAY 2
0749SP
WORLDCUP 3
WORLDCUP 6
SAESANG
WOOJUNG
1239SP
24:00 NEWS
1929SP
WORLDCUP HIGH
JAYEON
TEUK GI SBS
TEUK KIM
TEUK SANG
TEUK JO
BOM
SHINTO 2
12:00 NEWS
SORIMSA
INSANG 1
INSANG 3
1404SP
KUT
GURIUM
NEWS TUEKBO 1
JUNGUK
1334SP
20:00 SBS NEW
0109SP
1959SP
2029SP
BODO
2.60
4,545
2.50
3,101
2.99
4,545
2.20
4,523
6.70
4,523
3.30
6,244
1.50
4,110
2.60
4,110
2.10
4,459
4.89
7, 684
5.90
8,306
1.90
5, 865
2.50
4,264
3.70
2,599
0.60
1,406
1.80
1, 376
4.10
6,314
2.80
5,370
4.40
5,119
2.90
5,764
6.10
3, 656
4.00
3,851
1.90
4,328
4.40
4,691
2.30
4,264
0.60
4,736
4.40
4,748
4.20
5,269
1.97
1,335
3.40
766
1.20
1,470
1.80
1,628
0.50
6, 056
3.28
4,264
4.10
6,314
0.60
3,161
0.95
2, 889
3.89
6,056
1.66
896
6.40
5,340
1.00
2,738
1.80
4,658
4.40
3, 604
0.50
1, 346
5.20
3, 968
4.50
3,968
4.40
3,604
2.50
2,528
7.20
2,524
1.10
1,096
1.20
4,571
2.40
2,363
1.70
1,421
3.01
1,331
3.50
1,096
2.08
5, 644
2.70
206
1.78
1,881
1.74
1, 881
1.00
5,220
58.27 1
41.35 2
50.67 1
68.53 1
22.50 1
63.07 4
91.33 1
52.69 1
70.78 1
52.38 5
46.93 1
102.89 1
56.85 1
23.41 1
78.11 1
25.48 1
51.33 1
63.93 1
38.78 1
66.25 1
19.98 1
32.09 1
75.93 1
35.54 1
61.80 1
263.11 1
35.97 1
41.82 1
22.59 10
7.51 4
40.83 1
30.15 1
403.73 2
43.33 3
51.33 1
175.61 1
101.37 2
51.89 2
17.99 1
27.81 1
91.27 1
86.26 1
27.30 1
89.73 1
25.44 1
29.39 1
27.30 1
33.71 1
11.69 1
33.21 1
126.97 1
32.82 1
27.86 1
14.74 6
10.44 1
90.45 4
2.54 1
35.22 4
36.03 3
174.00 1
4,545
0.9
6,202
1.2
4,545
0.9
4,523
0.9
4,523
0.9
24,976
5.0
4,110
0.8
4,110
0.8
4,459
0.9
38,420
7.7
8,306
1.7
5, 865
1.2
4,264
0.9
2,599
0.5
1,406
0.3
1,376
0.3
6,314
1.3
5, 370
1.1
5,119
1.0
5,764
1.2
3, 656
0.7
3, 851
0.8
4,328
0.9
4, 691
0.9
4,264
0.9
4,736
1.0
4,748
1.0
5,269
1.1
13,350
2.7
3, 064
0.6
1,470
0.3
1, 628
0.3
12,112
2.4
12,792
2.6
6,314
1.3
3,161
0.6
5,778
1.2
12,112
2.4
896
0.2
5,340
1.1
2,738
0.6
4,658
0.9
3, 604
0.7
1,346
0.3
3,968
0.8
3,968
0.8
3, 604
0.7
2,528
0.5
2,524
0.5
1, 096
0.2
4,571
0.9
2,363
0.5
1, 421
0.3
7, 986
1.6
1, 096
0.2
22,576
4.5
206
0.0
7,524
1.5
5, 643
1.1
5,220
1.0

190
2 MAGAZINES Totals:
Sisa Journal
6.60
2,250
Han Kyurei 21
4.40
2, 000
Jugan Chosun
5.40
2,250
News Plus
2.80
1,875
Jugan Hankuk
1.60
1,500
News Week
2.80
2,250
News people
2.00
1,500
News Maker
1.70
1,500
Jugan MaeKyun
1.60
1,750
Economist
1.00
1,750
HanKyung Busi
0.50
1,625
Shin Dong-Ah
5.30
1,875
WalKan Chosun
5.30
1, 875
Win
1.20
2,875
Totals:
, 65
14
26,875
5.4
,47
1
2,250
0.5
,29
1
2,000
0.4
, 68
1
2,250
0.5
,78
1
1, 875
0.4
,29
1
1,500
0.3
53
1
2,250
0.5
,43
1
1,500
0.3
, 03
1
1,500
0.3
,17
1
1,750
0.4
, 67
1
1,750
0.4
,52
1
1, 625
0.3
,69
1
1, 875
0.4
, 69
1
1,875
0.4
,17
1
2,875
0.6
,45
139
497,533
100.0
60
32
43
39
63
89
76
71
84
104
166
309
33
33
228
45

