LITERATURE REVIEW (12-16 pages) This essay is basically synthesizing 10 sources to support the new research im proposing that needs to be done. The sources needs to be categorized into 3/4 subheadings

ABSTRACT The effective media strategy in advertising is gradually becoming the premise of company that lives in the competition of marketing. Due to the rapid growth of new advertising media and technologies, it is possible for a firm to precisely target advertising to the potential consumer segment within a market. This research explores the extent to which an advertiser should regulate the quality of its targeting and effect on consum- ers’ surplus and social welfare. The authors present a theoretic model measuring the targeting quality of internet-based targeted advertising with two measures termed accuracy and recognition. Accuracy measures the possibility of correct prediction in the target segment, while recognition is defined as the probability that any member of the targeted segment is identified. The authors demonstrate, within a monopolistic framework, the online targeted advertising might lower or increase both consumers’ surplus and social welfare compared with mass advertising, which depends on different range of accuracy and recognition in the social media.

The recognition of internet-based targeted advertising plays a positive role on equilibrium price, whereas the accuracy plays a negative role in the regulation of advertising intensity and equilibrium price. Therefore, it is believed that the accuracy and recognition of the online targeted advertising can be used as a lever for the strategic segmentation of a market. Bidirectional Role of Accuracy and Recognition in Internet-Based Targeted Advertising Jiang Zhao, School of Economics and Management, Southeast University , Nanjing, China Shu-e Mei, School of Economics and Management, Southeast University , Nanjing, China Wei-jun Zhong, School of Economics and Management, Southeast University, China Keywords: Accuracy, Internet-Based Targeted Advertising, Mass Advertising, Price Competition, Recognition 1. INTRODUCTION Advertising is one of the most critical marketing strategies for a firm, and the largest element of advertising spending is media purchasing (Aaker, 1968). The informative view of advertis - ing claims that in many markets, firms invest in advertising to create awareness for products, prices, attributes and special offers to otherwise uninformed potential consumers. Therefore, it is always a great challenge to buy the media effectively and not direct toward the “wrong people”. As a classic concern goes by John Wanamaker in 19 th century, “Half the money I DOI: 10.4018/jeco.2015010104 Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 51 spend on advertising is wasted and the trouble is I don’t know which half” (Dellarocas, 2012, p.1178). Traditionally, the objective in media planning is to minimize wasted advertising by reducing the quantity of advertising sent to the consumers who will probably never buy the product. However, throughout a variety of technological advances, the internet provides many firms with an unparalleled ability to precisely target advertising towards the po- tential consumer segment within a market. In particular, the targeting improves the quality of the match between the consumer and the advertisement message (Bergemann& Bonatti, 2011). Currently, within the internet-based tar- geting techniques including IP address tracking and the analysis of cookies, the consumers’ preference and potential demands can be eas- ily obtained across many different websites and search engines (Rutt, 2012). That means the firm can send the internet-based targeted advertising via online platforms to the potential consumers with great accuracy. For instance, social network websites, such as Google and Facebook, can roughly reveal an individual’s current interests by tracing his/her recent web- browsing activities, and then directly provide the advertisements of relevant products that might interest him/her. As reported in a previous study, a potential consumer cannot be an actual buyer unless the firm invests in advertising for the sales (Stahl, 1994). The effective media strategy in advertis - ing is gradually becoming the critical premise of the company that lives in the competition of marketing. However, due to the expensive spending on mass advertising through national TV or newspapers, firms are increasingly active in the use of internet-based targeted advertising while selling products to the loyal consumers (Iyer et al., 2005; Gal-Or et al., 2006).Given that advertisers are motivated to pursue the internet-based targeted advertising, the objec - tive of our study is to understand the following questions: when the firms have the ability to choose different targeting levels to different consumer segments, how they allocate their advertising? Whether the higher of targeting ability is, the better social benefits can obtained when compared to mass advertising? What is the primary role of targeting quality of internet- based targeted advertising on the products?

For these purposes, we first present a model to analyze the role of recognition and accuracy of internet-based targeted advertising on firm’s equilibrium price and profit in monopolistic framework, and then develop the model on the optimization of the investment on the two measures of imperfect targeting. A first output of this paper is to discover the bidirectional role of the two measures of internet-based targeted advertising, and the others, the effect of targeted advertising is not always better than mass advertising, which depends on the optimal choice of targeting qualities.

In fact, even in the monopolistic market, the informative advertising is used for the introduction of products’ characteristics so as to enlarge the potential market and avoid the potential rival’s entry in the market. For instance, Microsoft’s operating system is nearly occupied the total personal computer market in China, still quite a large number of advertisements are sent to the consumers in China every year. The model addressed in this article fits well with previous studies on the impact of informative advertising on product information and pricing.

