Citation |

- Permanent Link:
- http://ufdc.ufl.edu/UFE0045264/00001
## Material Information- Title:
- A Study of New Electronic Commerce Models
- Creator:
- Yang, Yu-Chen
- Publisher:
- University of Florida
- Publication Date:
- 2013
- Language:
- English
## Thesis/Dissertation Information- Degree:
- Doctorate ( Ph.D.)
- Degree Grantor:
- University of Florida
- Degree Disciplines:
- Business Administration
Information Systems and Operations Management - Committee Chair:
- Cheng, Hsing
- Committee Members:
- Horowitz, Ira
Aytug, Haldun Shugan, Steven Mark - Graduation Date:
- 5/4/2013
## Subjects- Subjects / Keywords:
- Betting ( jstor )
Business models ( jstor ) Consumer prices ( jstor ) Consumer sectors ( jstor ) Customers ( jstor ) Marketing strategies ( jstor ) Merchants ( jstor ) Prices ( jstor ) Retail stores ( jstor ) Sales rebates ( jstor ) ec group-buying rebate
## Notes- General Note:
- Electroniccommerce is the trades of products or services over Internet or computernetworks. Social commerce is one of the hottest topics in e-commerce. Consumersobtain a discount deal when the group of buyers reaches a required minimumnumber. My main interest in this dissertation is to explore these businessmodels in e-commerce. In the firstessay, Groupon is themost notable social commerce websites that take advantage of the transformativeeffect of social media on businesses. As one of the hottest sectors ine-commerce, these social commerce websites offer consumers a significantreduction in the price of products or services when the group of buyers reachesa required minimum number. The objective of this research is to provide usefulguidelines to the merchants on whether it is in their best interests to offer adeal on Groupon; and if yes, when is the optimal timing to do so, and what arethe optimal price and the optimal predetermined threshold number of buyers forthe deal to become valid. The second essay considers the oBaz businessmodel, which can be thought as aGroupon in reverse, and provides user-generated group deals. The objective of this research to investigate whether it is in the merchantâ€™s best interests to offer a deal on oBaz; and if yes,what is the optimal price to do so. Furthermore, our study shows thatoBazâ€™s segmentation on consumers outperforms than the typical marketingsegmentation when the proportion of informed consumers is intermediate.
## Record Information- Source Institution:
- UFRGP
- Rights Management:
- All applicable rights reserved by the source institution and holding location.
- Embargo Date:
- 5/31/2015
## UFDC Membership |

Downloads |

## This item has the following downloads: |

Full Text |

PAGE 1 1 A STUDY OF NEW ELECTRONIC COMMENCE MODELS By YU CHEN YANG 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 UN IVERSITY OF FLORIDA 2013 PAGE 2 2 2013 Yu Chen Yang PAGE 3 3 To my family, for a lifetime of love and support PAGE 4 4 ACKNOWLEDGMENTS I would like to express my gratitude to my advisor Dr. Hsing Kenny Cheng whose expertis e, understanding, and patience, added considerably to my graduate experience. I appreciate his vast knowledge and skill in many areas and his assistance in writing paper s My thanks also go to my committee members, Dr. Ira Horowitz, Dr. Haldun Aytug and Dr Steve Shugan for their comments and constructive suggestions. Finally, this work is dedicated to my parents and grandparents. I would also like to thank my family for the support they provided me through my entire lif e PAGE 5 5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 ABSTRACT ................................ ................................ ................................ ..................... 9 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 11 Electronic Commerce ................................ ................................ .............................. 11 Structure of the Dissertation ................................ ................................ ................... 11 2 OPTIMAL SOCIAL COMMERCE: A STUDY OF GROUPON ................................ 13 What is Groupon? ................................ ................................ ................................ ... 13 Literature Review ................................ ................................ ................................ .... 15 Models and Analyses ................................ ................................ .............................. 17 The Model of Groupon ................................ ................................ ..................... 17 Case N N: N o Groupon P romotion in B oth P eriods ................................ ........... 20 Case GN: Running Groupon in The First Period ................................ .............. 22 Case N G : Running Groupon in The Second Period ................................ ......... 25 Empirical Analyses ................................ ................................ ................................ 26 Summary ................................ ................................ ................................ ................ 29 3 CONSUMER GENERATED GROUP DEALS: A STUDY OF OBAZ ...................... 36 Background ................................ ................................ ................................ ............. 36 Literature Review ................................ ................................ ................................ .... 37 Models and Analyse s ................................ ................................ .............................. 38 Case N: N o oBaz Promotion ................................ ................................ ............ 41 Case J : Join oBaz ................................ ................................ ............................ 42 Case N Ve rsus Case J ................................ ................................ ..................... 48 Empirical Analyses ................................ ................................ ................................ 51 Summary ................................ ................................ ................................ ................ 52 4 A STUDY OF NEW E COMMERCE BUSINESS MODELS BASED ON ................................ ................................ ............. 58 StickK and Gym Pact ................................ ................................ .............................. 58 Literature Review ................................ ................................ ................................ .... 59 Models and Analyses ................................ ................................ .............................. 60 PAGE 6 6 Model R : Standard Retailer Model ................................ ................................ ... 61 Model S : StickK Business Model ................................ ................................ ...... 62 Model GP : Gym Pact Business Model ................................ ............................. 65 Comparisons among Standard Retailer Model, StickK Model and Gym Pact Model ................................ ................................ ................................ ............ 68 Summary ................................ ................................ ................................ ................ 70 5 CONCLUSION ................................ ................................ ................................ ........ 74 APPENDIX: PROOF OF LEMMAS AND PROPOSITIO NS ................................ .......... 76 LIST OF REFERENCES ................................ ................................ ............................. 139 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 142 PAGE 7 7 LIST OF TABLES Table page 2 1 Summary of Notation ................................ ................................ .......................... 31 2 2 Summary of Proposition 2 3 ................................ ................................ ............... 31 2 3 Summa ry of Proposition 2 4 ................................ ................................ ............... 32 2 4 Baseline parameters for numerical analysis ................................ ....................... 32 3 1 Summary of notation ................................ ................................ .......................... 54 3 2 Baseline parameters for numerical analysis ................................ ....................... 54 4 1 Summary of notation ................................ ................................ .......................... 72 PAGE 8 8 LIST OF FIGURES Figure page 2 1 An example of Groupon promotion ................................ ................................ ..... 33 2 2 Example of the m erchant s etting a l ow r equired m inimum s old u nits ................. 33 2 3 Optimal profit versus proportion of informed customers ( ) ............................... 34 2 4 Optimal profit versus reservation price of informed customers ( ) .................... 34 2 5 Optimal profit versus price sensitivity of uninformed customers ( ) ................. 35 3 1 An example of the Bag it game ................................ ................................ ........... 54 3 2 An example of the oBaz discount ................................ ................................ ....... 55 3 3 Optimal profit under negligible marginal cost versus proportion of informed consumers ( ) ................................ ................................ ................................ ... 55 3 4 reservation price of informed co nsumers ................................ ............................ 56 3 5 reservation price of informed consumers ................................ ............................ 56 3 6 Optimal profit versus proportion of informed consumers ( ) ............................. 57 3 7 Optimal profit versus price sensitivity of informed consumers ( ) ..................... 57 4 1 The standard retail model ................................ ................................ ................... 72 4 2 The StickK business model ................................ ................................ ................ 73 4 3 The Gym Pact business model ................................ ................................ ........... 73 PAGE 9 9 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 A S TUDY OF NEW ELECTRONIC COMMENCE MODELS By Yu Chen Yang May 2013 Chair: Hsing K. Cheng Major: Business Administration Electronic commerce is the trades of products or services over Internet or computer networks. Social commerce is one of the hottest topi cs in e commerce. Consumers obtain a discount deal when the group of buyers reaches a required minimum number. My main interest in this dissertation is to explore these business models in e commerce. In the first essay, Groupon is the most notable social commerce websites that take advantage of the transformative effect of social media on businesses. As one of the hottest sectors in e commerce, these social commerce websites offer consumers a significant reduction in the price of products or services when the group of buyers reaches a required minimum number The objective of this research is to provide useful guidelines to the merchants on whether it is in their best interests to offer a deal on Groupon; and if yes, when is the optimal timing to do so, and what are the optimal price and the optimal predetermined threshold number of buyers for the deal to become valid. The second essay considers the oBaz business model, which can be thought as a Groupon in reverse, and provides user generated group deals. T he objective of this research to investigate whether it is in the merchant s best interests to offer a deal on PAGE 10 10 oBaz ; and if yes, what is the optimal price to do so Furthermore, our study shows that oBaz s segmentation on consumers outperforms than the ty pical marketing segmentation when the proportion of informed consumers is intermediate. In the third essay, we aim to study two n ew e c ommerce b usiness m odels the StickK and Gym Pact models. These two business models work very similarly, and both are base d on be tting on c onsumer w agers The research purpose is to examine whether the retailer will use the StickK model or the Gym Pact model. We use a traditional retail model as the benchmark and find that these two models allow consumers to have an opportuni ty to win their money back. It becomes an incentive to drive consumer demand. However, most consumers are unable to take the money back, especially when the effort cost is high, and thus retailers benefit from the consumers who lose money. PAGE 11 11 CHAPTER 1 INTR ODUCTION Electronic Commerce Electronic commerce is the trades of products or services over Internet or computer networks. Electronic commerce provides the merchants a different business model to sell or promote their products. Consumers also seek for a pr omotional deal over the Internet. Groupon, for example, creates the business model of group buying and offers the daily deals of various products and services with a discount of 50 percent off regular prices if a certain number o f people sign up for the de al oBaz, online bazzar works as a Groupon in reverse. oBaz offers a consumer oriented service and does the haggling for their consumers while comparing to the above group buying websites The objective of the this dissertation is to analyze the pricing strategy for these electronic commerce models. Structure of the Dissertation The dissertation is organized as a collection of three essays, each of which covers one of several aspects of the entire study. Each chapter corresponds to an essay which is compl etely by itself. Due to this self contained style of preparation, some redundancies across chapters may arise. This section is intended to give an outline of this dissertation. In Chapter 2, t he objective is to provide useful guidelines to the merchants on whether it is in their best interests to offer a deal on Groupon; and if yes, when is the optimal timing to do so, and what are the optimal price and the optimal predetermined threshold number of buyers for the deal to become valid. W e develop a two perio d model to investigate the optimal pricing strategy for the merchant to provide a Groupon PAGE 12 12 deal. The merchant faces a capacity constraint in each of the two periods. The merchant can run a Groupon deal either in the first period, the second period, or none at all. In Chapter 3 we analyze the oBaz business model and aim to provide managerial insights to the merchants W hat is the optimal price to do so ? Furthermore, when comparing to the typical marketing segmentation, is there any different result with oBaz s segmentation? We employ a linear discount function to study the optimal pricing strate gy. In Chapter 4, we aim to study two n ew e c ommerce b usiness m odels the StickK and Gym Pact models. These two business models work very similarly, and both are based on be tting on c onsumer w agers The research purpose is to examine whether the retailer will use the StickK model or the Gym Pact model. The discussion of the results and future work is concluded in Chapter 5. PAGE 13 13 CHAPTER 2 OPTIMAL SOCIAL COMMERCE: A STUDY OF GROUPON What is Groupon? G roup buying, offering consumers a significant reduction in the price of products or services when the group of buyers reaches a required