APPENDIX J
ADPLUS RESULTS FOR NETWORK TELEVISION AND MAGAZINES
COMBINED IN MARCH

192
ADplus(TM) RESULTS: NETWORK TV, MAGAZINES
Hyunsoo Park
University of Florida
MAR. 17 ~ APR. 16, 1996
Target: 2,884,983 f
Korean men 20-49 years
old who live in Seoul. 0
1
Message/vehicle =93.2% 2
3
4
5
6
7
8
9
10+
20+
Summary Evaluation
Reach 1+ (%)
Reach 1+ (000s)
Reach 3+ (%)
Reach 3+ (000s)
Gross rating points (GRPs)
Gross impressions (000s)
Average frequency (f)
Cost-per-thousand (CPM)
Cost-per-rating point (CPP)
Cost-per-net reach point (CPRP)
Cost-per-response (CPR)
Vehicle List
Rating
Ad Cost
1 NETWORK TV
Totals:
JAE 3 NOON
2.99
4,545
2249SP
4.57
4,736
MINI SERIES
4.33
6,570
SHIMYA
2.50
3,101
PANKWAN 1
3.30
6,244
1749SP
2.90
1, 811
PAKWAN 3
6.00
3, 021
1354SP
2.10
6,314
WURI
1.10
4,879
2054SP
6.23
3,103
1159SP
0.30
2,414
TEUKSUN
2.40
4,110
TOWERING 1
4.80
1, 856
TOWERING 2
3.30
1,463
WUJUNG MBC
3.28
4,264
JAYEON
1.66
896
0749SP
3.40
766
Frequency (f) Distributions
Vehicle Message
% f
% f+
% f
% f+
3.8
100.0
5.0
100.0
9.1
96.2
10.7
95.0
14.9
87.1
16.5
84.4
17.8
72.2
18.4
67.9
16.9
54.5
16.5
49.5
13.6
37.6
12.6
33.0
9.7
23.9
8.6
20.3
6.2
14.2
5.3
11.7
3.7
8.0
3.1
6.4
2.1
4.3
1.7
3.3
2.2
2.2
1.7
1.7
0.0
0.0
0.0
0.0
96.2%
95.0%
2,887.3
2,850.8
72.2%
67.9%
2,167.3
2,037.3
402.2
374.7
12,064.5
11,241.6
4.2
3.9
35.24
37.82
1, 057
1,135
4,417
4,474
0.15
0.15
CPM-MSG Ads
Total Cost Mix %
36.88
128
398,253
93.7
50.67
5
22,725
5.3
34.54
2
9,472
2.2
50.58
4
26,280
6.2
41.35
3
9,303
2.2
63.07
4
24,976
5.9
20.82
2
3, 622
0.9
16.78
1
3,021
0.7
100.22
1
6,314
1.5
147.85
1
4,879
1.1
16.60
2
6,206
1.5
268.22
1
2,414
0.6
57.08
1
4,110
1.0
12.89
1
1, 856
0.4
14.78
1
1,463
0.3
43.33
5
21,320
5.0
17.99
9
8,064
1.9
7.51
4
3,064
0.7

193
WORLDCUP
3.89
6, 056
JABAN
4.93
2,508
21:00 NEWS DE
4.89
7,684
1929SP
0.95
2,889
BAKSANG
1.20
5,951
KASEUM
2.70
2,336
BULHANG
2.30
5,239
1259SP
2.31
2,348
BBONGBAT
1.80
2,288
NEWS TODAY 2
1.97
1,335
TEUK NEWSTODA
3.50
1,534
TEUK NEWSTODA
3.20
1, 856
MYUNGTAE 2
4.80
3,668
CHONGSUN 1
6.30
7,238
CHONG 7
6.80
2, 948
9
1.00
2, 948
10
0.40
2,948
11
3.00
2,948
12
3.70
1,658
13
5.30
2,190
0559SP
3.00
1,039
NEWS TEUKBO
0.50
2,535
JUNKUK
3.01
1,331
20:00 SBS NEW
2.08
5, 644
1959SP
1.78
1,881
2029SP
1.74
1, 881
WORLDCUP KI
2.70
3, 653
SPARKMAN
0.50
4,155
96 CHONG
0.70
2,809
CHUNG 6
1.50
2, 861
7
1.00
2, 861
8
0.50
2,861
CHUNG 9
0.01
2,861 9
CHUNG 10
0.30
2,861
COMEDY
8.60
213
TOMATO
2.80
3, 604
MANAM
2.75
730
1349SP
1.10
1,096
0729SP
2.73
318
MANGANG
2.30
730
MYUNGSA
2.30
2,130
BADUK
0.40
2,430
JUNGMYUN
0.70
3, 604
14:35 NEWS
6.60
4,571
AHDUL
0.60
4,155
0159SP
1.00
206
2 MAGAZINES
Totals:
Sisa Journal
6.60
2,250
Han Kyurei 21
4.40
2,000
Jugan Chosun
5.40
2,250
News Plus
2.80
1, 875
Jugan Hankuk
1.60
1,500
News Week
2.80
2,250
News people
2.00
1,500
News Maker
1.70
1,500
Jugan MaeKyun
1.60
1,750
Economist
1.00
1,750
89
4
24,224
5.7
96
3
7,524
1.8
38
4
30,736
7.2
37
3
8, 667
2.0
31
1
5,951
1.4
84
1
2,336
0.5
93
1
5,239
1.2
88
2
4,696
1.1
37
1
2,288
0.5
59
3
4,005
0.9
61
1
1,534
0.4
33
1
1,856
0.4
47
1
3,668
0.9
30
1
7,238
1.7
45
1
2,948
0.7
27
1
2, 948
0.7
67
1
2, 948
0.7
76
1
2,948
0.7
94
1
1,658
0.4
77
1
2,190
0.5
54
1
1, 039
0.2
00
1
2,535
0.6
74
10
13,310
3.1
45
5
28,220
6.6
22
4
7,524
1.8
03
4
7,524
1.8
10
2
7,306
1.7
00
1
4,155
1.0
76
1
2,809
0.7
58
1
2, 861
0.7
37
1
2, 861
0.7
73
1
2,861
0.7
67
1
2,861
0.7
89
1
2,861
0.7
83
1
213
0.1
90
1
3, 604
0.8
85
2
1,460
0.3
21
1
1,096
0.3
88
2
636
0.1
58
1
730
0.2
87
1
2,130
0.5
50
1
2,430
0.6
62
1
3, 604
0.8
09
1
4,571
1.1
83
1
4,155
1.0
87
1
206
0.0
65
14
26,875
6.3
47
1
2,250
0.5
29
1
2,000
0.5
68
1
2,250
0.5
78
1
1, 875
0.4
29
1
1,500
0.4
53
1
2,250
0.5
43
1
1,500
0.4
03
1
1,500
0.4
17
1
1,750
0.4
67
1
1,750
0.4
51
16
52
101
165
28
75
33
42
22
14
19
25
38
14
98
245
32
14
13
11
169
14
90
35
36
45
277
133
63
95
190
536
317
0
42
8
33
3
10
30
202
171
23
230
6
60.
32
43
39
63,
89,
76
71,
84,
104,
166.