In particular, Butters (1977) first proposes a message-sending model in markets for ho - mogenous goods, where advertising provides information about the existence of products and their characteristics. Later, Grossman and Shapiro (1984) extend the Butter’s model to the markets with horizontal differentiation via a circle model, and show that the decreased ad- vertising costs may reduce profits by increasing the severity of price competition. The similar models are established by Stahl (1994), Bester and Petrakis (1995), Hamilton (2004) and Celik (2007). One of the assumptions in these papers is that the advertising is uniform throughout the market. Hernández-García (1997) compares the targeted advertising and mass advertising within a monopolistic framework, and considers that the targeting might lower both consumers’ surplus and social welfare. Then Roy (2000) Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

52 Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 establishes a model, within a homogeneous market, firms can target consumers, and com- pete on prices after observing the competitor’s advertising. Moreover, an infinite number of equilibrium is generated, which depends on significant coordination between firms. Later, Esteban et al. (2001) consider different levels of advertising to be directed at different segments within a market, and argue that the monopolist will direct heavier advertising weights to the consumers who are willing to pay more for the product, in such a way that the overall level of advertising falls with targeting. By contrast, this paper proposes similar results with Iyer et al.

(2005), who demonstrate that the use of internet- based targeted advertising increases the market price and leads to higher profits when compared with mass advertising. Similarly, Johnson (2013) examines how the increasing ability of firms target their advertisements to particular individuals influences market outcomes when consumers have access to advertising avoidance tools, and confirms that the firms generally ben- efit from increased targeting. All these results indicate that the targeted advertising is better than mass advertising in cost savings due to its accurate targeting ability. However, there are still some controversies on the role of targeting ability in the firms’ equilibrium prices and profits while adopting targeted advertising for the sales. Galeotti and Moraga-Gonzalez (2003) present a strategic game of pricing and targeted advertising, and prove that the market is segmented only from time to time and presents substantial price dis- persion across segments. Gal-Or et al. (2006) introduce a bidimensional descriptor to evaluate the quality of the targeting termed accuracy and recognition, and demonstrate that the adver - tiser should choose to reduce its investment in recognition and increase its investment in ac- curacy while facing a competitor that pursues a strategy to improve its recognition of potential customers. Recently, Ben Elhadj-Ben Brahim et al. (2011) construct a model in which two horizontally differentiated firms compete in prices and targeted advertising on an initially uniformed market, and they prove that firm’s equilibrium profits may be lower with targeted advertising relative to mass advertising when the firm can perfect differentiate between cus- tomers choosing a possibly different advertising intensity for each consumer type. Likewise, Esteves and Resende (2013) also suppose that when the firms have the perfect targeting ability, the overall welfare and consumer welfare falls when firms use targeted advertising instead of mass advertising.

Obviously, the assumption of perfect targeting is an ideal condition. In most con- dition, the firm only has imperfect targeting quality due to the consumers’ imperfect in- formation. Therefore, in this article, we use the accuracy and recognition to describe the targeting quality of imperfect internet-based targeted advertising, which avoids some limits of aforementioned assumption on the perfect targeted advertising and reveals better on the features of internet-based targeted advertising.

Fortunately, previous studies on targeted pricing show us some clues. Chen et al. (2001) prove that when individual marketing is feasible, but imperfect, improvements in targetability by either or both competing firms can result in win-win competition for both even if both players behave noncooperatively. Then Chen and Iyer (2002) demonstrate that low address- ability and high levels of market differentiation both help reduce price competition, which facilitates symmetric choice of addressability by the firms in equilibrium. Likewise, Esteves (2010) investigates the competitive and welfare effects of information accuracy improvements in markets where firms can price discriminate after observing a private and noisy signal about a consumer’s brand preference, and argues that the firms charge more to customers they believe have a brand preference for them, and the price has an inverted-U shaped relationship with the signal’s accuracy. In similar results, Shaffer and Zhang (2002) explore the competitive effects of one-to-one promotions in a model with two com - petition firms where the firms differ in size and consumers have heterogeneous brand loyalty.

They suggest that the optimal way to manage customer churn is to engage in both offensive Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 53 and defensive promotions with the relative mix depending on marginal cost of targeting. Then, Chen Y et al. (2009) consider that dynamic targeted pricing can benefit competing firms, when they actively pursue customer recognition based on consumer purchase history. However, the dynamic targeted pricing may not benefit for a monopolist. In a recent study, Zhang and Katona (2012) analyze the strategic aspects of contextual advertising, considering an interme- diary who has access to a content base, sells advertising space to advertisers who compete in the product market. Their results indicate that the intermediary lowers the targeting ac- curacy such that the consumers see less relevant advertisements when competition is high. That gives us a hint that the accuracy of advertising might be used as a lever for price competition. In this study, we are focused on the dif- ference between the internet-based targeted advertising and mass advertising in a monopo- listic framework. The remainder of the article is organized as follows. Section 2 introduces a basic model of mass advertising and imperfect internet-based targeted advertising. Section 3-6 explores the bidirectional role of accuracy and recognition in the equilibrium price, regulation of advertising intensity, consumers’ welfare and social welfare. Section 7 investigates the optimization of imperfect internet-based targeted advertising with mismatch lose. Sec- tion 8 extends the analysis of the optimized investment on the accuracy and recognition of internet-based targeted advertising. Section 9 concludes. The Appendix collects some formal proofs and the definitions of all the symbols.