minimum number has become one of the hottest sectors in e commerce G rou pon, a name combi n ation of first launched in Chicago in November 2008, is a popular group buying website that caters to major cities in the North America, Europe South America and Asia According to Forbes, Groupon is one of the fastest growing compan ies in e commerce history and has around 3,000 business owners signed up to offer deals and more than 35 million subscribers in more than 250 markets ( Steiner 2010 ) Groupon usually offers the daily deals of vario us products and services with a discount of 50 percent off regular prices if a certain number o f people sign up for the deal. For example, consumers can pay $10 to purchase a $ 20 worth of movie tickets if the predetermined number of buyers say 100, sign u p for the deal in a day. Many merchants have jumped on the Groupon bandwagon by offering deals of the day on its website with an aim to attract new customers and increase the sales volume. However, they fail to consider certain negative effects before deci ding to offer seekers and bargain shoppers, not the average customers who are less sensitive to price changes, which is actually the segment of the market the merchants want to attract ( Hernandez 2010 ) Furthermore, the deep discounts more often than not lead to a decrea down the road ( Salkever 2010 ) Finally, many merchants do not have the sufficient PAGE 14 14 capacity to handle the surge of demand arising from deep discounts, leading to disgruntled customers and dilution of their brands. For example while Japanese were celebrating the New Year in 2011, Groupon Japan provided a deal which offered a 56% out dinner, and about 2 5 0 subscribers ha d purchased the tickets in a day. However, the deal provider was unable to handle the demand and delivered the dinner to many customers late and in terrible condition ( Koh 2011 Mokey 2011 ) The objective of this research is to provide useful guidelines to the merchants on whether it is i n their best interests to offer a deal on Groupon; and if yes, when is the optimal timing to do so, and what are the optimal price and the optimal predetermined threshold number of buyers for the deal to become valid. To this end, we build a two period ana lytical model to reflect the salient features of the reality involved in the merchant faces a capacity constraint in each of the two periods. The merchant can run a Groupon deal either in the first period, the second period, or none at all. The customer population under consideration consists of two segments of customers termed informed and uninformed customers. Both the informed and s product or service. Both segments of customers may purchase the product or service from the merchant without a Groupon deal. However, the informed consumers are more price sensitive than the uniformed consumers, and only the informed consumers are the de al seekers and bargain hunters as characterized by Varian (1980). Therefore, only informed customers are exposed to the promotion the merchant runs on Groupon. Furthermore, these PAGE 15 15 informed customers will become more price sensitive after being exposed to a promotion. independent of the required number of buyers for a Groupon deal to become valid, a key feature developed by Groupon. The merchant should set this required number of buye rs lower than the optimal demand from the informed customers. Empirical evidence suggests that many merchants set this number close to zero, effectively bypassing this key feature developed by Groupon. The best situation for the merchant to offer a Groupon deal is when the reservation price of the informed customers is low and/or the uninformed customers are less price sensitive. If the merchant is better off offering a Groupon deal, it does not make much difference when the Groupon deal in introduced in th e first period or the second period in the absence of an advertising effect. However, it is best to introduce a Groupon deal in the first period when there exists an advertising effect from running a Groupon promotion. Literature Review The work of Jing an d Xie (2011) and Edelman et al. (2010) are most relevant to our research. Jing and Xie (2011) compared two individual selling strategies margin strategy and volume strategy with a group buying strategy where the seller sells only to the informed customer s in the margin strategy, and sells to both the informed and the less informed customers in the volume strategy. In the group buying strategy, the seller sells at a discounted price only after a certain required group size is achieved. Jing and Xie (2011) find that the group buying strategy is most effective when the information gap between the informed and less informed customers is not too big and not too small, and the interpersonal information sharing is relatively efficient. PAGE 16 16 Although both Jing and Xie (2011) and our research examine whether group buying is more profitable for the merchant, the different inner workings of group buying platforms (Fundable.com versus Groupon.com) perhaps lead to a different characterization of the informed and the less inf ormed customers in Jing and Xie (2011) and the informed and uninformed customers in our study. The informed customer in than the less informed customer, and the less inf ormed customer has a lower valuation of the product due to the information gap. The informed customers in our setting have consider the informed customers as those who are more informed about the group selling deal than the product per se to reflect the empirical evidence that Groupon users are mostly deal seekers and bargain hunters. A key finding of our study is that the required number of purchases requirement in Groupon is n optimal pricing policy and optimal profit. Empirical evidence suggests that the merchant oftentimes sidestep this requirement by setting the required number of purchases as low as possible. This implies that the merchant offer s a deal in Groupon to price discriminate than to disseminate information of the product. Edelman et al. (2010) build a model of repeat experience good purchase to study the profitability of a merchant issuing discount vouchers through Groupon.com. They pa rtition the consumers into two different groups, each of which has a different cumulative distribution function describing the valuations of the consumers. The valuations of one group are systematically lower than those of the other group. They find that t he group of consumers who are offered the discount vouchers must be more PAGE 17 17 price sensitive for the vouchers to generate profitable price discrimination. Further, the higher the proportion of low valuation group of consumers, the more difficult it is for the discount vouchers to be profitable for the merchant. Our research incorporates the finding of Edelman, Jaffe, and Kominers (2010) by assuming that the informed customers who will access the Groupon deals in our model are more price sensitive than the unifo rmed customers, and that the reservation price of the informed customers is lower than that of the uninformed customers. We take a different perspective from that of Edelman, Jaffe, and Kominers (2010) in modeling the advertising effect. They assume that u ninformed consumers are unaware of the existence of the firm offering discount vouchers. In our study, both the informed and uninformed consumers are aware of the firm. The difference lies in the awareness of the existence of the discount vouchers. Models and Analyses The Model of Groupon There are two prominent features for the deals promoted on the Groupon website 1 First, most merchants usually offer half off their regular prices when they run Groupon promotions. Second, each merchant is required to set a minimum number of sold units. The Groupon deal becomes valid only if the consumer demand exceeds the required threshold number. For example, in F igure 2 1, the Groupon promotion by Q Saxon Judd a purveyor of custom clothing offers the discounted price of $100, 50% of the regular price of $200, and sets the required minimum number of sold units at 10. 2 Figure 2 1 shows that only three customers place orders which implies that at least 1 http://www.groupon.com 2 http://qcustomclothier.com/index.aspx PAGE 18 18 seven more buyers need to sign up for th is Groupon deal; otherwise, the deal is invalid and Q Saxon Judd makes no sales even at the discounted price To reflect the current practice of Groupon, we let be the regular price of the product or service under consideration, the discounted price on Grou pon be where and the minimum number of sold units required for Group deal to become valid be The merchant faces the critical decision problem of whether to run a promotion in Groupon. If the merchant decides to offer a Groupon deal, when is the optimal timing of running this deal and what are the optimal price and minimum number of sold units period problem. The merchant has three strategic options in this case. The first option is to not run Groupon promotion in either period, which we denote as the NN strategy ( N o Groupon in the first period, N o Groupon in the second period). The second option is to run Groupon in the f irst period denoted by GN strategy ( G roupon in the first period, N o Groupon in the second period). Finally, the NG strategy denotes running Groupon in the second period. T he merchant under consideration has a capacity (or perishable inventory) constraint o f in each of the two period s ( Koh 2011 ) Empirical evidence shows that Groupon deals attract mostly deal seekers and bargain hunters (Levine 2010), and we mod el such segment of consumers as the ( Varian 1980 ) The average consumers unaware of the Groupon represent the proportion of customers who are informed customers attracted to the G roupon deal, while is the proportion of uninformed customers The demand function for the informed customers is modeled as PAGE 19 19 ( 2 1) where and represen t the potential market size and the price sensitivity parameter of informed customers respectively. Salkever ( 2010 ) points out that customers will quickly get used to the new price after a promotion and the price will become e xtremely difficult to rise again. Hence, the price sensitivit y parameter of the informed customers will increase from to in the second period if the merchant offer s a Groupon promotion in the first peri od; otherwise, it stays the same in the second period. For uninformed customers the demand function is ( 2 2) w here denotes the potential market size of uninformed customers and captures uninformed customers price sensitivity. Since informed customers are more price sensitive than uninformed customers, one has Table 2 1 summarizes the notation used in our model. The reservation price (i.e. maxim um willingness to pay) of the uninformed consumers is by setting Eq. ( 2 2) to zero. Likewise, t he reservation price of the informed consumers is b efore and during the Groupon promotion, and equals a fter a Groupon promotion. Since the informed consumers have a lower willingness to pay than that of uninformed consumers it follows that and That is, and ( 2 3) PAGE 20 20 In addition the reservation price s of both informed and uninformed customers are greater than the marginal cost i.e. and ( 2 4) Case N N: N o Groupon P romotion in B oth P eriods We first consider the benchmark case where there is no Groupon promotion in both periods The merchant faces the following problem. ( 2 5) subject to ( 2 6) ( 2 7) ( 2 8) ( 2 9) Eq. ( 2 5) function in the NN case where t he proportions of informed and uninformed c ustomer s are represented by and respectively Since there is no Groupon promotion in both periods, t he merchant thus has two identical profit structure s in these two period s Inequality ( 2 6) ensures th at the total customer demand will not exceed the capacity in each period. Inequalities ( 2 7) and ( 2 8) ensure that the demands of informed and uninformed consumers are nonnegative. Finally, Inequality ( 2 9) requires that the regu lar price be nonnegative. Solving the above problem leads to the following proposition P ROPOSITION 2 1. In the benchmark case of NN the optimal price and profit for the merchant are as follows: PAGE 21 21 1a a nd if a nd 1b and if 2a and if and 2b and if and Proof. See the proof in the Appendix Proposition 2 1 shows the merchant s optimal pricing policy in Case NN where t he capacity and the informed customer base are the two key factors influencing how the merchant price the product or the service. A h igh enough capacity and large enoug h informed customer base lead s the merchant to implement pricing policy ( 1a ) and a n intermediate capacit y result s in policy ( 1 b) An intermediate capacity and an intermediate informed customer bases lead s to pricing policy ( 2a) while a lower PAGE 22 22 capacit y and a lower informed customer base result in policy ( 2 b) The merchant will have demand from both informed and uninformed customers in p olicies (1a) and (1b) and demand only from the uninformed customers in policies (2a) and (2b). Case GN: R unning Gr oupon in T he F irst P eriod In this GN case, the merchant offer s the Groupon promotion in the first period. profit function in the GN case becomes ( 2 10) The first term gives the profit the merchant makes in the first perio d when running a Groupon offer The second term describes the profit for the merchant in the second period. Recall that the informed customers will increase their price sensitivities to in the second period after a Groupon promot ion in the first period. Thus, the merchant solves the following problem. ( 2 11) subject to ( 2 12) ( 2 13) ( 2 14) ( 2 15) ( 2 16) ( 2 17) ( 2 18) PAGE 23 23 ( 2 19) Inequalities ( 2 12) and ( 2 13) describe the capacity (or perishable inventory) constraints in the first and seco nd period. Inequality ( 2 14) ensures that the Groupon deal becomes valid. Inequalities ( 2 15) ( 2 17) follow from the requirement that the demands of informed and uninformed consumers are nonnegative. Inequalities ( 2 18) and ( 2 19) give the feasible range f or the decision variables and A key finding of the Case GN is that the optimal price and the optimal profit do not depend on the minimum number of buyers for the Groupon deal to become valid, The optimal can take on any positive value not greater than The following proposition summarizes the key findings of the GN Case. P ROPOSITION 2 2 (Case GN ) The optimal price and the optimal profit are independent of the minimum number of buyers for the Groupon deal to become valid. The optimal can take any value in the range of Proof. See the proof in th e Appendix N ote that the optimal minimum number of required purchases described in Proposition 2 2 is found through the Kuhn Tucker condition of Inequality ( 2 14) the constraint of valid deal requirement. If the merchant sets th is value too high, it will lead to an invalid Groupon deal and the merchant incurs a loss of potential sales that should have been realized under a lower valid deal requirement. Furthermore, the optimal profit is independent of the value of Hence, it is i n the merchant low value of Empirical evidence shows that many merchants follow the prescription of Proposition 2 2 by setting the threshold value as