194
HanKyung Busi
0.50
1, 625
309.52
1
1, 625
0.4
Shin Dong-Ah
5.30
1, 875
33.69
1
1, 875
0.4
WalKan Chosun
5.30
1,875
33.69
1
1, 875
0.4
Win
1.20
2,875
228.17
1
2,875
0.7
Totals:
37.82
142
425,128
100.0

APPENDIX K
ADPLUS RESULTS FOR NETWORK TELEVISION AND MAGAZINES
COMBINED IN APRIL

196
ADplus(TM) RESULTS: NETWORK TV, MAGAZINES
Hyunsoo Park
University of Florida
APR. 17 ~ MAY 16, 1996
Frequency (f) Distributions
Vehicle
Message
Target: 2,884,983 f
Korean men 20-49 years
old who live in Seoul. 0
1
Message/vehicle =89.9% 2
3
4
5
6
7
8
9
10+
20+
Summary Evaluation
Reach 1+ (%)
Reach 1+ (000s)
Reach 3+ (%)
Reach 3+ (000s)
Gross rating points (GRPs)
Gross impressions (000s)
Average frequency (f)
Cost-per-thousand (CPM)
Cost-per-rating point (CPP)
Cost-per-net reach point (CPRP)
Cost-per-response (CPR)
Vehicle List
Rating
Ad Cost
1 NETWORK TV
Totals:
MINI SERIES
4.33
6,570
SHIMYA
2.50
3,101
PANKWAN 1
3.30
6,244
PANKWANG 3
6.00
3, 021
2054SP
6.23
3,103
JAE 3 NOON
2.99
4,545
MOKYOKTANG
2.20
3,158
AHCHIM 2
1.22
1,564
SPORTS
3.47
6, 053
TV INSANG
3.65
3,394
JOKWANGJO?
1.40
6,570
NEWSTODAY 2 M
1.97
1,335
WORLDCUP HIGH
3.89
6, 056
21:00 NEWS DE
4.89
7,684
1259SP
2.31
2,348
WUJUNG
3.28
4,264
1929SP
0.95
2,889
% f
% f+
% f
% f+
9.6
100.0
12.7
100.0
19.1
90.4
21.7
87.3
22.3
71.2
22.9
65.6
19.2
48.9
18.1
42.7
13.4
29.7
11.8
24.6
8.1
16.3
6.7
12.9
4.4
8.2
3.4
6.2
2.1
3.8
1.6
2.8
1.0
1.7
0.7
1.2
0.4
0.7
0.3
0.5
0.3
0.3
0.2
0.2
0.0
0.0
0.0
0.0
90.4%
87.3%
2,
711.2
2,
617.8
48.9%
42.7%
1,
467.3
1,
280.6
271.4
244.0
8,
142.0
7,
319.1
3.0
2.8
37.93
42.20
1,
138
1,
266
3,
417
3,
539
0.11
0.12
CPM-MSG
Ads
Total Cost
Mix %
41.01
85
281,957
91.3
50.58
2
13,140
4.3
41.35
2
6,202
2.0
63.07
2
12,488
4.0
16.78
1
3,021
1.0
16.60
4
12,412
4.0
50.67
2
9,090
2.9
47.85
1
3,158
1.0
42.73
7
10,948
3.5
58.15
3
18,159
5.9
31.00
2
6,788
2.2
156.43
2
13,140
4.3
22.59
8
10,680
3.5
51.89
2
12,112
3.9
52.38
2
15,368
5.0
33.88
4
9,392
3.0
43.33
2
8,528
2.8
101.37
3
8,667
2.8

197
CHERUNOVILLE
3.80
5,051
SPORTSHIGHLIG
4.00
6,056
19:00 NEWS LI
1.50
4,804
SAKWA
4.10
6,458
0729SP
2.85
1,123
1959SP SBS
1.78
1, 881
2029SP
1.74
1, 881
0729SP SBS
2.73
318
JUNKUK
3.01
1,331
20:00 SBS NEW
2.08
5, 644
POKSO
1.20
4,013
0629SP
1.55
155
2 MAGAZINES
Totals:
Sisa Journal
6.60
2,250
Han Kyurei 21
4.40
2, 000
Jugan Chosun
5.40
2,250
News Plus
2.80
1, 875
Jugan Hankuk
1.60
1,500
News Week
2.80
2,250
News people
2.00
1,500
News Maker
1.70
1, 500
Jugan MaeKyun
1.60
1,750
Economist
1.00
1,750
HanKyung Busi
0.50
1, 625
Shin Dong-Ah
5.30
1, 875
WalKan Chosun
5.30
1, 875
Win
1.20
2, 875
Totals:
31
1
5, 051
1.6
47
2
12,112
3.9
76
3
14,412
4.7
50
3
19,374
6.3
13
2
2,246
0.7
22
2
3,762
1.2
03
4
7,524
2.4
88
4
1,272
0.4
74
6
7,986
2.6
45
4
22,576
7.3
47
3
12,039
3.9
33
2
310
0.1
.65
14
26,875
8.7
.47
1
2,250
0.7
.29
1
2,000
0.6
.68
1
2,250
0.7
.78
1
1, 875
0.6
.29
1
1,500
0.5
.53
1
2,250
0.7
.43
1
1,500
0.5
. 03
1
1,500
0.5
.17
1
1,750
0.6
. 67
1
1,750
0.6
.52
1
1,625
0.5
. 69
1
1,875
0.6
.69
1
1,875
0.6
. 17
1
2,875
0.9
.20
99
308,832
100.0
44
50
106
52
13
35
36
3
14
90
111
3
60
32
43
39
63
89
76
71
84.
104.
166.
309
33.
33
228,
42.