2. LITERATURE AND MODEL SETUP 2.1. A Basic Model of Mass Advertising and Imperfect Internet- Based Targeted Advertising The informative advertising is used to stimulate the consumer’s demands. Even in the monopo- listic framework, the advertising is used to build the brand image of the product for the demands of consumers, which could be made as a bar - rier for the entry of other potential firms.

Therefore, in order to simplify the study, we develop a mass advertising model of a market with only one monopolist as previous reported (Hernández-García, 1997). Suppose the firm chooses a price p for the product, and the potential demands for the product can be defined as q x p = ( ), but this will never be realized until the consumers are informed of its existence by advertising. In this framework, we are aimed to study whether the shift from mass advertis - ing to internet-based targeted advertising can affect the monopolistic firm’s supply of the products. We assume that the mass advertising is sent to the customers randomly and all con- sumers have the same probability of φ receiv - ing an advertisement. Moreover, A ( ) is de- fined as the per capita cost needed to reach the fraction φ and the marginal cost of advertising is increasing with the number of consumers. In addition, the marginal cost of advertising is increasing with the number of consumers reached, i.e. A ( ) 0 and A ( ) 0. These assumptions are in accordance with previous reports (Butters, 1977; Grossman and Shapiro, 1984). In this situation, both the price and the level of advertising are obtained in the firm’s strategy space. Then, the firm produces the product at a constant marginal cost of c. The firm’s maximum profit with mass advertising r ( ) is defined as follows: Ma x p c xp A r = ( ) ( ) ( ) (1) The first order condition (FOC) of the profits on price p and the fraction are shown in the following equations: = + = r p p c x p x p 0 0 ; ( ) () ( ) (2) Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

54 Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 = = r p c x p A 0 0 ; ( ) () ( ) (3) In order to evaluate the impact of advertis - ing on the reaching consumers more accu- rately, it requires that we look at a parametric example of the cost function A ( ) φ . The sim- plest example is a quadratic functional form, which equals to multiplication of constant a and square of fraction φ as previous reported (Bester & Petrakis, 1996): A a ( ) = 2 2 (4) We denote the optimal price termed pr * and optimal reaching fraction as * for mass advertising. The firm’s pricing strategy is de- scribed in Equation (2), which shows a simple inverse elasticity rule. Obviously, the optimal price of the product does not depend on the level of advertising. Therefore, a dichotomy between the pricing strategy and the advertising strategy is available. Under imperfect informa - tion, reducing the price increases the full infor - mation demands, however, the monopolist needs to cover the new segment of the market with additional advertisement so as to realize that sale. We assume that all consumers are hetero - geneous and they are located in a linear city in the presence of positive transportation costs.

The length of the city which means the total demands of the consumers is normalized to 1.

All of them derive the same gross surplus v from consumption and incur a transportation cost tper unit of distance. The utility function can be defined as following equation of U v p tx = , where x [ , ] 0 1 represents the physical distance between consumers lo- cated at the point and the monopolist. In fact, due to the normalized unit total demands, the value of physical distance x equals to the potential demands of x x p = ( ). Therefore, consumers’ potential demands, firm’s equilib - rium price and advertising intensity with mass advertising can be expressed in the following expressions. Moreover, the procedure of deriva - tion is expressed in Section 1 of the Appendix: x p v p t p v c v c at r ( ) , ( ) * * = =+ = 2 4 2 and (5) Here, the expression of s v c = repre- sents the social surplus gross of transportation costs (Hernández-García, 1997, p.134). That means: *=s at 2 4 (6) Then, we consider the description of the internet-based imperfect targeted advertising.

We use a spatial analogy for the description of heterogeneity for the existing consumers lo- cated in a linear city. Each consumer will purchase only one unit of the products from the firm at a distance of x after receiving an ad- vertisement from the firm. Therefore, x is considered to be the marginal consumer. How- ever, due to the imperfect targeting ability, there is a probability of α committing a Type I error while sending advertising to the wasted popu- lation. That means when the firm aimed to provide internet-based targeted advertising to the objective consumers, only a segment of x ( )1 can be correctly recognized, and only this segmentation of consumers will purchase the product after receiving the advertising.