small as possible. For exa mple, the minimum number of required purchases in Figure 2 2 is one, a number that is as close PAGE 24 24 to zero as permitted by Groupon The same finding of Proposition 2 2 applies to the case when the merchant runs the Groupon deal in the second period, the NG cas e to be discussed later. The optimal pricing policy and profit when the merchant runs the Groupon deal in the first period (the GN case) are far more complex than those in the benchmark NN case. In particular, the optimal pricing policy depends on whether there is demand from the informed customers in the second period (the non Groupon period), the relationship between the reservation prices of the uninformed and informed customers, and the capacity of the merchant. The results are summarized in the followi ng proposition. P ROPOSITION 2 3 When the merchant runs the Groupon deal in the first period the optimal price and profit are summarized in Table 2 2. Proof. See the proof in the Appendix Depending on the characteristics of underlying parameters of the problem at uninformed consumers population), there are ten different optimal pricing policies when the merchant opts to run a Groupon deal in the first period. The first c haractering condition in Proposition 2 3 is whether there exists non negative demand from informed consumers in both periods. Case 1 describes the optimal pricing policy and profit for the merchant when demands from informed consumers in both periods are n on negative, while Cases 2 and 3 occur when there is no demand from informed consumers in the non Groupon period with the reservation price of uninformed consumers greater than twice that of the informed consumers separating Case 2 from Case 3. Within each of PAGE 25 25 the three cases, whether there exists excess capacity in the Groupon period or both periods further divides it into more sub cases. Case N G : R unning Groupon in T he S econd P eriod In the case of NG the merchant runs the Groupon promotion in the second p eriod, and the profit function becomes ( 2 20) The merchant faces the following problem : ( 2 21) subject to ( 2 22) ( 2 23) ( 2 24) ( 2 25) ( 2 26) ( 2 27) ( 2 28) ( 2 29) The constraints in Inequalities ( 2 22) ( 2 29) and their interpretation s are the same as those in Inequalities ( 2 12) ( 2 19). The results of C ase NG are summarized in the following proposition. PAGE 26 26 P ROPOSITION 2 4 When the merchant runs the Groupon deal in the second period the optimal price and profit are summarized in table 2 3. Proof. See the proof in the Appendix Notice that the characterizing conditions for the NG case in Proposition 2 4 are the same as those for the GN case in Proposition 2 3, although the optimal price and profit are different for the NG case. Empirical Analyses To obtain further insights Groupon we conduct ed extensive computational analys e s using empirical Groupon data in the restaurant category since restaurants exhibit many characteristics of a servic e industry where there exists a capacity constraint and the excess capacity in one period cannot be stored and used in the next period. We collected Groupon promotion data in Chicago, Los Angeles, and New York City from September, 2009 to November, 2010. F or the baseline values of parameters in our model, ( Talluri and Ryzin 2004 ) estimated t he elasticity of demand for restaurant meals at 2.3. We found that the p roportion of informed customers is 0.34 for coffee coupons, and 0.094 for tissue coupons ( Neslin et al. 1985 ) which we used as the range of poss ible values for in our models. We explore d the impact of several key parameters on the profitability of each Groupon strategy including (1) the p roportion of informed customers (2) the price sensitivi ty parameter of informed customers before a Groupon promotion (3) the potential market size for informed customers and (4) the capacity in each period, Table 2 2 shows the baseline values and the range of parameters in our analyses. The price sensitivity parameter of the uninformed PAGE 27 27 consumers, and that of the informed consumers after a Groupon promotion, are set in relat ion to For example, indicates that the uninformed consumers are 30% less price sensitive than the informed consumers. T he potential market size of uninformed customers is s et in relation to in a similar fashion. Our computational analyses results are show n in Figure s 2 3 through 2 5 Figure 2 3 shows how the proportion of informed customers influences the merchant s optimal profit for all three po ssible cases A s the proportion of informed customers increase s, t he merchant s optimal profit is decreasing in the cases of GN and NG while the merchant s optimal profit is slightly increasing in Case NN Our computation analyses also indicate that the o ptimal regular price in the cases of NN is very close to the optimal regular price in the case of GN However, the profit earned in Case GN is far below than that in Case NN E ven though much more customers are attracted by cutting the price in half in Cas e GN it actually leads to a lower profit when the merchant runs the Groupon promotion in the first period. To further deepen the loss in the Case GN the informed customers become more sensitive to price in the second period. The optimal regular price in Case NG is slightly higher than those in Case NN and Case GN The profit earned in the Groupon period of Case NG is greater than the Groupon period of Case GN but less than each period of Case NN Figure 2 4 indicates the relationship between the merchan t s optimal profit and reservation price of informed customers. In the case of NN the merchant would concentrate on uninformed customers and give up informed customers when the reservation price of informed customers is less than the threshold R1 in Figur e 2 4 Only when informed customers reservation price is larger than R1 will t he merchant find it PAGE 28 28 profitable to target both uninformed and informed customers in C a se NN However, the merchant can run a Groupon deal ( the case s of NG and GN ) to attract info rmed customers when their reservation price is low. Running a Groupon promotion either in the first period (Case GN ) or the second period (NG) will generate a higher profit than the benchmark case of NN in the region of low reservation prices of informed c onsumers. A s the reservation price of informed customers becomes large r than the threshold R2 i t is not beneficial for the merchant to run any Groupon promotion. Figure 2 5 sensitivity parameter of uninformed consumers while holding all other parameters decreases as the uninformed consumers become more price sensitive. Furthermore, running a Groupon promoti on generates a higher profit when uninformed consumers are less price sensitive, while not running any Groupon promotion works best when uninformed consumers are highly price sensitive. A lower price sensitivity parameter of uninformed consumers for a fixe d price sensitivity parameter of informed consumers amounts to the situation where informed consumers become more price sensitive for the a fixed price sensitivity parameter of uninformed consumers. The finding of Figure 2 5 complements that of Figure 2 4 In sum, we find that running a Groupon promotion generates a higher profit for the merchant when the reservation price of the informed consumers is low and/or the informed consumers are very price sensitive. Without a Groupon promotion, the merchant will find it more profitable to give up the informed consumers segment and PAGE 29 29 simply target the uninformed consumers in these situations. The Groupon promotion achieves the effect of price discriminating two different segments of the consumers. Our extensive compu tational analyses in Figures 2 3 through 2 5 indicate that running Groupon in the second period (the NG case) always dominates running Groupon in the first period (the GN case). This finding derives from the fact that the informed consumers will become mor e price sensitive after a Groupon promotion. In the next subsection, we explore the advertising effect of running a Groupon promotion in the first period where a portion of the informed consumers who purchased the Groupon deal are pleased with the merchant unchanged in the second period. Summary Social commerce electronic commerce that involves usin g social media online media that supports social interaction and user contributions, 3 has experienced explosiv e growth since its inception with Groupon as the most notable example of social commerce websites. Groupon provide s merchants opportunities to offer discount vouchers as featured daily deals in their websites to increase their profitability. The most criti cal questions to the merchant are whether offering a daily deal promotion results in a higher profit and if the answer is yes, when is the optimal timing of do so. In this paper, we develop analytical models that reflect the unique features of this social commerce website to address these questions. 3 http://en.wikipedia.org/wiki/Social_comme rce PAGE 30 30 We find that the merchant will realize a higher profit by offering a Groupon deal when the reservation price (i.e., willingness to pay) of informed customers, the segment of the consumers who are deal seekers an d bargain hunters, is low and/or when uninformed customers, the segment of the consumers who are unaware of the Groupon deal, are less price sensitive. Moreover, offering the same deal in Groupon will lead to a higher profit than in LivingSocial in the cur rent U.S. market where the proportion of informed consumers is less than 30 percent of the consumer population. We use half of the regular price as the Groupon price in our model to reflect the current Groupon business practice. A worthy future research di rection is to explore the option for the merchant to determine an optimal promotion price that is not tied to half of the regular price. Another potential extension to our research is to model the decision whether to run a Groupon promotion as a repeated g ame in an infinite planning horizon. PAGE 31 31 Table 2 1 Summary of Notation Parameter Description Regular price per unit Discounted price per unit Min imum value of sold units for Groupon deal to become valid P roportion of informed customers participating in promotions, M aximum potential market s izes for informed customers ( ) and for uninformed customers ( ) P rice sensitivities of the informed customers P rice sensitivities of the un informed customers C apacity of the merchant in units per period of time Marginal cost Table 2 2. Summary of Proposition 2 3 See the Appendix for proof and detailed expressio ns of optimal prices and profits in each case. Characterizing Conditions Optimal Price Optimal Profit Case 1: Non negative demand from informed customers in both periods Excess capacity in both periods Binding capacity in the Groupon period Case 2: Zero demand from informed customers in the non Groupon period and N on negative demand from informed customers in the Groupon period Excess capacity in both periods Binding capacity in the Groupon period Zero demand from informed customers in the Groupon period Excess capacity in both periods Binding capacity in both period s Case 3: Zero demand from informed customers in the non Groupon period and N on negative demand from u ninformed customers in both periods Excess capacity in both periods Binding capacity in the Groupon period Zero demand from uninformed customer s in both periods Excess capacity in the Groupon period Binding capacity in the Groupon period PAGE 32 32 Table 2 3. Summary of Proposition 2 4 See the Appendix for proof and detailed expressions of optimal prices and profits in each case. Table 2 4 Baseline parameters for numerical analysis Parameter Baseline values Range of value P roportion of informed customers participating in promotions, P rice sensitivities (informed) an d (uninformed) P otential market sizes (informed) and (uninformed) C apacity in units per period of time, Marginal cost Characterizing Conditions Optimal Price Optimal Profit Case 1: Non negative demand from informed customers in both periods Excess capacity in both periods Binding capacity in the Groupon period Case 2: Zero demand from informed customers in the non Groupon period and N on negative dem and from informed customers in the Groupon period Excess capacity in both periods Binding capacity in the Groupon period Zero demand from infor med customers in the Groupon period Excess capacity in both periods Binding capacity in both period s Case 3: Zero demand from informed customers in the non Groupon period and N on negative demand from uninformed customers in both periods Excess capacity in both periods Binding capacity in the Groupon period Zero demand from uninformed customers in both periods Excess capacity in the Groupon period Binding ca pacity in the Groupon period PAGE 33 33 Figure 2 1 An example of Groupon promotion (reprinted from http://www.groupon.com) Figure 2 2 Example of the m erchant s etting a l ow r equired m inimum s old u nits (reprinted f rom http://www.groupon .co.uk ) PAGE 34 34 Figure 2 3 Optimal profit versus proportion of informed customers ( ) Figure 2 4 Optimal profit versus reservation price of inf ormed customers ( ) R 2 R 1 PAGE 35 35 Figure 2 5 Optimal profit versus price sensitivity of uninformed customers ( ) PAGE 36 36 CHAPTER 3 CONSUMER GENERATED GROUP DEALS: A STUDY OF OBAZ Background Group buying is a shopping strategy to greatly reduce prices on the condition that enough consumers would make the purchase. The group buying business is rising lately. Groupon 1 launched on November 2008, owns more than 50 million subscribers and offers more than 22 million daily deals in North America. Groupon further help consumers save about $980 million 2 A nother three year old company, LivingSocial 3 serves more than 580 markets across 25 countries. LivingSocial has about 45 million members, and has provide d over 21 million deals already. These examples demonstrate that group buying brings huge potential business. oBaz 4 which stands for online bazaar offers a consumer oriented service and does the haggling for their consumers, as opposed to the a bove group buying websites. oBaz can be thought as a Groupon in reverse that provides user generated group deals. People can either create their own group by posting what they want at oBaz, or participate in another group of similarly desiring consumers. T hen oBaz s team will try to strike a deal with the merchant, based on the product or services the group wants. Members of group who think the deal is good enough to meet their needs would have to pay oBaz a fee to obtain the coupon and decide whether to pu rchase within 24 hours ( Eaton 2011 ) OBaz also helps merchant s segment their consumers by providing a discount deal. Bas ed on the concept of marketing segmentation, a monopoly can obtain increased profits by charging higher prices to those segments willing to pay more 1 http://www.groupon.com 2 http://www.groupon.com/pages/press kit 3 http://livingsocial.com 4 http://www.obaz.com/ PAGE 37 37 and charging less to tho se who are more sensitive to price However, will oBaz actually bring merchants mor e profits even if it attracts some consumers who need such products or