APPENDIX L
ADPLUS RESULTS WITH FLOWCHART

199
ADplus(TM) FLOWCHART For AS IANA Campaign
FEB. 17 THRU MAY 16, 1996
Target audience: 2,884,983, Korean men 20-49 years old who live in Seoul
Prepared by: Hyunsoo Park, University of Florida.
MEDIA Total
Vehicles
Ads
Cost (000)
FEB
MAR
APR
NETWORK TV
338
1,150.9
125
128
85
1259SP
7
18.8
1
2
4
2249SP
6
28.4
4
2
0
Mini Series
10
65.7
4
4
2
Jae 3 Noon
8
36.4
1
5
2
Pankwan 1
10
62.4
4
4
2
21:00 News Des
11
84.5
5
4
2
NEWS TODAY 2
13
17.4
10
3
0
0749SP
8
6.1
4
4
0
1929SP
8
23.1
2
3
3
JAYEON
10
9.0
1
9
0
20:00 SBS NEWS
13
73.4
4
5
4
1959SP
8
15.0
4
4
0
2029SP
11
20.7
3
4
4
SHIMYA
5
15.5
0
3
2
2054SP
6
18.6
0
2
4
JUNKUK
16
21.3
0
10
6
0729SP
4
2.9
0
2
2
0959SP KBS
2
4.2
2
0
0
1339SP
1
6.3
1
0
0
13:00 News
1
4.5
1
0
0
SOS Haeyang
1
4.0
1
0
0
Dawon Bok 1
1
4.5
1
0
0
Dawon Bok 2
1
4.5
1
0
0
Bamkwa 2
2
6.2
2
0
0
Bob Roberts
1
4.0
1
0
0
Sulnal 1
1
1.3
1
0
0
Sulnal 2
1
1.9
1
0
0
Kum 2
1
8.0
1
0
0
Kum 3
1
5.0
1
0
0
Saekye 2
1
3.1
1
0
0
Sagi Masul
2
9.0
2
0
0
Star wa Go
1
5.1
1
0
0
Sigan 1
1
4.5
1
0
0
Simya
2
6.2
2
0
0
Queeze Norae
1
4.5
1
0
0
Queeze Myung
1
4.5
1
0
0
Pankwan 4
1
4.1
1
0
0
Pakwan 5
1
4.1
1
0
0
MBC Grand
1
4.5
1
0
0
Rocky V
1
8.3
1
0
0
Kajok 1
1
5.9
1
0
0
TEUK WOOJUNG
1
4.3
1
0
0
BLUE
1
2.6
1
0
0
ONGOJIP
1
1.4
1
0
0
MBC NEWS TODAY
1
1.4
1
0
0
1019SP
1
6.3
1
0
0
CHUNHANG 1
1
5.4
1
0
0
CHUNHANG 2
1
5.1
1
0
0

200
SEOCHO 1
1
5.8
1
0
0
SEOCHO 2
1
3.7
1
0
0
HWANG YA 7
1
3.9
1
0
0
HONGKILDONG
1
4.3
1
0
0
12:55 NEWS
1
4.7
1
0
0
BEST
1
4.3
1
0
0
1459SP
1
4.7
1
0
0
JUNGLEBOOK
1
4.7
1
0
0
SEHAE BOK
1
5.3
1
0
0
WORLDCUP 3
1
1.5
1
0
0
WORLDCUP 6
1
1.6
1
0
0
SAESANG
2
12.1
2
0
0
WOOJUNG
3
12.8
3
0
0
1239SP
1
6.3
1
0
0
24:00 NEWS
1
3.2
1
0
0
WORLDCUP HIGHL
2
12.1
2
0
0
TEUK GI SBS
1
5.3
1
0
0
TEUK KIM
1
2.7
1
0
0
TEUK SANG
1
4.7
1
0
0
TEUK JO
1
3.6
1
0
0
BOM
1
1.3
1
0
0
SHINTO 2
1
4.0
1
0
0
12:00 NEWS
1
4.0
1
0
0
SORIMSA
1
3.6
1
0
0
INSANG 1
1
2.5
1
0
0
INSANG 3
1
2.5
1
0
0
1404SP
1
1.1
1
0
0
KUT
1
4.6
1
0
0
GURIUM
1
2.4
1
0
0
NEWS TUEKBO 1
1
1.4
1
0
0
JUNGUK
6
8.0
6
0
0
1334SP
1
1.1
1
0
0
0109SP
1
0.2
1
0
0
BODO
1
5.2
1
0
0
1749SP
2
3.6
0
2
0
PAKWAN 3
1
3.0
0
1
0
1354SP
1
6.3
0
1
0
WURI
1
4.9
0
1
0
1159SP
1
2.4
0
1
0
TEUKSUN
1
4.1
0
1
0
TOWERING 1
1
1.9
0
1
0
TOWERING 2
1
1.5
0
1
0
WUJUNG MBC
5
21.3
0
5
0
WORLDCUP
4
24.2
0
4
0
JABAN
3
7.5
0
3
0
BAKSANG
1
6.0
0
1
0
KASEUM
1
2.3
0
1
0
BULHANG
1
5.2
0
1
0
BBONGBAT
1
2.3
0
1
0
TEUK NEWSTODAY
1
1.5
0
1
0
TEUK NEWSTODAY
1
1.9
0
1
0
MYUNGTAE 2
1
3.7
0
1
0
CHONGSUN 1
1
7.2
0
1
0
CHONG 7
1
2.9
0
1
0
9
1
2.9
0
1
0
10
1
2.9
0
1
0
11
1
2.9
0
1
0
12
1
1.7
0
1
0
13
1
2.2
0
1
0
0559SP
1
1.0
0
1
0