Meanwhile, there is a probability of β com- mitting Type II error targeting the consumers for the firm purchasing the online advertising via social media organization such as Facebook, Google etc. It means the population of ( )1 x consumers that are misrecognized as the objec - tive consumer. However, all these consumers will not purchase any of the products after re- ceiving the advertising. Hereinafter, recognition Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 55 ( )R and accuracy ( )T are used for the better description of the quality of online targeting technology. The relationship between recogni - tion ( ) R and accuracy ( )T for a given adver - tiser is shown as follows: R = 1 (7) Tx x x = + ( ) ( ) ( ) 1 1 1 (8) In our assumption, all of the consumers derive the same gross surplus from consumption and incur a transportation cost t per unit of distance. Therefore, their utility function can be expressed as U v P tx = 0, which means only consumers with U 0 will pur - chase the product, so the monopolist’s potential sale at price p are given by the following expression for the internet-based targeted ad- vertising: x p v p t ( ) (9) Let ϕ considered as the fraction of con- sumers while receiving the advertisements from the firm. Then the firm’s profit using the online advertisement at the range of [ , ] 0 x with Type I and Type II error are presented in the follow- ing equation: Max p c x a x x t = + ( ) ( ) ( ) ( )   1 2 1 1 2 (10) Substituting the constraint condition of Equation (7) and (8) into the objective Func- tion (10), and the maximum profit of the firm while purchasing online targeted advertising to the potential consumers can be simplified as follows: Ma x p c xR a x R T t = ( ) 2 2 (11) The FOC of the profits on price p and the fraction yields (see Section 3 of the Ap- pendix): = + + [ ] + = t p R t p v c pRa tT 2 20 2 ( ) (12) That means: p v c a Tt*= + + 2 4 (13) = = t v p t R p ca T ( ) 0 (14) That implies: * ( ) = p c T a (15) Moreover, the second order condition of the profits on price p and the fraction implies the existence of the max value.

Therefore, we demonstrate that fraction of advertising is correlated with the accuracy of social media. Here, we find that target - ing really raises the price of the product.

When the price is decreased by the mo - nopolist, the dichotomy between pricing and advertising is broken by the accuracy of the targeting. As a result, the marginal cost of reducing the price can be increased by targeting accuracy. Then, our objective is to evaluate the mar - ket outcomes from a social perspective. In our Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

56 Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 model, advertising is socially desirable, when the accuracy is increased, the market size might be changed. So the relevant issue is to analyze whether that internet-based targeted advertising for the firm in monopolistic framework is always socially optimal. In a previous study, Shapiro (1980) demonstrates that a monopolist using a random technology is always undersupplied by informative advertising. Next, we extend the result to the case of targeting with the welfare standard of consumers’ surplus plus producer surplus. From the spatial analogy as mentioned above, the total social surplus under targeting can be shown as follows:

W st xxR a xR T t= ( ) 2 2 2    (16) The FOC for the socially optimal level of advertising is derived as follows: = = W s tx xR a xR T t 0 2 0 ; ( ) (17) Although the online targeting technology raises the price of the product, the monopolist has a higher incentive to provide internet-based targeted advertising than mass advertising.

However, the targeting also plays distinct roles affecting welfare in opposite directions. First, the full information demand is reduced by the higher price, which decreases welfare. Secondly, the wasting of advertisement can be avoided by the online targeting, which increases welfare.

In the following section, we are aimed to de- termine which effect is dominant and whether the bidirectional role is correlated with the two parameters of online targeting. 2.2. Role of Accuracy in the Regulation of Firm’s Equilibrium Price In order to evaluate the impact of targeting on social welfare, the equilibrium price of prod- ucts using mass advertising and online targeted advertising are compared from aforementioned equation (5) and (13), which indicated that the imperfect targeting might raise the equilibrium price of the product due to distinct accuracy.

Indeed: P P v c a T v c a Tt r* * = + + + = 2 4 2 4 (18) So the monopolist has the ability to raise the equilibrium price above marginal cost by chang - ing the intensity of advertising and the accuracy.

From the above results, we can conclude that the internet-based targeted advertising might benefits the monopolist, so the relevant issue is to analyze the implications of the different roles of targeting intensity for both consumers and society. In the following study, the role of accuracy and recognition in the regulation of advertising intensity, the demands of consum- ers, consumers’ surplus and the social welfare are described separately.

Proposition 1: The accuracy of internet-based targeted advertising plays a negative role in the regulation of equilibrium price while compared with mass advertising. How- ever, the accuracy plays a bidirectional role in the regulation of advertising intensity when compared with mass advertising.

Indeed, if T v p t ( ) / , the number of ads rise, whereas if T v p t > ( ) /, and then the num- ber of ads decreases.

Proof: By contradiction. Let us suppose that T v p t > ( ) and * *< . This implies t h a t ( ) ( )( ) p c T a p c v p at > . I t means * *> , which contradicts with the aforementioned assumption.

The level of advertising might be reduced because, although the targeting makes each advertising much more efficient depending on correct targeting. The wrong targeting not only induced a great cost loss, but also increase Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 57 consumers’ burden. Finally, the effect of online targeted advertising might be lower than mass advertising.

2.3. Bidirectional Role of Accuracy and Recognition of Internet- Based Targeted Advertising in the Demands of Consumers Next, we begin to analyze the different ways in which targeting affects consumers. First, when the accuracy and recognition is too large, that means the price of the product might rise to a high level, which reduces the number of consumers who buy the product. As the objec- tive of the targeting advertising is to obtain the potential consumers and reduce the waste of advertisements by accuracy and recognition, the two parameters of targeting on consumers demand needs to be studied. Then, the number of consumers who buy the products after receiv - ing the mass advertising and internet-based targeted advertising is compared. The original demands of consumers are described as x s atr = 3 2 8 / while the monopolist adopt- ing mass advertising for the sales. In addition, the demands of consumers while receiving the online targeted advertising can be expressed as follows: x R s t R p c T tat = 2 4 2 ( ) (19) Therefore, we can conclude Proposition 2 after the comparison of the demands of consum - ers with different advertising strategies.