services? This study aims to provide useful guidelines to merchants regarding whether or not it is in their best interests to offer a deal on oBaz ; and if yes, what is the optimal pric e? Furthermore, when compared to the typical marketing segmentation, is there any different result with oBaz s segmentation? Literature Review Social commerce is a recent marketing innovation and the marketplace of social commerce has some significant char acteristics, e.g., sellers are usually individual merchants, sellers are connected in online social network ( Stephen and Toubia 2010 ) and group buying organizers who are looking for price based incentives for volume purchases can aggregate individual buyers via the social web ( Anand and Aron 2003 ) The work of Anand and Aron (2003 ) is most relevant to our research. The study investigates the web based group buying mechanism and t he posted price mechanism via the models of demand uncertainty. The model of parallel demand regimes finds that the quantity demanded under one demand regime dominates that demanded under the other, while the model of intersecting demand regimes shows that neither demand regime absolutely dominates the other. They find that group buying beats posted prices in the model of intersecting demand regimes, but group buying can never exceed the revenues from post prices. In the model of intersecting demand regimes the sellers are able to set non linear price quantity schedules under group buying and thus maximize total expected revenue. PAGE 38 38 Some other works are related to social commerce, such as Jing and Xie (2011 ) and Edelman et al. (2011 ) Jing and Xie (2011 ) compare margin and volume s trategies with a group buying strategy, and show that the group buying strategy is more effective when the information gap between the informed and less informed customers is intermediate and the interpersonal information sharing is relatively efficient. Edelman et al. (2011 ) set up a model of rep eat experiences of good purchases and examine the profitability of merchant s issuing discount vouchers through Groupon.com They divide the consumers into two groups high valuation group and low valuation group and find that the group of consumers who are offered the discount vouchers become too sensitive to price to generate profitable price discrimination. Also, a high proportion of low valuation group of consumers makes the discount vouchers less profitable for the merchant. Our research incorporates t he key assumption of Edelman et al. (2011 ) that informed customers who will access oBaz deals are more sensitive to price than uniformed customers, and that the reservation price of informed customers is lower than that of uninformed customers. Our model also follows the model of Anand and Aron (2003 ) by normalizing the price sensitivity of uninformed consumers t o 1, and lets the price sensitivity of informed consumers be greater than 1, since the informed consumers are more price sensitive and the demand curve is steeper than that of uninformed consumers. Models and Analyses T he deal promoted on the oBaz website ( www.obaz.com ) has the following features First, after consumers log in to the website, they are presented with a series of game on the PAGE 39 39 website. Consumer s can button, i.e., to bag it, or the red to skip to the next product For example, in Figure 3 1, c onsumers would see the picture of the Metallic Druzy Ring when signing up for the game They either want the product or skip it Then oBaz negot iate s a group discount price with the merchant based on the number Finally, oBaz shows the finalized deal for consumers to purchase, such as when the merchant agrees to offer the Metallic Druzy Ring at the di scounted price of $ 70 3 9 % of the regular price $ 115 ( see Figure 3 2 ). Let the total number of consumers be normalized to 1. The number of the consumers who for a particular oBaz deal under consideration is denoted by while the number of the consumers who do not is oBaz uses the number of consumers who want the deal to negotiate with the merchant. The merchant applies a standard linear quantity di scount scheme ( Jucker and Rosenblatt 1985 ) in arriving at the discount price. Let be the reg ular price of the product under consideration, and the discount price offered to oBaz be which correspond s to linear quantity discount For example, which means 10% of consumers who click the Want it button. Then, the merchant offers 10% discount for the consumers who click the Want it! button, 20% discount for and so on. We note that in the beta version of oBaz, consumers pay the fee to obta in a promotion code that can help purchase the product from the merchant ( Eaton 2011 ) This fee is now included in the di scount price. Consumers pay oBaz the discount price, which includes the fee to receive the product directly from oBaz. PAGE 40 40 We describe the consumers who who click the Want it! button. as informed consumers since oBaz will inform th em about the promotional deal after they sign up for the deal. However, consumers who do not click will never be informed by oBaz and are ignorant of the promotional deal and the discounted price. Hence, informed consumers are more price sensitive than uni nformed consumers. Let the demand function for the informed consumers be ( 3 1) where represents the price sensitivity parameter of informed consumers. For uninformed consumers, the demand function is ( 3 2) where captures uninformed consumer price sensitivity, and since informed consumers are more sensitive to price than uninformed consumers. Without loss of generality, (e.g. Anand and Aron, 2003), we normalize to 1 and where The empirical study by Tellis ( 19 88 ) shows that the price sensitivity of informed consumers is no larger than four times the price sensitivity of uninformed consumers Table 3 1 summarizes the notation used in the model. The reservation price (i.e., maximum wil lingness to pay) of the informed consumers equals b y setting Eq ( 3 1) to zero, and that of the uninformed consumers is by setting Eq. ( 3 2) to zero Since the uninformed consumers are ignorant about th e promotional deal and the discounted price, they have higher reservation price to pay for the product than the informed consumers do. We let t he PAGE 41 41 informed consumers willingness to pay be lower than that of the uninformed consumers which implies ( 3 3) In addition the reservation price of consumers, including informed and uninformed consumers, are greater than the marginal cost i.e., and ( 3 4) In the next two sections, the merchant is required to consider whether or not to offer a deal on oBaz. Two scenario are discussed in the following. In the first case, Case N: No oBaz Promotion the merchant only provides the regular deal to consumers, while both the oBaz deal and the regular deal will be offered to consumers in the second case, Case J: Join oBaz Case N: N o oBaz P romotion We first consider the benchmark case where t he merchant does not offer a deal on oBaz and thus needs to solve the followin g optimization problem: ( 3 5) subject to ( 3 6) ( 3 7) ( 3 8) Eq. ( 3 5) function. The merchant makes a profit o f per unit of sale and the demand for the product is Inequalities ( 3 6) and ( 3 7) ensure that the demands of informed and uninformed PAGE 42 42 consumers are nonnegative. Finally, Inequality ( 3 8) requires that th e regular price be nonnegative. Solving the above problem leads to the following proposition P ROPOSITION 3 1. In the benchmark case of Case N: N o oBaz P romotion the optimal price and profit for the merchant are as follows: 1 and if 2 and if Proof. See the proof in the Appendix. Proposition 3 1 shows the s opt imal pricing policy in Case N where the informed consumer base is the key factor influencing how the merchant optimally price s the product. A higher proportion of informed consumer base leads the merchant to implement pricing po licy (1), where the merchant s demand comes from both informed and uninformed consumers. A lower proportion of informed consumer base results in policies (2). Pricing policy (2) takes effect in the case where the merchant derives demand from uninformed con sumers only. Case J : J oin oBaz In the first stage in Case J oBaz uses the number of signing up consumers, to negotiate with the merchant and then the merchant sends back a discounted price Every inform ed consumer needs to pay oBaz an extra activation fee for the deal and the discount price for the item. The demand of informed consumers in Case J will then be Hence, oBaz solve s the following optimization problem at the beginning PAGE 43 43 ( 3 9) subject to ( 3 10) ( 3 11) Eq. ( 3 9) describes oBaz profit function. Inequalities ( 3 10) ensure that the demand s of informed consumers are nonnegative while Inequalities ( 3 11) require that the regular price be nonnegative too. At the second stage, the merchant solves the optimization problem to join oBaz if choosing to join oBaz: ( 3 12) subject to ( 3 13) ( 3 14) ( 3 15) Eq. ( 3 12) profit function as well as the objective function. The merchant faces the demand of uninformed consumers and that of informed consumers while offering an oBaz deal. Each uninformed consumer pays for the item while each informed consumers pay Inequalities ( 3 13) and ( 3 14) ensure that the demands of informed and uninformed consumers are nonnegative. Finally, Inequality ( 3 15) is similar to Inequalities ( 3 8) and ( 3 11), to ensure the regular price be nonnegative. PAGE 44 44 Inequality ( 3 13) ensures the nonnegative demand of uninformed consumers. It also implies no uninformed consumers are willing to pay if the merchant charges the regular price greater than Similarly, Inequality ( 3 14), which is corresponding to the demand of informed consumers, implies that informed consumers are not willing to make the purchase if the merchant charges the regular price over For the merchant, one of these two inequalities is tighter. It thus leads to the following lemma. L EMMA 3 1. In Case J: Join oBaz if Inequality ( 3 13) is the tighter bound; otherwise, Inequality ( 3 14) is the tighter bound. Proof. See the proof in the Appendix. Lemma 3 1 sho ws that either Inequality ( 3 13) or ( 3 14) is the tighter bound. With low price sensitivity of informed consumers both informed and uninformed consumers are willing to purchase the products if they are charged less than Only informed consumers are willing to pay if the merchant prices the product between and and no consumers make the purchase if charged more than On the othe r hand, in case of high price sensitivity of informed consumers Inequality ( 3 14) is tighter than Inequality ( 3 13). With the price less than informed consumers are prepared to make the purchase and so are uninformed consumers. With the price between and u ninformed consumers are willing to pay for the product but informed consumers will not. No consumers are willing to pay for the price more than Lemma 3 1 reveals which constraint is required, and thus leads to the PAGE 45 45 following lemma that displays the timing that either informed, uninformed, or both informed and uninformed consumers are willing to pay. L EMMA 3 2. In Case J: Jo in oBaz 1. Both informed consumers are willing to pay if either and or and 2. Only informed consumers are willing to pay if and 3. Only uninformed consumers are willing to pay if and where and are the only real root s in (0,1) of and respectively and Proof. See the proof in the Appendix. Lemma 3 2 shows the timing for the merchant to entice informed and/or uninformed consumers to make the purc hase in Case J: Join oBaz Lemma 3 1 shows only one tighter bound either Inequality ( 3 13) or Inequality ( 3 14) is required. If informed consumers are more willing to pay, i.e., Inequality ( 3 13) is tighter then either inform ed consumers make the purchase or both informed and uninformed are willing to pay for the product. Under the condition that informed consumers are more willing, both the informed and uninformed are willing to purchase the product with the low proportion of informed consumers and only informed consumers make the purchase if the proportion of informed consumers is high Likewise, under the condition that uninformed consumers are more willing to pay ( ), both the informed and uninformed willingly make the purchase if PAGE 46 46 and only uninformed consumers are willing to pay if the proportion of informed consumers is less than Solving the ab ove optimization problem with Lemma 3 1 and Lemma 3 2 leads to the following proposition P ROPOSITION 3 2 In Case J : Join oBaz the pric ing strategy and the optimal profit for the merchant are as follows: 1. Both informed and uninformed consumers are willin g to pay: and 2. Only informed consumers are willing to pay: and 3. Only uninformed consumers are willing to pay: and where Proof. See the proof in the Appendix. Proposition 3 2 shows the merchant s optimal pricing policy in Case J : Join oBaz The merchant adopts pricing policy (1) if both informed and uninformed consumers are willing to make the purchase. Pricing policy (2) takes effect in the case where the merchant derives demand from informed consumers only, and Pricing policy (3) works in the case where only uninform ed consumers demand the product. Proposition 3, on the other hand, displays oBaz s pricing strategy in Case J : Join oBaz PAGE 47 47 P ROPOSITION 3 3 In Case J : Join oBaz the optimal profit for oBaz is if it is able to derive demand from i nformed consumers. The pricing strategy for oBaz is as follows: (1) both informed and uninformed consumers are willing to pay: oBaz charges and (2) only informed consumers are willing to pay: oBaz charges and the optimal profit for the merchant is same as that for oBaz, i.e., Proof. See the proof in the Appendix. Proposition 3 3 shows oBaz s optimal pricing strategy, and the optimal profit is proportional to the square of the fee and the price sensitivity of informed consumers. With respect to the marginal cost the optimal fee is decreasing as the marginal cost increases no matter whether uninforme d consumers are willing to pay or not ( and ) 5 The larger the marginal cost, the lesser the fee oBaz is able to charge and the lesser the profit oBaz can obtain. Likewise, the price sensitivity of inform ed consumers is the same as the marginal cost. It indicates that the optimal fee decreases in the price sensitivity of informed consumers ( and ). Considering the p roportion of informed consumers and display the positive relationship between the optimal fee and the proportion oBaz s optimal profit is thus increasing in the proportion of informed consumers 5 See Appendix for the details of Comparative Statics. PAGE 48 48 Case N V ersus Case J We further figure out the optimal profit and the optimal timing. The comparison under negligible marginal cost leads to Lemma 3 3 & 3 4. L EMMA 3 3. With negligible marginal cost, if where is the only real root in (0,1) of and Proof. See the proof in the Appendix. L EMMA 3 4. With negligible marginal cost, if where is the only real root in of and Proof. See the proof in the Appendix. Lemma 3 3 displays the comparison between Case N and Case J while both informed and uninformed consumers are willing to pay. With the demand from thes e two types of