201
NEWS TEUKBO
1
2.5
0
1
0
WORLDCUP KI
2
7.3
0
2
0
SPARKMAN
1
4.2
0
1
0
96 CHONG
1
2.8
0
1
0
CHUNG 6
1
2.9
0
1
0
7
1
2.9
0
1
0
8
1
2.9
0
1
0
CHUNG 9
1
2.9
0
1
0
CHUNG 10
1
2.9
0
1
0
COMEDY
1
0.2
0
1
0
TOMATO
1
3.6
0
1
0
MAN AM
2
1.5
0
2
0
1349SP
1
1.1
0
1
0
MANGANG
1
0.7
0
1
0
MYUNGSA
1
2.1
0
1
0
BADUK
1
2.4
0
1
0
JUNGMYUN
1
3.6
0
1
0
14:35 NEWS
1
4.6
0
1
0
AHDUL
1
4.2
0
1
0
0159SP
1
0.2
0
1
0
PANKWANG 3
1
3.0
0
0
1
MOKYOKTANG
1
3.2
0
0
1
AHCHIM 2
7
10.9
0
0
7
SPORTS
3
18.2
0
0
3
TV INSANG
2
6.8
0
0
2
JOKWANGJO?
2
13.1
0
0
2
NEWSTODAY 2 MB
8
10.7
0
0
8
WORLDCUP HIGHL
2
12.1
0
0
2
WUJUNG
2
8.5
0
0
2
CHERUNOVILLE
1
5.1
0
0
1
SPORTSHIGHLIGH
2
12.1
0
0
2
19:00 NEWS LIN
3
14.4
0
0
3
SAKWA
3
19.4
0
0
3
1959SP SBS
2
3.8
0
0
2
0729SP SBS
4
1.3
0
0
4
POKSO
3
12.0
0
0
3
0629SP
2
0.3
0
0
2
AGAZINES
42
80.6
14
14
14
Sisa Journal
3
6.8
1
1
1
Han Kyurei 21
3
6.0
1
1
1
Jugan Chosun
3
6.8
1
1
1
News Plus
3
5.6
1
1
1
Jugan Hankuk
3
4.5
1
1
1
News Week
3
6.8
1
1
1
News people
3
4.5
1
1
1
News Maker
3
4.5
1
1
1
Jugan MaeKyung
3
5.3
1
1
1
Economist
3
5.3
1
1
1
HanKyung Busin
3
4.9
1
1
1
Shin Dong-Ah
3
5.6
1
1
1
WalKan Chosun
3
5.6
1
1
1
Win
3
8.6
1
1
1
OTALS
380
1,231.5
139
142
99

202
3 Month
COVERAGE
Gross
Mean
FEB
MAR
APR
Vehicle:
Reach 1+
286.1
95.3
96.4
97.1
92.6
Reach 3+
209.9
69.9
72.0
78.1
59.8
Reach 4+
160.7
53.6
53.5
63.5
43.7
Reach 5+
116.3
38.8
35.9
49.0
31.4
GRPs
1273.3
424.4
392.3
494.0
387.0
Frequency
Message:
4.4
4.3
4.1
5.1
4.2
Reach 1+
281.4
93.8
95.2
96.1
90.1
Reach 3+
196.5
65.5
67.5
74.5
54.5
Reach 4+
146.0
48.7
48.3
59.1
38.9
Reach 5+
103.2
34.4
31.2
44.5
27.5
GRPs
1176.6
392.2
364.9
460.1
351.6
Frequency
4.2
4.2
3.8
4.8
3.9
MONTHLY COST (000s)
NETWORK TV
470.7
398.3
282.0
MAGAZINES
26.9
26.9
26.9
TOTALS
497.5
425.1
308.8
The advertising carry-over fixed function rate is 23.4%.
The message/vehicle ratio for magazine advertisements is 35%.

REFERENCES
Aaker, D. A., Batra, R., and Myers, J. G. (1992). Advertising management (4th
edition). Prentice Hall, Englewood Cliffs, NJ.
AD DATA (1995). AD DATA 1995 MBC AD Comm., Seoul, Korea
Agostini, J. M. (1961). How to estimate unduplicated audiences Journal of Advertising
Research 1. 11-14.
Agresti, A. and Finlay, B. (1986). Statistical methods for the social sciences (2nd
edition). Dellen Publishing Co., San Francisco, CA.
Assmus, G„ Farley, J. U., and Lehmann, D. R. (1984). How advertising affects sales:
Meta-analysis of econometric results. Journal of Marketing Research 21. 65-74.
Bass F. M. and Clarke, D. G. (1972). Testing distributed lag models of advertising
effect. Journal of Marketing Research 9. 298-308.
Biehal, G„ Stephens, D., and Curio, E. (1992), Attitude toward the ad and brand
choice. Journal of Advertising 21. 19.
Brown, G. (1993). Attention and memory of TV and magazines. Admap (December),
15-20.
Brown, L. O., Lessler, R. S., and Weillbacher, W. M. (1957). Advertising Media.
Ronald Press, New York, NY.
Burke, W. L , and Schoeffler, S. (1980). Brand awareness as a tool of profitability.
Working Paper, The Strategic Planning Institute. Boston, NY.
Burke, M. C., and Edell, J. A. (1986). Ad reactions over time: Capturing changes in the
real world. Journal of Consumer Research 13. 114-118.
Burke, M. C., and Edell, J. A. (1989). The impact of feelings on ad-based affect and
cognition. Journal of Marketing Research 26. 69-83.
203