Proposition 2: The accuracy and recognition of internet-based targeted advertising plays a bidirectional role in the demands of consumers. The expression of: RT a ts as t a pc > ( ) / () 4 2 2 3 2 means the demands of consumers after receiv - ing internet-based targeted advertising might be lower than that of mass advertising. However, when: RT a ts a s t a pc < ( ) / () 4 2 2 3 2 the demands of consumers after receiving the online targeted advertising might be even higher. From the aforementioned analysis, it is clear that the monopolist benefits from the new advertising only in specific range, whereas the consumers are likely to be worse off. Then, we are determined to verify whether the targeting show similar results to consumers’ surplus. 2.4. Bidirectional Role of Accuracy and Recognition of Internet- Based Targeted Advertising in the Consumers’ Surplus Hereinafter, we compare the consumers’ surplus after the consumers receiving mass advertising from the monopolist and that incepting of in- ternet-based targeted advertising. It is known that the consumers’ surplus CS( ) can be depicted as the following expression of CS P P Q = ( ) / max * 2 . Therefore, the con- sumers’ surplus of online targeted advertising CS t ( ) and mass advertising CS r ( ) are com- pared. We can obtain the results as the follow- ing Proposition 3.

Proposition 3: The accuracy and recognition of internet-based targeted advertising is cor - related with the consumers’ surplus, when: RT s ats atp ac s t p c s p c > + ( ) ( ) ( ) 8 4 4 2 2 the consumers’ surplus after online targeted advertising is higher than mass advertising.

However, the opposite result of: RT s a ts atp a c s t p c s p c + ( ) ( ) ( ) 8 4 4 2 2 Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

58 Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 will lead to a lower consumers’ surplus when compared with mass advertising.

Proof: See following equations:

CSvv c s at s at r= + = 1 2 2 8 32 3 2 3 2 ( ) CS vv c p cs t R p c T at t= + 1 2 2 4 24 ( )( ) Then CS t and CS r are compared in the following in-equation. If CS CS t r≥ , then: s at v v c p c s t R p c T at3 2 32 1 2 2 4 2 4 0 + × ( ) ( ) else if:

CS CS t r< s at v v c p c s t R p c T at3 2 32 1 2 2 4 2 4 0 + > ( ) ( ) That’s the results. QED.

From the analysis, we can conclude that the consumers’ surplus largely depends on accuracy and recognition of internet-based targeted advertising, which plays a bidirectional role in the regulation of consumers’ surplus.

The too high accuracy and recognition might reduce the consumers’ surplus compared with mass advertising.

2.5. Bidirectional Role of Accuracy and Recognition of Internet-Based Targeted Advertising in Social Welfare In the following study, we examine whether the targeted advertising with different targeting quality is desirable from a social point of view.

The results can be obtained from Proposition 4.

Proposition 4: The targeting might not be socially desirable which depends on the product of accuracy and recognition of internet-based targeted advertising. When: RT s v p pc s v p > + 4 2 21 2 / ( )()( / ) the social welfare with imperfect internet-based targeted adverting is higher than mass advertis - ing, and the opposite results of: RT s v p pc s v p < + 4 2 2 1 2 / ( )()( / ) will lead to a lower social welfare when com- pared with mass advertising.

Proof: The comparison between the levels of welfare under both mass advertising and internet-based targeted advertising is shown as follows: W s at r= 4 2 16 and: W s v p v pp cTR at t= + ( ) ( )( ) 2 2 1 2 When: RT s v p p cs v pW W t r > + > 4 2 2 1 2 ( ) ( )( ), However, when: RT s v p p cs v pW W t r < + < 4 2 2 1 2 ( ) ( )( ), Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 59 That also means too high accuracy and recognition of the online targeting might reduce the social welfare .QED.

From the above results, we can realize that a more efficient targeting technology might not be always socially desirable, which is correlated with the accuracy and recognition. The intuition is as follows: Targeting raises the price of the good when the accuracy and recognition are upgraded. If the accuracy and recognition are too high, the cost of targeting might increase a lot, therefore, the firm must raise the price at a high level so as to make profits. Therefore, the targeting is really harmful for consumers.