consumers, a low proportion of informed consumers leads the merchant to offer a more profitable deal on oBaz. By Proposition 3 1, we know that in Case N the demand is derived only from uninformed consumers if the proportion of informed consu mers is low, i.e., and the merchant obtains the optimal profit Then, Lemma 3 4 shows the comparison between and with negligible marginal cost, and is better with a high proportion of informed consumers. By Lemma 3 2, we know PAGE 49 49 that In Case J the demand is possibly derived either from informed consumers if and or from uninfo rmed consumers if and Lemma 3 1 determines who is more willing to pay an oBaz deal. Thus, Proposition 3 4 exhibits the comparison between Case N and Case J under the scenario that informed consumers are more willing to purchase an oBaz deal. P ROPOSITION 3 4. With negligible marginal cost and the following two conditions are satisfied, i.e., and the merchant is able to make a more profitable deal on oBa z for Proof. See the proof in the Appendix. Proposition 3 4 demonstrates that the optimal promotion timing for the merchant is based on the proportion of informed consumer s, and Figure 3 3 diagrams Proposition 3 4. The condition from Lemma 3 1 means that informed consumers are more willing to pay an oBaz deal. In Case N the tipping point is which means only the uninformed consumers are willing to pay if and both the informed and uninformed consumers are willing to pay if Likewise the tipping point in Case J is Case J outperforms Case N only as the proportion of informed consumers is intermediate under a scenario of negligible marginal cost With informed consumers high willingness to buy an oBaz deal, an intermediate proportion of informed consumer s leads the merchant to be more profitable in Case J On the other hand, a regular deal should be offered with either high or low proportion of informed consumers. PAGE 50 50 Figure 3 4 and Figure 3 5 demonstrate whether it is the merchant s best interest to provide an oBaz deal. Both Figure 3 4 and Figure 3 5 plot the benefit and the loss with an oBaz deal. T he merchant s optimal strategy is the case of No oBaz when the optimal price is far lower than the reservation price of informed consumers (Figure 3 4) and the case of Join oBaz when the optimal price approaches informed consumers maxi mum willingness to pay (Figure 3 5). The line with price elasticity of demand means the demand curve for the output of the merchant offering a regular deal. Likewise, the line with represents the demand of informed consumers with an oBaz deal, and that with is for uninformed consumers. In Figure 3 4, if the merchant joins oBaz and charges as the regular price, the quantity demanded moves from A to B, which incurs a loss since informed consumers, informed by the oBaz promotions, would not pay at a regular price (Area ABFG). Informed consumers who would make the purchase pay for the product because oB az already negotiate a discounted price with the merchant. However, the informed consumers realize that the actual price is the discounted price plus the activation fee. The actual demand of informed consumers locates not in Q but in D. Hence, the profit f rom informed consumers is area OHMN and the merchant suffers a loss, MNSQ. In Figure 3 4, the benefit cannot cover the loss since area OHMN is smaller than the sum of area MNSQ and ABFG. On the other hand, Figure 3 5 exhibits the merchant s best strategy i s oBaz. The benefit (OHMN) is large enough to compensate for the loss (MNSQ and ABFG). PAGE 51 51 Empirical Analyses To obtain further insights into the merchant s optimal strategy on oBaz, we conducted numerical analyses with empirical data. Proposition 3 4 only sho ws the finding under the scenario of negligible marginal cost, and the numerical analyses further estimate the comparison with marginal cost. We collected oBaz promotion data from September 2011 to May 2012. For the baseline values in our model, the price sensitivity of informed consumers is set as and the range is between 1 and 4 because of the empirical study by Tellis ( 1988 ) The proportion of informed consumers is set .35 as the baseline value. We explore the impact of these key pa rameters, including (1) the proportion of informed consumers, and (2) the price sensitivity parameter of informed consumers, on the profitability of the C ase N (No oBaz promotion) and the Case J (Join o Baz). Table 3 2 shows the baseline values of parameters in our model. Our numerical analyses results are shown in Figures 3 6 and 3 7. Figure 3 6 shows how the proportion of informed consumers influences the merchant s optimal profit for both cases. The me rchant can make better profits by providing oBaz promotions if the proportion of informed consumers is between ; otherwise, using oBaz to promote the products is not as beneficial as selling by the merchants themselves. Too few or too many informed consumers cannot guarantee that the merchant will make more profits by oBaz. The merchant is encouraged to run the oBaz promotion only in the case of an intermediate amount of informed consumers, and the finding is consistent with Propos ition 3 4. PAGE 52 52 Figure 3 7 plots the merchant s optimal profit with respect to the price sensitivities of informed consumers. Considering the price sensitivity of informed consumers, the merchant s optimal strategy is to provide an oBaz deal when the informed c onsumer price sensitivity is high enough, and to provide a regular deal with low price sensitivity of the informed consumers. If the price sensitivity of informed consumers becomes high, the oBaz segment of consumers becomes significant for the merchant. Since only informed consumers know the oBaz promotion, oBaz can attract more informed consumers when their price sensitivity is high enough. Summary In this paper, we developed and tested a set of models designed to examine whether the merchant is able to offer the products on oBaz or not. The oBaz website provides a new business model in social commerce. Our focus has investigated on the effect of oBaz on the merchant and consumers. Our results conclude that it is workable for the merchant to provide a de al on oBaz. Our findings imply that oBaz is most profitable as an intermediate proportion of informed consumers (Figure 3 3 and Figure 3 6). T he merchant is unable to drive the demand of informed consumers when the proportion of informed consumers is low. Then oBaz is not a good option for the merchant to increase demand and profit. On the other hand, the high proportion of informed consumers leads to high discounts. The merchant is required to provide a higher discount on oBaz for informed consumers. It m ay drive more informed consumers to make a purchase, but it reduces the profit dramatically. Our results also indicate how oBaz charges activation fees. If oBaz is able to drive demand of informed consumers, oBaz s profit is proportional to the square of t he PAGE 53 53 activation fee. Furthermore, the merchant s profit is identical to oBaz s profit if only informed consumers are willing to purchase the product. Hence, if oBaz charges the fee too high, the merchant and oBaz is unable to make any profit from the promoti onal deal. We only consider the linear quantity discount, but it may have different results if other discount policies such as all units quantity discount or incremental units quantity discount. A worthy future research direction will be to explore differe nt discount policies. PAGE 54 54 Table 3 1. Summary of notation Parameter Description Regular price per unit P roportion of informed consumers P rice sensi tivit y of the informed consumers F ee paid by the merchant to oBaz Marginal cost Table 3 2 Baseline parameters for numerical analysis Parameter Baseline values Range of value P roportion of informed consumers P rice sensitivit y of informed consumers Marginal co st Figure 3 1 An example of the Bag it game (reprinted from http://www.obaz.com) PAGE 55 55 Figure 3 2 An example of the oBaz discount (reprin ted from http://www.obaz.com ) Figure 3 3 Optimal profit under negligible marginal cost versus proportion of informed consumers ( ) PAGE 56 56 Figure 3 4 Merchant s benefit and loss when the optimal price is lower than the reservation price of informed consumers Figure 3 5 Merchant s benefit and loss when t he optimal price approaches the reservation price of informed consumers PAGE 57 57 Figure 3 6 Optimal profit versus proportion of informed consumers ( ) Figure 3 7 Optimal profit versus price sensiti vity of informed consumers ( ) PAGE 58 58 CHAPTER 4 A STUDY OF NEW E COMMERCE BUSINESS MODELS BASED ON CONSUMER S PRESENT BIAS StickK and Gym Pact A rebate is a type of promotion used as incentives to sales. It is required for the consume rs to make the purchase first, and then a rebate is redeemed by way of reduction or refund on what has already been paid. The sellers are able to earn profits from those consumers who make the purchase without redeeming their rebates. StickK 1 is a way to a chieve your goals. Stick is for stick to it and the second K is for the contract. P eople would get better motivated to meet their goals if they had something to lose, Dean Karlan, Yale Economics Professor and co founder & president of StickK.com explains At StickK consumers put the money on the line, and make a commitment. If they achieve their goals, they get the money back. If they fail, they lose the money. For example, my goal is to lose 1 pound a week for 20 weeks. I make a commitment to losing $100 for every week I fail and put up $2000. If I achieve the goal lose 20 pounds in 20 weeks I am able to win $2000 back; otherwise, StickK wins $2000 ( Sittenfeld 2008 ) Gym Pact 2 founded by a 2010 graduate of Harvard University, offers motivation fees customers agree to pay more if th ey miss their schedule workouts Gym Pact negotiated a group rate with a gym and then paid the membership fees for participants, who in return for a free membership agreed to work out at least four times per week. If they fail to follow the schedule in any one week, the participants pay $25, the regular 1 http://www.stickk.com/ 2 http://www.gym pact.com PAGE 59 59 membership fee is only $30 per month. If they leave the program for any reason other than injur y or illness, they will pay $75 ( Bernard 20 12 ; Johnston 2011 ) StickK works like a rebate consumers pay first, redeem their rebates and get the money back later. Likewise, Gym Pact is another rebate model consumers join for free but will pay for the penalties if failing to follow the workout schedule. StickK and Gym Pact provide the new business models for retailers. Are the new models more profitable than the traditional retail model? Our research aims to examine whether it is in the retaile s best interests to offer such a rebate as StickK or Gym Pact. If yes, what is the optimal price for retailers to charge their consumers? Literature Review According to present biased preferences individuals give more weight to immediate moment when con sidering a trade off betwe en the immediate moment and any future moment A simple model is described as ( 4 1) where is the instantaneous utility a person receives at time an d represents his/her intertemporal preference at time The parameter represents long run, time consistent discounting, and possible time inconsistent preference for present peri od utility compared to all future periods is represented by Anyone who has a would be present biased ( O'Donoghue and Rabin 1999 ) Gilpatric (2009 ) proposes the model of slippage in rebate programs based on the theory of present biased preferences. The timeline for a rebate program is usually as follows: at time 1 consumers decide whether to make the purchase, at time 2 consumers make the decision as to whether to redeem the rebate, and at time 3 rebate PAGE 60 60 checks are delivered to consumers if the rebates were submi tted. Consumers perceive the difference between the timing of purchase and the timing of redemption when choosing a deal associated with the rebate program. Slippage usually happens in rebate programs if consumers fail to redeem the rebate after making a p urchase. Models and Analyses Following the model setup in Gilpatric (2009 ) the total number of consumers is normalized t o 1, and the time consistent discounting parameter is simplified as 1. A consumer values a service at and let be the value of the refund that a consumer can obtain after the r ebate submission. The consumer makes the purchase at time 1, and then it will cost this consumer the effort cost to redeem a rebate at time 2. Table 4 1 summarizes the notation used in the paper. The model includes two types of players, a retailer and its consumers. In a market where a rebate is offered, the retailer earns profits from those naively present biased consumers, who mistakenly believe that they will redeem the rebates after the purchase but actually they will not bec ause they do not expect the present biased preferences. For time consistent consumers or present biased but sophisticated consumers, they realize their future preferences and predict their future actions correctly when choosing w hether to make the purchase or not. At time 1 consumer s purchase decisions are influenced by how consumers believe they will redeem the rebate later. At time 2 consumers decide to redeem the rebate if they perceive the discounted refund value greater tha n their effort cost: ( 4 2) PAGE 61 61 However, n ave consumers at time 1 believe that they will redeem the rebate later if the refund value simply exceeds the effort cost: ( 4 3) Hence, a rebate offer that sati sfies the following condition, ( 4 4) will lead the nave buyers to believe that they will redeem the rebate at time 2 when making the purchase but they will actually fail at time 2. Let the retailer charge consumer s the regula r price The following condition will lead consumers to purchase the service at time 1: ( 4 5) Consumers have different various degrees of present bias. The cumulative function is used to indicate the present biased preferences in the population, and is the associated probability density function. Model R : S tandard R etailer M odel We first consider the benchmark model where no rebate is offered. Cons umers pay the regular price for the service. Note that the retailer offers no rebate under the standard retailer model, and the value of refund is zero ( ). The rebate condition is not considered because no consumer will redeem a rebate. Considering the purchase condition, consumers will purchase the service at time 1 if: ( 4 6) The retailer chooses the regular price and such choices will determine the proportion that purchases the service. Consumers will purchase at time 1 if PAGE 62 62 by rearranging Eq. ( 4 6), and the proportion who make the purchase are (see Figure 4 1 ). Under the standard retailer model th e retailer determines the regular price to maximize the profits: ( 4 7) where the retailer pay the wholesale price for each consumer, and is the pro fit per unit sale. The service value is greater than the wholesale price, i.e., If the service value is lower than the wholesale price, which implies then no consumers are willing to pay for the service Hence, the retailer is always making no profit and unwilling