204
Calder, B. 1, and Stemthal, B (1980). Television commercial wearout: An information
processing view. Journal of Marketing Research 17. 173-186.
Choi, B. (1994). The Character of Korean Society. Nutinamoo, Seoul, Korea.
Clark, E. M, Brock, T. C, and Stewart, D. W. (1993). Attention, attitude, and affect in
response to advertising. LEA publishers, Hillsdale, NJ.
Cohen, D. (1988). Advertising. Scott, Foreman and Company, Glenview, IL.
Cox, D. S., and Cox, A. D. (1988). What does familiarity breed? Complexity as a
moderator of repetition effects in advertisement evaluation. Journal of Consumer
Research 15. 111-116.
Craig, C. S., and Ghosh, A. (1986). The development of media models in advertising :
An anthology of classic articles Garland Pub., New York, NY
Craig, C. S., and Ghosh, A. (1993). Using household-level viewing data to maximize
effective reach. Journal of Advertising Research 33. 38-48.
DeKimpe, M. G, and Hanssens, D. M. (1995a) The persistence of marketing effects
on sales Marketing Science 14. 1-21.
DeKimpe, M. G., and Hanssens, D. M. (1995b). Empirical generalizations about
market evolution and stationarity. Marketing Science 14. G109-G121
Donius, J. F. (1986). Marketplace measurement: Tracking and testing advertising and
other marketing effects. ANA Inc., New York, NY.
Dutka, S. (1995). DAGMAR. defining advertising goals for measured advertising
results (2nd edition). NTC Business Books, New York, NY.
Elliott, J. (1985). How Advertising Frequency Affects Advertising Effectiveness:
Indications of Change, ADMAP.
Falcheck, K. A. (1995). Estimating the impact of advertising media plans: Media
executives describe weighting and timing factors. Unpublished thesis. University of
Florida. Gainesville, FL
Fletcher, D. A., and Bowers, A. T. (1988). Fundamentals of advertising research (3rd
edition). Wadsworth Publishing Co., Belmont, CA.
Hansen, F (1995). Recent developments in the measurement of advertising
effectiveness: the third generation. Marketing and Research Today ^Netherlands) 23.
259-69,

205
Hastak, M., and Olson, J. C. (1989). Assessing the role of brand-related cognitive
responses as mediators of communication effects on cognitive structure. Journal of
Consumer Research 15, 444-456.
Higie, R. A., and Sewall, M. A. (1991). Using recall and brand preference to evaluate
advertising effectiveness. Journal of Advertising Research 31. 56-63.
Holstius, Karin (1990). Sales response to advertising. International Journal of
Advertising 9. 38.
Homer, P. (1990). The mediating role of attitude toward the ad: Some additional
evidence. Journal of Marketing Research 27. 76-86.
Joyce, T. (1984). Page Exposures. Mediamark Research Inc., New York, NY.
Ju, K. H. (1991). Simple approaches to modeling advertising media exposure.
Unpublished doctoral dissertation, The University of Texas at Austin., Austin, TX.
Ju, K. H., and Leckenby, J. D (1990). Performance of a simple reach/frequency model
In P.A. stout (ed.), Proceedings of the 1989 Conference of the American Academy of
Advertising. (RC27-32). American Academy of Advertising , Richmond, VA.
Kang, J. (1995). An Evaluation of Methodology to Estimate the Carryover Effect of
Advertising Communication: the case of Korean confectionery industry. Unpublished
master’s thesis. Korean University of Foreign Languages, Seoul, Korea.
Kim, C. (1990). Rush of transnational advertising agencies into the Korean advertising
industry. Kwangko Chongbo (December), 60-65.
Kim, K. (1996). Advertising in Korea: International challenges and politics. Advertising
in Asia edited by Frith, K. T., Iowa State University, Iowa.
Koo, H. (1993). Labor and economic growth in five Asian countries. Business History
Review 67. 702-703.
Korean Gallup Research Institute. (1992). Measuring Systems for TV Advertising
Effects. Korean Gallup, Seoul, Korea.
Kreshel, P. J., Lancaster, K. M„ and Toomey, M. A. (1985). How leading advertising
agencies perceive effective reach and frequency. Journal of Advertising 14. 32-38.
Krugman, H. E. (1972). Why Three Exposures May Be Enough. Journal of Advertising
Research 12. 3-9.

206
Kwango Yeonkam (1995). Kwango Yeonkam 95. Cheil Communications Co., Seoul,
Korea.
Lancaster, K. M. (1993). ADplus: For Multi-media Advertising Planning. Media
Research Institute Inc., Gainesville, FL.
Lancaster, K. M., and Helander, P. E. (1987). Forecasting advertising effects using
media exposure distribution models: Some test market results. Proceedings of the 1987
Annual Conference of the American Statistical Association. Section on Business and
Economic Statistics, 580-585.
Lancaster, K. M., and Katz, H. E. (1989). Strategic Media Planning. National
Textbook Co., Lincolinwood, IL.
Lancaster, K. M., Kreashel, P. I, and Harris, J. R (1986). Estimating the impact of
advertising media plans: Media executives describe weighting and timing factors.
Journal of Advertising 15. 21-29.
Lancaster, K. M., Martin, T. C. (1988). Estimating the impact of advertising media
plans: Media executives describe weighting and timing factors. Journal of Advertising
15, 21-29.
Lancaster, K M., Peiati, V., and Cho, J (1991). Perceptions of leading media directors
about advertising repetition effects. Journal of Media Planning 6. 3-16
Leckenby, J. D., and Boyd, M. M. (1984). An improved binomial reach/frequency
model for magazines. Current Issues and Research in Advertising 1. 1-24.
Leckenby, J. D., and Kim, H. (1994). How media directors view reach/frequency
estimation: Now and a decade ago. Journal of Advertising Research 34. 9-21.
Leckenby, J. D., and Kishi, S. (1982a). Performance of four exposure distribution
models. Journal of Advertising Research 22. 35-42.
Leckenby, J. D., and Kishi, S. (1982b). How media directors view reach/frequency
estimation. Journal of Advertising Research 22. 64-69.
Leckenby, J. D, and Rice, D. M. (1985). A beta binomial network TV exposure model
using limited data. Journal of Advertising 14. 25-31.
Lee, K. Y. (1995). Evaluating the Effectiveness of Television Advertising Schedules in
Terms of Advertising Exposure. Unpublished dissertation, University of Florida,
Gainesville, FL.