However, the process of advertising is becoming efficient by targeting when the accuracy and recognition is not too high. In this situation, the targeting generates substantial savings in cost and shows better marketing effect than mass advertising. Generally, we can conclude that if the accuracy and recognition are not too high, the monopolist can raise the price of the product by only a little and both consumers’ surplus and social welfare rise. By contrast, if the ac- curacy and recognition of the monopolist is too high, this positive effect might be weaken, and might lower the social welfare. Therefore, the monopolist should make its targeting accuracy and recognition in a specific range to pursue the optimized profit. The proper accuracy and recognition of targeting by the firm can be used as a tool for segmentation of a market. 2.6. Optimization of Imperfect Targeting with Mismatch Lose In the above model, we haven’t considered the loss of the firm while purchasing the social media and sending the targeted advertising to the wrong consumers. In fact, as there is a probability of committing Type II error in sending internet-based targeted advertising to the wrong consumers by the firm, all these consumers will never buy the product even receiving the advertising. However, these wrong consumers are considered as the firm’s poten- tial consumers, and these advertisements are wasted, which we call it mismatch lose for the firm. In our following model, we further con- sider the coefficient of the mismatch lose as c1 for one unit wrong consumer, therefore, the mismatch lose of the firm while purchasing the online advertising and sending to the consum- ers of ( ) 1 x at a probability of is expressed as c x1 1 ( )  , and then the profit of the firm is shown in the following expression: Max p c x a x x c x t = + ( ) ( ) ( ) ( ) ( )    1 2 1 1 1 2 1 (20) Substituting the constraint condition of (7) and (8) into the objective function of (20), and the max profit of the firm using internet-based targeted advertising can be expressed in the following equations: Ma x p c xR a x R T c T x R T t = ( ) ( ) 2 1 2 1 (21) The FOC of the profits on price p and the fraction yields: = + + + + = t p R p v c t a R tT c T R tT 0 2 2 1 0 2 1 [ ( )] ( ) (22) That means: p v c a T c T T v c T c a T* ( ) ( ) = + + + + + 2 4 1 2 2 1 1 21 1 (23) Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

60 Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 The equation works when: aT c T T c T a 4 1 2 2 1 1 1 = = ( ) ( ) = = t p c xRaxR T 0 0 ( )   (24) That means: * ( ) = p c T a Likewise, the second-order condition of the profit is also available. Therefore, we continue to compare the targeting level, the demands of consumers, as well as the welfare. From the equation (24), we can c o n c l u d e t h a t t h e e q u i l i b r i u m l e v e l o f advertising will not be changed and isn’t correlated with mismatch lose.

Proposition 5: The targeting level is cor - related with the accuracy of internet- based targeted advertising, but isn’t correlated with its mismatch lose .The optimized price arises with mismatch lose, and decreases with the improve - ment of accuracy. The minimum opti - mized price is: v c Tc a T + + 2 1 1 2 1( ) and the minimum optimized advertising level is considered as the following ex - pression of: 2 11c T a ( ) 2.7. Optimization of Imperfect Targeting after Investment on Accuracy and Recognition with Mismatch Lose In our model, when the firm chooses the online targeted advertising from the social media or - ganization for the marketing, the potential consumers might be changed into the actual consumers. Moreover, the accuracy and recog- nition of the internet-based targeted advertising might directly affect the firm’s profit and social welfare. If the accuracy and recognition of the internet-based targeted advertising are not well pursued, the effects of the internet-based tar - geted advertising might be less than mass ad- vertising. Therefore, the firm must make the decision simultaneously on the investment for a proper accuracy and recognition of the inter - net-based targeted advertising. Specifically, we assume that the cost of acquiring the accuracy and recognition is a quadratic function. Both k1 and k2 are constants in the function: f R T k R x k T x x ( , ) ( ) ( ) = + + 12 2 2 1 1    (25) The expression can be simplified in the following expression: f R T k R xk TRx ( , )= + 1 2 2 (26) Then, the monopolist’s profit with invest- ment on targeting qualities generates: Max p c xR a x R T c T x R T k R x k TRx t = ( ) ( )      2 1 2 1 12 2 (27) Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 61 The FOC yields: = + + + + + + = t p R p t a R T c T R T k R k TR 0 2 2 1 0 2 1 12 2 [ ( v c)] ( ) That means: p v c a T c T T k R k T v c T a c T kRTk T * ( ) ( ) = + + + + + + + + + 2 4 1 2 2 2 1 11 1 2 1 12 2 2 2 (28) = + + + + v c ac T k R ac T k T 2 2 2 2 1 2 1 1 2 The equation works only: a T c T T k R T k T T c T k RT k T a 4 1 2 2 2 1 1 122 1 122 = + + = + + ( ) ( ) (29) = = t p c xRaxR T 0 0 ( )   (30) That means: * ( ) = p c T a Likewise, the second-order condition of the profit is also existed. Moreover, we continue to compare the targeting level, the demands of consumers, as well as the welfare with the mass advertising. The same conclusions are also de- rived. From the equation (29), we can conclude that the level of advertising will not be changed and isn’t correlated with the targeting cost on the accuracy and recognition. From equation (28), we can conclude that the optimized price showed positive correlation with recognition of internet-based targeted advertising, but negative correlation with the accuracy of internet-based targeted advertising. That is why when the firm improves the targeting, both the recognition and accuracy will improved, which showed opposite effect to the firm. The higher recognition will enhance the equilibrium price, as well as the demands of consumers. However, the higher accuracy will decrease the equilibrium price, as well as the demands of consumers. The final effects derive the relative role of recognition and accuracy on the firm’s equilibrium price.