to offer the service. Let present biased preference follow a uniform distribution then the optimization problem is solved as Proposition 1. P ROPOSITION 4 1. Under the standard retailer busines s model the retailer sets the optimal price as and the optimal profit is if present biased preference follows a uniform distribution Proof. See the proof in the Appendix. Prop osition 4 1 shows the optimal price and the optimal profit for the retailer under the standard retailer model with the uniform present biased preference. We discuss the StickK business model and the Gym Pact Business model in the following sections. Model S : StickK B usiness M odel Under the StickK business model, at time 1 consumers puts up the wager and at time 2 they either take back or lose it. According to the theory of mental PAGE 63 63 accounting, the loss fun ction is steeper than the gain function ( Thaler 1985 ) With something to lose people perceive to pay more than what they actually pay. T he par ameter is introduced to capture mental accounting. If consumers lose at time 2, they perceive to lose In other words, consumers view the value of the refund as At time 1 consu mers who believe they will win the wager and purchase the service if the value of the service plus the discount refund value net of effort cost exceeds the price: ( 4 8) At time 2 consumers choose to take the wager back if the di scounted refund value exceeds the effort cost: ( 4 9) Figure 4 4 shows the number of participants and the number of participants who win the bet. The number of participants is obtained from Eq. ( 4 8), a nd the number of winners is which comes from Eq. ( 4 9). Note that under the StickK business model the retailer makes profits only from the participants who lose the wager, and the number of losers is the number of participants m inus the number of winners Hence, the retailer s profit function is or ( 4 10) Since each individual consumer chooses to put up the wager under the StickK business model, we thus have following p roposition by solving the problem. PAGE 64 64 P ROPOSITION 4 2 Under the StickK business model every consumer participates, puts up but no one is able to win the wager. The retailer s optimal profit is Proof. See the proof in the Appendix. Proposition 4 2 shows the optimal wager that each consumer post and the optimal profit for the retailer. Each consumer seems to take the best option for himself /herself and mistakenly believes s/he is able t o win the wager. However, the truth is that no one can get the wager back and the retailer is the only winner. We keeping considering another situation that the wager is posted not by each consumer but by the retailer. The retailer solves the following max imization problem if the retailer is able to choose the wager: ( 4 11) subject to ( 4 12) ( 4 13) ( 4 14) ( 4 15) Eq. ( 4 11) describes the retailer s objective function under the StickK business model. Inequality ( 4 12) ensures that the number of participants is no less than the number of winners. Inequality ( 4 13) ensures the number of participants is no more than PAGE 65 65 the total number of consumers. Inequality ( 4 14) guarantees that the number of winners is nonnegative. Inequality ( 4 15) requires that the wager has to be nonnegative. Let present biased preference follow a uniform distribution then Pro position 4 3 exhibits same result as Proposition 4 2. The optimal profit is unchanged if the wager is chosen either by the retailer or by consumers. P ROPOSITION 4 3 Under the scenario where the retailer prices the service, the retailer is able to use the StickK business model to win the optimal profit if 1. consumers perceive the service value is greater than the effort cost 2. the wholesale price is greater than the effort cost Proof. See the proof in the Appendix. Proposition 4 2 and 4 3 show that the retailer makes the same profit either if the wager is posted by consumers or by the retailer. At first glance, consumers are able to take the advantage by setting the wager themsel ves. However, there is no difference for the retailer between these two cases the wager decided by the retailer or by consumers. Model GP : Gym Pact B usiness M odel Under the Gym Pact business model, at time 1 consumers join the gym for free, and at time 2 they either keep the free membership or pay the penalty Likewise, T he parameter captures mental accounting, and consumers perceive the penalty to be more than what they actually pay if they fail to fol low the workout schedule. They feel like to pay for the penalty. Consumers are not required to pay for the membership if they follow the workout schedule. Hence, consumers value the refund as the membership fee At time 1 consumers who believe they will not be penalized and PAGE 66 66 participate in the workout program for free if the value of the service plus the discounted refund value (membership fee) net of effort cost exceeds the pe nalty: ( 4 16) At time 2 consumers who will follow the workout schedule if the discounted value of refund (membership fee) exceeds the immediate cost of effort: ( 4 17) Figure 4 3 shows the number of pa rticipants and the number of participants who are penalized (lose the bet). Under the Gym Pact business model the retailer makes profits only from the participants who fail to follow the schedule the number of losers is the number of participants minus th e number of participants who are not penalized Hence, the retailer s profit function is or ( 4 18) The retailer solves the following maximization problem under the Gym Pact business model: ( 4 19) s ubject to ( 4 20) ( 4 21) PAGE 67 67 ( 4 22) Eq. ( 4 19) describes the retailer s objective function under the Gym Pact business mode l. Inequality ( 4 20) ensures that the number of participants is no less than the number of winners. Inequality ( 4 21) ensures the number of participants is no more than the total number of consumers. Inequality ( 4 22) requires that the penalty is nonnegative. Inequality ( 4 17) and implies and thus the number of winners is definitely nonnegative. Let the present biased preference follows a uniform distribution, then Proposition 4 4 shows the result under the Gym Pact business model. P ROPOSITION 4 4 Under the Gym Pact business model, the retailer sets the optimal penalty as 1. and the optimal profit is if 2. and the optimal profit is if Proof. See the proof in the Appendix. Proposition 4 4 exhibits the pricing policies under the Gym Pact business model with uniform pres ent biased preference. T he service value is the key factor influencing how the retailer optimally price s the service. An intermediate service value leads the retailer to implement pricing policy (1). A higher service value result s in policies (2). Pricing policy (2) takes effect in the case where the retailer derives demand PAGE 68 68 from all the consumers. The low service value causes the consumers who make the purchase redeem their rebates and thus the retailer is unable to make any profi t form the consumers. Comparisons among S tandard R etailer M odel, S tickK M odel and Gym Pact M odel We further examine these three business models in the section and attempt to answer our research question: Is it the retailer s best interests to offer a Stick K like or Gym Pact rebate? For the retailer, under what condition is the StickK business model or the Gym Pact business model is more profitable than the standard retailer model? First we compare the StickK business model with the standard business model a nd find the insights in Proposition 4 5. PROPOSITION 4 5 For the retailer, StickK business model is more beneficial than the standard retailer model if the condition is satisfied. Proof. See the proof in the Appendix. Propositio n 4 5 exhibits that the StickK business model is able to bring the retailer more profits under the condition We further find and the effort cost makes the condition easily be satisfied. The higher effort cost leads the retailer to choose the StickK business model. First, under the StickK business model consumers have the opportunity to win their money back, and it becomes an incentive to drive consumer demand even if the effort cost is high. Then the high effort cost leads consumers to lose their money and the retailer thus makes profits from the consumers who lose. For the retailer, the high effort cost is the optimal condition to run the StickK business model. PAGE 69 69 Next, we compar e the Gym Pact business model with the standard business model and indicate the optimal timing for the Gym Pact business model. By Proposition 4, under the Gym Pact business model if only some of consumers join Gym Pact and part of these participants win t he bet, the optimal profit is Let the profit difference be then Lemma 1 shows the behavior of L EMMA 4 1 The function is 1. a strictly decreasing funct ion of and has only one real root where 2. a strictly increasing function of and has only one real root where Proof. See the proof in the Appendix. Lemma 4 1 exhibits the behavior of The function for and for With respe ct to the effort cost is positive for otherwise, the function is negative for We thus have proposition 6 to further indica te the optimal condition for the Gym Pact business model. P ROPOSITION 4 6 If some consumers buy the service and part of them redeem, then the Gym Pact business model is better than the standard retailer model if 1. the membership fee, is less than 2. the effort cost, is greater than and Proof. See the proof in the Appendix. The lower membership fee and the higher effort cost le ad the retailer to choose the Gym Pact business mode. T he low membership fee increases not only the number of participants but also the number of participants who follow the workout schedule. But the increase rate of the buyer size is far larger than that of the buyers who follow as the PAGE 70 70 membership fee is low. In other words, under the Gym Pact business model the low membership fee also enlarges the size of the buyers who fail, and the retailer is able to make profit from these buyers who fail. With respect to the effort cost, the Gym Pact business model is similar to the StickK business model. The high effort cost decreases the number of purchase under the standard retailer model and increases the possibility for the buyers to be penalized under the Gym Pact business model. Summary StickK and Gym Pact provide two new business models in rebate programs. Our research examines the feasibility of these new models and offers retailers the managerial insights. First of all, traditional retailers should adopt the St ickK business model or the Gym Pact business model if consumers are perceiving high effort cost. These two models create an incentive for consumers to win the money back or join for free. However, most consumers are unable to take the money back due to the high effort cost, and thus retailers are able to make profits. Secondly, retailers prefer the Gym Pact business model especially when the membership fee is low. The low membership fee decreases the number of participants and the number of losers. However, the decreasing rate of the number of participants is much higher than that of the number of the losers. The Gym Pact business model is profitable for retailers and able to make retailers earn better profits than the standard retailer model. But there are still some limitations in the study. First, the comparison is based on the uniform present biased preferences. The study let the presented biased preferences follow the uniform distribution. It may lead to different results with a normal distribution or o ther distributions. Secondly, the study only focuses on the service industries, which PAGE 71 71 have some unique characteristics such as no tangible goods. It may lead to some different results in other industries such as restaurants or grocery stores. PAGE 72 72 Table 4 1. Summary of notation Parameter Description Regular price under the standard retailer model Wager paid by a consumer under StickK business model Membership fee for a consumer to join the service A consumer values the service captures mental accounting captures a bias for the present and foll ows a distribution with probability density function Wholesale price paid by the retailer for each consumer Value of the refund Intertemporal preference at time Instantaneous utility a person receives at time Long r un, time consistent discounting Figure 4 1 The standard retail er model PAGE 73 73 Figure 4 2 The StickK business model Figure 4 3 The Gym Pact business model PAGE 74 74 C HAPTER 5 CONCLUSION In this dissertation, I investigated the new electronic commerce models. Specially, I concentrated on (1) social commerce; (2) consumer generated group deals and (3) a new rebate program. In Chapter 2, Groupon and LivingSocial are two most notable social commerce websites that take advantage of the transformative effect of social media on businesses. As one of the hottest sectors in e commerce, these social commerce websites offer consumers a significant reduction in the price of produ cts or services when the group of buyers reaches a required minimum number We find that offering the same deal in Groupon will lead to a higher profit than in LivingSocial in the current U.S. market where the proportion of informed consumers is less than 30 percent of the consumer population. In Chapter 3, t he oBaz model is a new E Commerce model that provides consumers user generated group deals and consumers are able to choose their favorite products. We employ an analytical model to study the model and the key result shows that it is workable for merchants to run the model only if the proportion of consumers who are informed by oBaz is intermediate. In Chapter 4, the StickK and Gym Pact models are based on be tting on c onsumer w agers We use a traditiona l retail model as the benchmark and find that these two models allow consumers to have an opportunity to win their money back. It becomes an incentive to drive consumer demand. However, most consumers are unable to take the money back, especially when the effort cost is high, and thus retailers benefit from the consumers who lose money. PAGE 75 75 This dissertation is organized as a collection of articles, each of which corresponds to one chapter of entire study. Each chapter is complete within itself and includes a c onclusion and future work section related with the aspects of the study covered in that specific chapter. Due to this self contained style of preparation, I ask the readers to refer to those sections for a detailed description of possible future works. PAGE 76 76 APPENDIX A PROOF OF LEMMAS AND PROPOSI T I ONS Proposition 2 1 Firstly note that the informed consumers have a lower reservation price than that of u ninformed consumers, i.e. Inequality (3), hence, implies and we only need to consider the constraint to ensure the nonnegative demands of both uninformed and informed consumers. Case (1) the merchant would consider charging both informed and uninformed consumers, t hen the Lagrange f unction is The Kuhn Tucker conditions are Subcase ( 1 a ) if and then from the first KKT condition, we have which implies We have from and from PAGE 77 77 Finally, we ha ve the optimal profit Subcase ( 1 b ) if and implies then we have