207
Lee, K., and Park, W. (1995). Advertising Effects and Media Planning. LG Ad. Co.,
Seoul, Korea.
Leone, R. P. (1983). Modeling sales-advertising relationships: An integrated time series
econometric approach Journal of Marketing Research 20. 291-295.
Liiien, G. L., Kotler, P., and Moorthy, K. S. (1992). Marketing models. Prentice Hall,
Englewood Cliffs, NJ.
Lodish, L. M., Abraham, M., Kalmenson, S., Livelsberger, J., Lubetkin, B.,
Richardson, B., and Stevens, M. E. (1995). How TV advertising works: A meta¬
analysis of 389 real world split cable TV advertising experiments. Journal of Marketing
Research 32, 125-139
Lowe, W. D. (1984). The use of intermediate measures in tracking and modeling an
advertising campaign. Journal of the Market Research Society 21. 46.
Machleit, K. A., and Wilson, D. R. (1988). Emotional feelings and attitude toward the
advertisement: The roles of brand familiarity and repetition. Journal of Advertising 17.
27-35.
Marquardt, R. A., and Murdoch, G. W. (1984). The sales/advertising relationship: An
investigation of correlations and consistency in supermarkets and department stores.
Journal of Advertising Research 24. 55-60,
McDonald, C. (1995). Advertising Reach and Frequency: Maximizing advertising
results through effective frequency (2nd Edition). NTC Business Books, Lincolinwood,
IL.
McDonald, C. (1982). Individual respondent analysis made better with complete single
source records. Effective Frequency: State of the Art. Advertising Research
Foundation, New York, NY, 181-194.
McDonald, C., and Carman, J. M. (1982). Are you overadvertising? Journal of
Advertising Research 22. 57-70.
McGann, A. F., and Russell, T. J. (1988). Advertising Media (2nd Edition). IRWIN,
Homewood, IL.
Mehta, A., and Purvis, S. C. (1994). Evaluating advertising effectiveness through
advertising response modeling. Paper presented at the thirteenth annual advertising and
consumer psychology conference on May 13-14, Minneapolis, MN.
Metheringham, R. A. (1964). Measuring the net cumulative coverage of a print
campaign. Journal of Advertising Research 4. 23-28.

208
Mitchell, A. A. (1993). Advertising exposure, memory, and choice. LEA Publishers,
Hillsdale, NJ.
Mitchell, A. A., and Olson, J. C. (1981). Are product attribute beliefs the only mediator
of advertising effects on brand attitude? Journal of Marketing Research 18. 318-332.
Montogomery, D. B., and Silk, A. J. (1972). Estimating dynamic effects of market
communications expenditures. Management Science 18 , B485-501.
Muehling, D. D., and Laczniak, R. N. (1988). Advertising's immediate and delayed
influence on brand attitudes: Considerations across message involvement levels. Journal
of Advertising 17. 23-34.
Naples, M. J. (1979). Effective frequency: The relationship between frequency and
advertising effectiveness. ANA Inc., New York, NY.
Nedungadi, P., Mitchell, A. A., and Berger, I. E. (1993). A framework for
understanding the effects of advertising exposure on choice in advertising exposure.
memory, and choice. Edited by Andrew A. Mitchell. Hillsdale, Erlbaum, NJ.
Ostrow, J. W. (1982). Setting effective frequency levels, effective frequency: The state
of the art. Advertising Research Foundation, Key Issues Workshop, New York, NY.
Palda, K. S. (1963). The measurement of cumulative advertising effects. Prentice-Hall,
Inc., Englewood Cliffs, NJ.
Park, C. W., and Young, S. M. (1986). Consumer response to television commercials:
The impact of involvement and background music on brand attitude formation. Journal
of Marketing Research 23. 11-24.
Park, W. (1995). Advertising Media Planning. Korean Advertising Research Institute,
Seoul, Korea.
Pechmann, C., and Stewart, D. W. (1988). Advertising repetition: A critical review of
wearin and wearout. Current Issues and Research in Advertising 11. 285-330.
Pieters, Rik G. M. and Klerk-Warmerdam, M. de (1996). Ad-evoked feelings: structure
and impact on A (sub a,d) and recall (attitude to advertisement). Journal of Business
Research 37. 105-14.
Poiesz, Theo B. C., and Robben, Henry S.J. (1994). Individual reactions to
advertising: Theoretical and methodological developments. International Journal of
Advertising 13. 25.

209
Rice, M. D (1985). Television exposure distribution models in advertising media
Unpublished dissertation, University of Illinois at Urbana-Champaign, 1L.
Russell, T , and Verrill, G. (1986). Otto Klenoner’s Advertising Procedure (9th
edition). Prentice Hall, Englewood Cliffs, NJ.
Rust, R. T. (1986). Advertising Media Models: A Practical guide. Health and
Company, Lexington, MA.
Rust, R. T., and Klompmaker, J. E. (1981). Improving the estimation procedure for the
beta binomial TV exposure model. Journal of Marketing Research 18. 442-448.
Ryu, P. (1993). Market Strategy and Competition. Bakyoungsa, Seoul, Korea.
Samuels, J. M. (1970 - 1971). The effects of advertising on sales and brand shares.
British Journal of Advertising 4. 187-207.
Schmarlensee, D. H. (1983). Today’s top priority advertising research questions.
Journal of Advertising Research 23, 53.
Schreiber, R. J., and Appel, V. (1991). Advertising evaluation using surrogate
measures for sales. Journal of Advertising Research 28. 27-31.
Schultz, D. E. (1989). Strategic advertising campaigns (2nd edition). NTC Business
Books, Lincolinwood, IL.
Schultz, D. E. (1994). Spreadsheet approach to measuring ROI for IMC. (return on
investment; integrated marketing communications). Marketing News 28. 12.
Schultz, D. E., and Barnes, B. E. (1994). Strategic advertising campaigns (4th edition).
NTC Business Books, Lincolinwood, IL.
Schultz, D. E., Martin, D., and Brown, W. P. (1984). Strategic Advertising Campaigns
(2nd edition). NTC Business Books, Lincolinwood, IL.
Schumann, D. W., Petty, R. E., and Clemons, D. S. (1990). Predicting the effectiveness
of different strategies of advertising variation: A test of the repetition-variation
hypotheses. Journal of Consumer Research 17. 192-202.
Shin, I. (1986). The history of advertising in Korea. Nanam Publications Co., Seoul,
Korea.
Shin, I. (1989). Advertising in Korea. Sisa-Yongosa Inc., Seoul, Korea.
Simon, J. L., and Arndt, J. (1980). The shape of the advertising response function.
Journal of Advertising Research 20. 11-30.