Proposition 6: The targeting level is correlated with the accuracy of internet-based targeted advertising, but isn’t correlated with its mismatch lose .The optimized price arises with the recognized cost, and decreases with the improvement of accuracy. The minimum optimized price of the firm is considered as the following expression of: v c Ta c T kRTk T + + + + 2 1 1 2 1 122( ) and the minimum optimized advertising level can be expressed as: 2 1 1 122c T k RT k T a ( ) + + The recognition of internet-based targeted advertising plays a positive role in the regulation of advertising level, however, the accuracy of internet-based targeted advertising plays a nega - tive role in regulating advertising level. The final effects derive the dominate role of recognition and accuracy on the advertising level. Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

62 Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 From the above model, it is clear that the ac- curacy and recognition of the online advertising play distinct roles in the regulation of optimized price. The recognition of internet-based targeted advertising plays a positive role in advertising level. That means a higher recognition on tar - geted advertising leads to a higher advertising level. However, the accuracy plays the opposite role in advertising level, and a higher accuracy might lead to a lower advertising level. The results reveal that when the firm decides to invest in the quality of advertising, the firm should invest the recognition and accuracy of advertising with different directions. 3. CONCLUSION In this article, we consider a model of monopolist purchasing the internet-based targeted adver - tising with different accuracy and recognition, which adds new insight into the impact of media proliferation sending to the potential consumers.

Though the internet-based targeted advertising is proved to be beneficial for the efficiency of the firms and can reduce message wastage well, still much more accurate targeting does not mean better welfare for the society or company. We demonstrate that within a monopolistic frame - work, the targeted advertising might lower both consumers’ surplus and social welfare compared with mass advertising, which is correlated with accuracy and recognition. Therefore, different quality of targeting could greatly affect the firms’ equilibrium profits, sometimes, its effect might be worse than the mass advertising. Further, our results indicate that the accuracy and recogni - tion play the opposite role in the regulation of advertising level, which means the recognition plays a positive role in the regulation of equi- librium prices whereas the accuracy plays a negative role, and the final effects derive from one of the relative role. Therefore, we believe the accuracy and recognition of internet-based targeted advertising can be used as a lever for strategic segmentation of a market. Otherwise, it is clear that there are still some limits in the study of a monopolist model in the scope of our results, and therefore it needs some extensions for the analysis of targeting in a competitive framework in the further studies.

However, the bidirectional role of accuracy and recognition of the online advertising suggest us that the pursuit of perfect targeting from social media organization might not be desirable, this partially explain the current controversies on whether the targeted advertising is always better than mass advertising. This will benefit from more clarifications of extensions for the affecting factors on targeting in a competitive framework in the future studies. ACKNOWLEDGMENT Financial support provided by National Key Technology Support Program (No.

2012BAH29F01), National Natural Science Foundation of China (No.71371050) and Re- search and Innovation Project for College Grad - uates of Jiangsu Province (No. CXZZ13_0137) are gratefully acknowledged.

REFERENCES Aaker, D. A. (1968). Subject: Media Selection Title:

A Probabilistic Approach to Industrial Media Selec - tion. Journal of Advertising Research, 8, 46–54.

Ben Elhadj-Ben Brahim, N., Lahmandi-Ayed, R., & Laussel, D. (2011). Is targeted advertising always beneficial? International Journal of Industrial Organization, 29(6), 678–689. doi:10.1016/j.ijin - dorg.2011.03.002 Bergemann, D., & Bonatti, A. (2011). Targeting in advertising markets: Implications for offline versus online media. The Rand Journal of Economics , 42 (3), 417–443. doi:10.1111/j.1756-2171.2011.00143.x Bester, H., & Petrakis, E. (1995). Price competition and advertising in oligopoly. European Economic Review , 39 (6), 1075–1088. doi:10.1016/0014- 2921(94)00099-L Butters, G. R. (1977). Equilibrium distributions of sales and advertising prices. The Review of Economic Studies, 44(3), 465–491. doi:10.2307/2296902 Çelik, L. (2008). Strategic informative advertising in a horizontally differentiated duopoly . CERGE-EI. Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 63 Chen, Y., & Iyer, G. (2002). Consumer addressability and customized pricing. Marketing Science, 21(2), 197–208. doi:10.1287/mksc.21.2.197.153 Chen, Y., Narasimhan, C., & Zhang, Z. J. (2001).

Individual marketing with imperfect targetability.

Marketing Science , 20 (1), 23–41. doi:10.1287/ mksc.20.1.23.10201 Chen, Y., & Zhang, Z. J. (2009). Dynamic targeted pricing with strategic consumers. International Journal of Industrial Organization, 27(1), 43–50.

doi:10.1016/j.ijindorg.2008.03.002 Dellarocas, C. (2012). Double marginalization in performance-based advertising: Implications and solutions. Management Science , 58(6), 1178–1195.

doi:10.1287/mnsc.1110.1474 Esteban, L., Gil, A., & Hernandez, J. M. (2001).