Plugging and back into the first KKT condition, we have which in turn implies Furthermore, we have from Finally, we have the optimal profit also means no informed customer demand. We keep considering the case of no informed customer demand in the following section. Case (2) the merchant would pay a ttention to uninformed consumers It means the merchant sets the price too high to attract informed consumers to make the purchase, i.e., We set the value of the informed customer demand to zero to avoid the negative demand of in formed customers. Then the merchant only faces uninformed customers, and the optimization problem will turn out to be PAGE 78 78 S ubject to The Lagrange functi on becomes The Kuhn Tucker conditions are Subcase ( 2a ) if Since hence, fr o m the first KKT condition, we have which implies We have from And note that in this case, and we thus have by plugging back in to Finally, we have the optimal profit Subcase ( 2b ) if PAGE 79 79 implies then we have Plu gging back into the first KKT condition, we have which in turn implies And note that in this case, and we thus have by plugging back in to Finally, we have the optimal profit Proposition 2 2 & 2 3 Similarly, implies and W e only need to consider the constraint Further note that implies and implies Since hence, In other words, implies PAGE 80 80 Hence, we consider the constraint to ensure that the consumer demand will not exceed the c apacity. We need to consider the following constraints when solving the problem: and Case (1) the merchant would consider charging both informed amd uninformed consumers, t he n the Lagrange function is The Kuhn Tucker conditions are PAGE 81 81 Note that we have from the second KKT condition ; thus the required minimum number of purchases, is optimal as long as Subcase ( 1 a ) if and then from the first KKT condition, we have which implies W e have from and from We also have that from Finally, we have the optimal profit PAGE 82 82 Subcase ( 1 b ) if and implies then we have Plugging and back into the first KKT condition, we have which in turn implies Furthermore, we have from We also have that from Finally, we have the optimal pr ofit PAGE 83 83 also means no informed customer demand in the second period. We keep considering the case of no informed customer demand in the second perio d in the following cases. Case (2) the merchant would pay attention to uninformed consumers, and i.e., It means the merchant sets the price too high to attract informed consumers in the second p e riod t o make the purchase, i.e., We set the value of the informed customer demand in the second period to zero to avoid the negative demand of informed customers. Then the optimization problem will turn out to be S ubject to PAGE 84 84 Note that in this case, hence, implies It is also clear that implies Hence, we consider these three constraints: and T hen the Lagrange function becomes The Kuhn Tucker conditions are PAGE 85 85 Note that we have from the second KKT condition ; thus the required minimum number of purchases, is optimal as long as Subcase ( 2 a ) if and then from the first KKT condition, we have which implies We have from and from We also have that from And note that in this case, and we thus have by plugging back in to Finally, we have the optimal profit PAGE 86 86 Subcase ( 2b ) if and implies then we have Plugging and back into t he first KKT condition, we have which in turn implies Furth ermore, we have from We also hav e that from PAGE 87 87 And note that in this case, and we thus have by plugging back in to Finally, we ha ve the optimal profit Now means no informed customer demand in the first period. We dicuss it in S ubcase s (2c) and (2d). Similarly, the problem turns out to be S ubject to The Lagrange function becomes The Kuhn Tucker conditions are PAGE 88 88 Subcas e ( 2c ) if Since hence, fr o m the first KKT condition, we have which implies We have from And no te that in this case, and we thus have by plugging back in to Finally, we have the optimal profit Subcase ( 2d ) if implies then we have Plugging back into the first KKT condition, we have which in turn implies And note that in this case, and we thus have by plugging back in to Finally, we have the optimal pro fit PAGE 89 89 Case (3) the merchant would pay attention to uninformed consumers, and i.e., It means the merchant sets the price too high to attract informed consumers in ther second p e ri od to make the purchase, i.e., We set the value of the informed customer demand in the second period to zero to avoid the negative demand of informed customers. Then t he optimization problem will turn out to be S ubject to Note that in this case, hence, implies It is also clear that implies Hence, we consider these three constraints: and T hen the Lagrange function becomes PAGE 90 90 The Kuhn Tucker conditions are Note that we have from the second KKT condition ; thus the required minimum number of purchases, is optimal as long as Subcase ( 3 a ) if and then from the first KKT condition, we have which implies PAGE 91 91 We have from and from We also have that from And note that in this case, and we thus have by plugging back in to Finally, we have the optimal profit Subcase ( 3b ) if and implies then we have PAGE 92 92 Plugging and back int o the first KKT condition, we have which in turn implies Fur th ermore, we have from We also have that from And note that in this case, and we thus have by plugging back in to Finally, we have the optimal profit Now means no un informed customer demand in the first period and the second period. We di s cuss it in S ubcase s (3c) and (3d). PAGE 93 93 Similarly, the problem tu rns out to be S ubject to The Lagrange function becomes The Kuhn Tucker conditions are Note that we have from the second KKT condition ; thus the required minimum number of purchases, is optimal as long as Subcase ( 3c ) if and PAGE 94 94 Since hence, frm the first KKT condition, we have which implies We have from We also have that from And note that in this case, and we thus have by plugging back in to Finally, we have the optimal profit Subcase ( 3d ) if and implies then we have Plugging and back into the first KKT condition, we have which in turn implies We also have that from PAGE 95 95 And note that in this case, and we thus have by plugging back in to Finally, we have the optimal profit For all case s (1) (3), we find that t he optimal price and the optimal profit are independent of and proving Proposition 2. Prop osition 2 3 summarizes the optimal prices and the optimal profits of cases (1) (3). Proposition 2 4 Similarly, implies and we only need to consider the constraint Further note that implies and implies Hence, In other words, implies Hence, we consider the constraint to ensure that the consumer demand will not exceed the capacity. We need to consider the following constraints when solving the problem: PAGE 96 96 and Case (1) the merchant would consider charging both informed and uninformed consumers, t hen the Lagrange function is The Kuhn Tucker conditions are Note that we have from the second KKT condition ; thus the required minimum number of pu r chases, is optimal as long as Subcase ( 1 a ) if and then from the firs t KKT condition, we have which imp lies PAGE 97 97 We have from and from We also have from Finally, we have th e optimal profit Subcase ( 1 b ) if and implies then we have PAGE 98 9 8 Plugging and back into the first KKT condit ion, we have which in turn implies Furthermore, we have from We also have that from Finally, we have the optimal profit also means no informed customer demand in the first period. We keep considering the case of no informed customer demand in the first period in the following cases. PAGE 99 99 Case (2) the merchant would pay attention to uninformed consumers, and i.e., It means the merchant set s the price too high to attract informed consumers in the first p e riod to make the purchase, i.e., We set the value of the informed customer demand in the first period to zero to avoid the negative demand of informed customers. Then the optimization problem will turn out to be S ubject to Note that in this case, hence, implies It is also clear that implies PAGE 100 100 Hence, we consider these three constraints: and T hen the Lagrange function becomes The Kuhn Tucker conditions are Note that we have from the second KKT condition ; thus the required minimum number of pu r chases, is optimal as long as Subcase ( 2 a ) if and then from the first KKT condition, we have PAGE 101 101 which implies We have from and from We also have that from And note that in this case, and we thus have by plugging back in to Finally, we have the optimal profit Subcase ( 2b ) if and PAGE 102 102 implies then we have Plugging and back into the first KKT condition, we have which in turn implies Furth ermore, we have from We also have that from And note that in this case, and we thus have by plugging back in to Finally, we have the optimal profit PAGE 103 103 Now means no informed customer demand in the second per iod. We dicuss it in S ubcase s (2c) and (2d). Similarly, the problem turns out to be S ubject to The Lagrange function becomes The Kuhn Tucker conditions are Subcase ( 2c ) if Since hence, fr o m the first KKT condition, we have which implies We have from PAGE 104 104 And note that in this case, and we thus have by plugging back in to Finally, we have the optimal profit Subcase ( 2d ) if implies then we have Plugging back into the first KKT condition, we have which in turn implies And note that in this case, and we thus have by plugging back in to Finally, we have the optimal profit Case (3) the merchant would pay attention to uninformed consumers, and i.e., It means the merchant sets the price too high to attract informed consumers in the first p e riod to make the purchase, i.e., We set the value of the informed PAGE 105 105 customer demand in the second period to zero to avoid the negative de mand of informed customers. Then the optimization problem will turn out to be S ubject to Note that in this case, hence, implies It is also clear that implies Hence, we consider these three constraints: and T hen the Lagrange function becomes PAGE 106 106 The Kuhn Tucker conditions are Note that we have from the second KKT condition ; thus the required minimum number of pu r chases, is optimal as long as Subcase ( 3 a ) if and then from the first KKT condition, we have which implies We have from and from We also have that from PAGE 107 107 And note that in this case, and we thus have by plugging back in to Finally, we have the optimal profit Subcase ( 3b ) if and implies then we have Plugging and back into the first KKT condition, we have which in turn implies PAGE 108 108 Furth ermore, we have from We also have that from And note that in this case, and we thus have by plugging back in to Finally, we have the optimal profit Now means no un informed customer demand in the first p eriod and the second period. We dicuss it in S ubcase s (3c) and (3d). Similarly, the problem turns out to be S ubject to PAGE 109 109 T he Lagrange function becomes The Kuhn Tucker conditions are Note that we have from the secon d KKT condition ; thus the required minimum number of puchases, is optimal as long as Subcase ( 3c ) if and Since hence, fr m the first KKT condition, we have which implies We have from PAGE 110 110 We also have that from And note that in this case, and we thus have by plugging back in to Finally, we have the optimal profit Subcase ( 3d ) if and implies then we have Plugging and back into the first KKT condition, we h ave which in turn implies We also have that from And note that in this case, and we thus have by plugging back in to Finally, we have the optimal profit Proposition 2 4 summarizes the optimal prices and profits of cases (1) (3). PAGE 111 111 Proposition 3 1 First note that the informed consumers have a lower reservation price than that of uninformed consumers H ence, implies and we only need to consider the constraint to ensu re the nonnegative demands of both uninformed and informed consumers. Case (1) : T he merchant consider s charging both informed and uninformed consumers. T hen the Lagrang ean function is The Kuhn Tucker conditions are If then from the first KKT condition, we have which implies We have from Finall y, we have the optimal profit If no informed consumers would make the purchase. The problem facing the merchant becomes Case (2). Case (2) : The merchant only faces uninformed consumers, and the optimiz ation problem is as follows. PAGE 112 112 Subject to The Lagrange an function becomes The Kuhn Tucker conditions are If then we have from the first KKT condition We have from and this condition is always satisfied by Eq. (4). Note that the merchant only faces uninformed consumers. Hence, means no informed consumers make the purchase. Finally, we have the optimal profit means that the no uninformed consumers would make the pur chase. Then the merchant make no profits since no informed and uninformed consumers would make the purchase Lemma 3 1 Note that oBaz and the merchant face the same informed consumers, which means Eq. ( 3 10) and Eq. ( 3 14) represent the same informed consu mers demand. Hence, the demand constraint Eq. ( 3 10) will be considered in the merchant s optimization problem. At the first stage we solve oBaz s optimization problem as an unconstraint problem: PAGE 113 113 Then we have From Inequality ( 3 13), we have From Inequality ( 3 14), since hence, we have Note that If which implies then Inequality ( 3 13) is the tighter bound; otherwise, Inequality ( 3 14) is the tighter bound. Lemma 3 2 Note that in Case J: Join oBaz the merchant and oBaz solve the optimization problems simultaneously, and we have from oBaz s optimization problem. At the second stage, the merchant solves the optimization problem (Eq. ( 3 12) ( 3 15)). By Lemma 3 1, if then Inequality ( 3 13) is the tighter bound; otherw ise, Inequality ( 3 14) is the tighter bound. Case (1) the constraint either Inequality ( 3 13) or ( 3 14) is not biding, which means both informed and uninformed consumers are willing to pay. Subcase (1a) if Inequality ( 3 13) is the tighter bound [ ], then PAGE 114 114 The Lagrange an function becomes The Kuhn Tucker conditions are If then we have from the f irst KKT condition. Note the merchant and oBaz solve the optimization problems simultaneously, and and we thus have and we have from Inequality ( 3 13). Let Now show that only if where is the only real root in (0,1) of PAGE 115 115 Hence, is strictly decreasing function. Also note that since and Hence, has the only real root and we thus have for and for Then only if i.e., Hence, both informed and uninformed consumers are willing to buy only if Note that means no demand from uninfo rmed consumers and only informed consumers are willing to pay and it will be discussed in Case (2). Subcase (1b) if Inequality ( 3 14) is the tighter bound [ ], then The Lagrange an function becomes The Kuhn Tucker conditions are If then we have from the first KKT condition. PAGE 116 116 Note the merchant and oBaz solve the optimizatio n problems simultaneously, and and we thus have and we have from Inequality ( 3 14). Let Now sho w that only if where is the only real root in (0,1) of Hence, is strictly increasing function. Also