210
Sissors, J. Z., and Bumba, L. (1996). Advertising media planning (5th edition). NTC
Business Books, Lincolinwood, IL.
Sissors, J. Z., and Surmanek, J. (1986). Advertising media planning (2nd edition). NTC
Business Books, Lincolinwood, IL.
SPI study. (1980). Brand awareness increases market share, profits. Marketing News
(November), 5.
Sterling, K. (1997). Not for the weary; Using effective frequency to gauge ad
exposure; Technology Information Marketing Computers 17. 60.
Stewart, D. W. (1989). Measures, methods, and models in advertising research. Journal
of Advertising Research 29. 54-60.
Stewart, D. W. (1986). The moderating role of recall, comprehension, and brand
differentiation on the persuasiveness of television advertising. Journal of Advertising
Research 26, 43-47.
Stewart, D. W., and Furse, D. H. (1986). Effective television advertising: A study of
1000 commercials. Lexington Books, Lexington, MA.
Sung, T. Y. (1989). A study of the establishment of a measure model for pre-estimation
and evaluation of advertising effect in TV-CM. Unpublished Dissertation, Sung-Keun-
Kwan University, Seoul, Korea.
Walker, D., and Von Gonten, M. F. (1989). Explaining related recall outcomes: New
answers from a better model. Journal of Advertising Research 29. 11-21.
Wilson, C. E. (1981). Procedure for the analysis of consumer decision making. Journal
of Advertising Research 21. 31-38.
Woodside, A. G. (1996). Measuring the Effectiveness of Image and Linkage
Advertising. Quorum Books, Westport, CT.
Yoon, H. (1980). A Comparative Study on Advertising Ethics and Consumerism
Among U S A,, Japan and R.O.K. Seokang University, Seoul, Korea.
Yoon, Y. (1994). Political economy of television broadcasting in South Korea. In:
Kim, C. W., & Lee, J. W., eds. Elite Media Amidst Mass Culture: A Critical Look at
Mass Communication in Korea. Nanam Publishing Flouse, Seoul, Korea.
Zuffyden, F. S. (1996). Linking advertising to box office performance of new film
releases - a marketing planning model. Journal of Advertising Research 36,29.

BIOGRAPHICAL SKETCH
Hyunsoo Park was born on March 22, 1966, in Seoul, Korea and spent most of his
life growing up in Seoul. After he received his Bachelor’s degree in Korean language and
literature in 1993 from Chung-Ang University located in Seoul, he attended the University
of Florida and received his Master’s degree in Mass Communications with an advertising
specialty in 1995. He was also admitted to the University of Florida to pursue a Ph. D. in
Mass Communications with an advertising specialty and is expected to receive his degree
in May of 1998.
After completing the Ph. D. program, Hyunsoo plans to return to Korea to teach
advertising at a Korean university. He also would like to work for an advertising agency
as a researcher or a consultant.
Hyunsoo worked for the CRC (Communication Research Center) at the University
of Florida as a research assistant. His Master’s study, “Evaluating the ethics of common
advertising practices, ” which was presented at the AEJMC in 1995, is under review for
publication in the Journal of Mass Media Ethics. His current research interests are
advertising media planning and predicting or measuring advertising effects. His two
studies related to media planning in South Korea will be submitted for publication.
211

I certify that I have read this study and that in my opinion it conforms to acceptable standards of
scholarly presentation and is fully adequate, in scope and quality, as a disserumerpfor the degree of Doctor
of Philosophy.
(mt M. Lancaster. Chairperson
Professor of Journalism and Communications
I certify that I have read this study and that in my opinion it conforms to acceptable standards of
scholarly presentation and is fully adequate, in scope and quality7as a dissertation for thoti^ree of Doctor
of Philosophy. / / //
â–  / -/*
/¿^ ■ /.¿Jtf-yx.
Jps^ph R. Pisani
professor of Journalism and Communications
I certify that 1 have read this study and that in my opinion it conforms to acceptable standards of
scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor
of Philosophy.
Marilyn Robert
Associate Professor of Journalism and
Communications
I certify that I have read this study and that in my opinion it conforms to acceptable standards of
scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor
of Philosophy.
p ^ p
Leonard P. Tipton
Professor of Journalism and Communications
I certify that I have read this study and that in my opinion it conforms to acceptable standards of
scholarly presentation and is fully adequate, in scope and qualityrlislulissertation for the degree of Doctor
of Philosophy. ^
^^Ríchard Lcv^s^íícafíbr^^^
Professor of Statistics
This dissertation was submitted to the Graduate Faculty of the College of Journalism and
Communications and to the Graduate School and was accepted as partial fulfillment of the requirements
for the degree of Doctor of Philosophy.
May, 1998
(-X^T-Y—i 74v |
Déan. College «f Journalism and
Communications
Dean. Graduate School

LD
1780
199fi
• Y2S
UNIVERSITY OF FLORIDA
3 1262 08556 6668

LD
1780
199fi
• Y2S
UNIVERSITY OF FLORIDA
3 1262 08556 6668