Informative advertising and optimal targeting in a monopoly. The Journal of Industrial Economics, 49(2), 161–180. doi:10.1111/1467-6451.00144 Esteves, R. B. (2010). Price discrimination with private and imperfect information. Documentos de Tradalho. Working paper series. URL:http:www.eeg.

uminho.pt/economia/nipe Esteves, R. B., & Resende, J. (2013). Competitive targeted advertising with price discrimination.

Documentos de Tradalho. Working paper series.

URL:http:www.eeg.uminho.pt/economia/nipe Gal-Or, E., Gal-Or, M., May, J. H., & Spangler, W. E.

(2006). Targeted advertising strategies on television.

Management Science , 52(5), 713–725. doi:10.1287/ mnsc.1050.0489 Galeotti, A., & Moraga-Gonzalez, J. L. (2003).

Strategic targeted advertising (No. TI 03-035/1).

Tinbergen Institute Discussion Paper Series. Grossman, G. M., & Shapiro, C. (1984). Infor - mative advertising with differentiated products.

The Review of Economic Studies, 51(1), 63–81.

doi:10.2307/2297705 Hamilton, S. F. (2004). Informative advertising in concentrated, differentiated markets. Manuscript, University of Central Florida, USA Food System Research Group.

Hernández-García, J. M. (1997). Informative advertising, imperfect targeting and welfare.

Economics Letters, 55(1), 131–137. doi:10.1016/ S0165-1765(97)00057-8 Iyer, G., Soberman, D., & Villas-Boas, J. M. (2005).

The targeting of advertising. Marketing Science, 24(3), 461–476. doi:10.1287/mksc.1050.0117 Johnson, J. P. (2013). Targeted advertising and ad- vertising avoidance. The Rand Journal of Econom- ics, 44(1), 128–144. doi:10.1111/1756-2171.12014 Roy, S. (2000). Strategic segmentation of a market. In- ternational Journal of Industrial Organization , 18 (8), 1279–1290. doi:10.1016/S0167-7187(98)00052-6 Rutt, J. (2012). Targeted Advertising and Media Market Competition. Available at SSRN 2103061 .

Shaffer, G., & Zhang, Z. J. (2002). Competitive one-to-one promotions. Management Science, 48 (9), 1143–1160. doi:10.1287/mnsc.48.9.1143.172 Stahl, I. I. II, & Dale, O. (1994). Oligopolistic pricing and advertising. Journal of Economic Theory, 64(1), 162–177. doi:10.1006/jeth.1994.1060 Zhang, K., & Katona, Z. (2012). Contextual advertis - ing. Marketing Science , 31 (6), 980–994. doi:10.1287/ mksc.1120.0740 Jiang Zhao is a PhD student. His current study is focused on electronic commerce.

Shu-e Mei is a PhD supervisor. Her research fields include electronic commerce, technology innovation, and management information systems.

Wei-jun Zhong is a PhD supervisor. His research fields include information security, electronic commerce, and management information systems. Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

64 Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 APPENDIX The Procedure of Derivation in Equation (5) of the Model We arrive at the following equation:

r p c v p t A p v cp vc A = = ++ ( ) ( ) ( ) ( ) ( ) 2 Therefore, the FOC yields: = + + = = < r r pp v c p 2 02 0 2 2 ( ) ; = = = < r r p cv p t a a ( ) ( ) ; 0 0 2 2 That means the firm’s optimal price and advertising intensity can be shown in the following expressions while sending the mass advertising to the consumers: p v c p v c pvc at v c v cv cr* * ;( ) ( ) = + = + + = + + + + 2 2 2 2 2 = vc at v c at ( ) 2 4 Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 65 Definitions and Implications Table 1 shows the function of each symbol in the model.

Table 1. The definitions and implications of the symbols in the model Definitions of SymbolsImplications of the Symbols r t, Firm’s profit using mass advertising/internet-based targeted advertising * *, Firm’s optimized advertising intensity using mass /internet-based targeted advertising v Gross surplus t Transportation cost for each unit. a Advertising costs constant x xp , ( ) , The distance between the marginal consumer and the left firm; The potential demands with different price. A ( ) φ Total advertising cost Pr*,Pt * Firm’s optimized price using mass advertising /internet-based targeted advertising R Abbreviation of recognition, which indicates the probability that any member of the targeted segment is identified. T Abbreviation of accuracy, which indicates the possibility of correct prediction in the target segment. c1 The coefficient of mismatch lose Wr,Wt The total social surplus using mass advertising or internet-based targeted advertising. s The social surplus gross of transportation costs k1,k2 The constant coefficient of investment on recognition and accuracy Copyright © 2015, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

66 Journal of Electronic Commerce in Organizations, 13(1), 50-66, January-March 2015 The Procedure of Derivation in Equation (13) of the Model Use the following equations to determine the procedure of derivation in the given model:

t p c v p t R a v p tT R = ( ) ( ) ( ) 2 2 Therefore, the FOC yields: = ++ + = ++ + t pR t p v caR tT R t p v ca T 0 2 22 2 2 ; () () = = < 0 2 0 2 2 ; t p That means the firm’s optimal price can be shown in the following expression of: p v c a T t*= + + 2 4 while sending the online targeted advertising to the consumers.