note that and Hence, has the only real root and we thus have for and for Then only if i.e., Hence, both informed and uninformed consumers are willing to buy only if Note that means no demand from informed consumers and only un informed consumers are willing to pay and it will be discussed in Case (3). By Subcase (1a) and (1b), we thus have both informed and uninformed consumers are willing to buy if either and or and PAGE 117 117 Case (2) if only informed consumers are willing to pay, then the merchant s optimization problem becomes Subjec t to The Lagrange an function becomes The Kuhn Tucker conditions are Note that the merchant and oBaz solves the optimizati on problems simultaneously and If we have and from the first KKT condition, and from which is always satisfied by Eq. ( 3 4). We have that only informed consumers are willing to pay if Inequality ( 3 13) is the tighter constraint [ ] and biding [ ]. Case (3) if only uninformed consumers are willing to pay, then the merchant s optimization problem becomes Subject to PAGE 118 118 The Lagrange an function becomes The Kuhn Tucker conditions are If we have from the first KKT condition and from which is always satisfied by Eq. ( 3 4). We have that only uninf ormed consumers are willing to pay if Inequality ( 3 14) is the tighter constraint [ ] and biding [ ]. Proposition 3 2 Case (1): if both informed and uninformed consumers are willing to pay, then from the pr oof of Lemma 3 2, we have and in either Subcase (1a) or (1b). The merchant s optimal profit is where PAGE 119 119 Case (2) if only info rmed consumers are willing to pay, then from the proof of Lemma 3 2, we have and The merchant s optimal profit is Case (3) if only un informed consumers are willing to pay, the n from the proof of Lemma 3 2, we have T he merchant s optimal profit i s Proposition 3 3 Case (1) If both informed and uninformed consumers are willing to pay, then from the proof of Lemma 3 2, we have Then oBaz s optimal profit is Case (2) If only informed consumers are willing to pay, then from the proof of Lemma 3 2, we have Then oBaz s optimal profit is PAGE 120 120 Also note that We thus have Comparative Statics 1. With respect to and 2. With respect to implies then we have 3. W ith respect to implies then we have PAGE 121 121 Lemma 3 3 Note that Let Then Show that if where is the only real root in (0,1) of Note that Now show that ( i ) (since ) ( i i ) (since ) ( i ii ) Show that for Let and we have PAGE 122 122 Thus the extreme points are and the inflection point is Note that and Hence, is increasing as and is dec reasing as Then we have that which implies that for Hence, for We thus have and is strictly decreasing as And note that and Hence, there exists only one real root in of and if We thus have if Lemma 3 4 Note that Let Show that if where is the only real root in of Note that PAGE 123 123 Now show that ( i ) (since ) ( i i ) (since ) ( i ii ) Show that for Note t hat the minimum point of is at we thus have that is increasing as Hence, which implies for We thus have for and is increasing as And note that and [Since ] Hence, there exists only one real root in of and if We thus have if Proposition 3 4 Note that and Since hence, by Proposition 3 2, the merchant s optimal profit in Case J is either for or for PAGE 124 124 First show that By Lemma 3 2, is the only real root in of Note th at is a strictly increasing function for and Since hence, which implies By Proposition 3 1 & 3 2, the merchant s optimal profits with in Case N and Case J are (1) for (2) and for (3) and for Hence, if then by Lemma 3 3, for and by Lemma 3 4, for We thus have Case J: Join oBaz is more profitable for the merchant than Case N: N o oBaz for Proposition 4 1 Since hence, the number of purchase From Eq. ( 4 3) the retailer s profit function is PAGE 125 125 We then have the optimal price and the optimal profit by setting Proposition 4 2 Unde r the StickK business model, every customer puts up whatever amount that s /he wants At time 1 a consumer puts up the wager and at time 2 s/he takes back or loses it. According to mental accounting, th ey perceive to win or lose Hence, the consumer perceives refund value as By Eq. ( 4 9), a consumer at time 1 believes s/he will redeem the rebate if which implies By Eq. ( 4 4), a consumer at time 1 believes s/he will redeem the rebate at time 2 but will fail to do so if and We thus have which implies By Eq. ( 4 8), a consumer at time 1 purchases the service if which implies Note that if then this consumer who makes the purchase at time 1 will always redeem the rebat e at time 2. Then the retailer makes no profit. If then this consumer makes no purchase. PAGE 126 126 Hence, the condition leads the customer to make the purchase. Since every customer is able to put up whatever amount that s /he wants hence, a rational customer will make the purchase and choose to pay the retailer his/her most beneficial wager, i.e., Since every consumer will pay the retailer and put up hence the retailer s optimal profit is Proposition 4 3 If the present biased preference follows a uniform distribution then the optimization problem becomes Subject to Part (1) First to figure out the feasible range for ( i ) Note that implies then we have PAGE 127 127 Now show that and are real numbers. Let be the discriminant, then Note that Subcase (a) if then Since hence, we have that which subsequently implies that Hence Subcase (b) if then We thus have that Hence, we know that from Subcase (a) and Subcase (b), and and are real numbers. ( ii ) Note that implies or and implies But we note that if then In other words will result in which means no customers make the purchase, and thus the retailer makes no profit. PAGE 128 128 ( iii ) Note that since Also note that We thus have Hence, we subsequently have Hence, we know We further have the following cases: Case (A) No feasible solution for and PAGE 129 129 Case (B) No customers make the purchase for Case (C) All the customers make the purchase for Case (D) Some customers make the purchase for Part (2) Now show that either no consumers or all the consumers make the purchase by the optimal wager. The Lagrange function is The Kuhn Tucker conditions are PAGE 130 130 Subcase (a) if and then from the first KKT condition, we have Now show that and are real numbers. Let be the discriminant. Note that and and we then have that and which implies that Hence, and are real numbers. Now show that no consumers make the purchase by the wager PAGE 131 131 Since i.e., hence, no customers would make the purchase by Then show that all the consumers make the purchase by the wager Since hen ce, we have and We thus have and note that all the customers would make the purchase by and the optimal profit become means the number of redemption equals to the number of purchase. means no consumers make the purchase. PAGE 132 132 means no consumer s redeem their rebates. However, this case implies and n ote that also implies that no customers would make the purchase. Hence, we have the optimal profit is either or Proposition 4 4 If the presen t biased preference follows a uniform distribution then the optimization problem becomes subject to Note that the constrai nts and imply Then the optimization problem becomes subject to PAGE 133 133 The Lagrange function is The Kuhn Tucker conditions are C ase ( 1 ) if and then from the first KKT condition, we have which implies that We have from and from which imply Finally, we have the optimal profit Note that [ ] means all customers would make the purchase, i.e, the number of purchase is 1, and [ ] means all PAGE 134 134 the customers who make the purchase redeem their rebates, i.e., the number of redemption equals to the number of purchase. T he retailer makes no profit if all the buyers redeem. Hence, we keep considering Case (2) the number of purchase is 1 in the following section but ignore the case of the number of redemption equals to the number of purchase. Case (2) all customers make the purchase and the number of purchase is 1, which implies Then the optimization problem will turn out to be subject to T hen the Lagrange function becomes The Kuhn Tucker conditions are We have from the first KKT condition and thus Finally, we have the optimal profit PAGE 135 135 Proposition 4 5 From Proposition 4 1, we have that the retailer s optima l profit under the standard retailer model is From Proposition 4 2 and Proposition 4 3, we know the retailer s optimal profit under the StickK business model is We thus have Hence, if then Lemma 4 1 If the present biased preference follows a uniform distribution then we have by Proposition 4 1 and by Proposition 4 4. Let Case (1) with respect to PAGE 136 136 Let then Note that and and we thus have Then which implies is a decreasing function of Note that Let be a small positive number such that then we have Since is a decreasing function of and hence, we can find one and only root and such that Case (2) with respect to Since and hence, Note that PAGE 137 137 Let be a small positive number such that then we have Since is an increasing function of and hence, we can find one and only root and such that Proposition 4 6 Case (1) with respect to the membership fee By Lemma 1, is a strictly decreasing function of and is the only one root. Hence, for and for Gym Pact business model is better than the standard retailer model if the membership fee is less than Case (2) with respect to the effort cost By Lemma 1, is a strictly increasing function of and is the only one root. Hence, for and for PAGE 138 138 Note that and If or then is always negative. Hence, we have that for and PAGE 139 139 LIST OF REFERENCES Anand, K. S., R. Aron. 2003. Group Buying on the Web: A Comparison of Price Discovery Mechanisms. Manage Sci 49 (11) 1546 1562. Bernard, T. S. 2012 Gym Pact Fines You for Not Exercising, New Yo rk Times (January 2), http://bucks.blogs.nytimes.com/2012/01/02/gym pact fines you for not exercising/ Dholakia, U. M. 2011 What Makes Groupon Promotions Pro fitable For Businesses? http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1790414 Eaton, K. 2011 Don't Have Time To Haggle Your Daily Deals? oBaz Can Help. Fast Company (August 8 ), http://www.fastcompany.com/1772325/obaz launches to haggle your daily deals for you Edelman, B., S. Jaffe, S. D. Kominers. 2010. To Groupon or Not to Gr oupon: The Profitability of Deep Discounts. Harvard Business School, http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1727508 Gilpatric, S. M. 2009. Slippage in Rebate Programs and Present Biased Preferences. Marketing Sci 28 (2) 229 238. Hernandez, B. A. 2010. Do Groupons hurt businesses?, The Christian Science Monitor (October 1), http://www.csmonitor.com/From the news wires/2010/1001/Do Groupons hurt busi nesses Jing, X. J. Xie. 2011. Group Buying: A New Mechanism for Selling Through Social Interactions. Manage Sci 57 (8) 1354 1372. Johnston, S. 2011. Harvard grads turn gym business model on its head; fitness plan The Boston Globe (January 24), http://articles.boston.com/2011 01 24/business/29342190_1_membership fees planet fitness workouts Jucker, J V., M. J. Rosenblatt 1985. Single period inventory models with demand uncertainty and quantity discounts: Behavioral implications and a new solution procedure Nav al Research Logistics Quarterly 32 (4) 537 550. Koh, Y. 2011 After Mochi, Osechi: The Fir st Food Brouhaha of the New Year, The Wall Street Journal (January 5), http://blogs.wsj.com/japanrealtime/2011/01/05/after mochi osechi the first food brouhaha of the new year/ Koh, Y. 2011. Groupon Learns Japan Ropes, The Wall Street Journal (January 18), http://b logs.wsj.com/japanrealtime/2011/01/18/from new year meals to drag shows groupon learns japan ropes/ PAGE 140 140 Levine, R. B. 2010 Do Groupons Hurt Businesses? The Christian Science Monitor ( October 1 ), http:// www.csmonitor.com/From the news wires/2010/1001/Do Groupons hurt businesses Little, J. D. C. 1975 BRANDAID: A Marketing Mix Model, Part 1: Structure. Operations Research 23 (4) 628 655. Lodish, L. M., M. M. Abraham, J. Livelsberger, B. Lubetkin, B. Rich ardson, M. E. Stevens 1995. A Summary of Fifty Five In Market Experimental Estimates of the Long Term Effect of TV Advertising. Marketing Sci 14 (3) G133 G140. MacMillan, D. 2011. Groupon May Gain as 19% of Non Users Plan to Try Coupon Service. Bloomberg Businessweek (June 9), http://www.businessweek.com/technology/content/jun2011/tc2011069_738419.ht m MacMillan, D. 2011. LivingSocial Aims to Be Different from Grou pon. Bloomberg Businessweek (September 22), http://www.businessweek.com/magazine/livingsocial aims to be different from groupon 09222011.html Metallic Druzy Ring (2012, January 6). Retrieved from https://www.obaz.com/game/product/1699/0/share/pinterest/by/67659 Metallic Druzy Ring (2012, January 6). Retrieved from https://www.obaz.com/game/product/1699/1665 Mokey, N. 2011. Groupon gripes: Are daily deals headed for disaster? Digital Trends (March 10), http://www.digitaltrends.com/opinion/groupon gripes are daily deals headed for disaster/ Neslin, S. A., C. Henderson, J. Quelch. 1985. Consumer Promotions and the Acceleration of Product Pu rchases. Marketing Sci 4 (2) 147 165. O'Donoghue, T., M. Rabin. 1999 Doing It Now or Later. The American Economic Review 89 (1) 103 124. Overnight Stay For Two With Scottish Gourmet Breakfast for Â£69 at Banchory Lodge Hotel (Up to Â£205 Value) (2011, Octob er 6). Retrieved from http://www.groupon.co.uk/deals/aberdeen/banchory lodge/811643 Q Saxon Judd Tulsa (2011, October 5). Retrieved from http://www.groupon.com/deals/q custom clothier tulsa ok Salkever, A. 2010 Is Groupon Changing What People Are Will ing to Pay? DailyFiance (August 10), http://www.dailyfinance.com/2010/08/16/is groupon changing what people are willing to pay/ PAGE 141 141 Sittenfeld, L. R. 20 08. StickK.com: Take a Contract Out On Yourself. CNBC (March 11), http://www.cnbc.com/id/23575559/StickK_com_Take_a_Contract_Out_On_Your self Steiner, C. 2010. Meet the fastest growing company ever, Forbes (August 30), http://www.forbes.com/forbes/2010/0830/entrepreneurs groupon facebook twitter next web p henom.html Stephen, A., O. Toubia. 2010. Deriving Value from Social Commerce Networks. Journal of Marketing Res earch 47 (2) 215 228. Talluri, K. T., G. J. v. Ryzin. 2004. The theory and practice of revenue management Springer. Tellis, G. J. 1988. The Pri ce Elasticity of Selective Demand: A Meta Analysis of Econometric Models of Sales. Journal of Marketing Research 25 (4) 331 341. Thaler, R. 1985. Mental Accounting and Consumer Choice. Marketing Sci 4 (3) 199 214. Var ian, H. R. 1980. A Model of Sales. Ameri can Economic Review 70 (4) 651 659. PAGE 142 142 BIOGRAPHICAL SKETCH Yu Chen Yang received his doctoral degree in business administration at the University of Florida in May, 2013. He also received an M.B.A. in management information system and a B.S. in applied math ematics. Prior to graduate study, he served two years in the Republic of China Air Force and was a research assistant at the Institute of Technology Management of National Tsing Hua University in Taiwan. His research interests are in the areas of economics of information systems electronic commerce, and social media H is teaching interests include da tabase management, statistics and data analysis. |