Frenemies in Platform Markets: TheCase of Apple’s iPad vs. Amazon’sKindleRon AdnerJianqing ChenFeng ZhuWorking Paper 15-087
Frenemies in Platform Markets: TheCase of Apple’s iPad vs. Amazon’sKindleRon AdnerDartmouth CollegeJianqing ChenUniversity of Texas at DallasFeng ZhuHarvard Business SchoolWorking Paper 15-087Copyright 2015, 2016 by Ron Adner, Jianqing Chen, and Feng ZhuWorking papers are in draft form. This working paper is distributed for purposes of comment and discussion only. It maynot be reproduced without permission of the copyright holder. Copies of working papers are available from the author.
Frenemies in Platform Markets:The Case of Apple’s iPad vs. Amazon’s Kindle Ron AdnerDartmouth CollegeEmail: [email protected] ChenUniversity of Texas at DallasEmail: [email protected] ZhuHarvard UniversityEmail: [email protected] 1, 2016 We thank Carliss Baldwin, Andrei Hagiu and participants at the NYU Economics of Strategy Workshop, thePlatform Strategy Research Symposium at Boston University, the Conference of Information Systems and Technology,and the Theory in Economics of Information Systems Workshop for helpful discussion.
Frenemies in Platform Markets:The Case of Apple’s iPad vs. Amazon’s KindleAbstractWe study the compatibility decisions of two competing platforms that generate profits throughboth hardware sales and royalties from content sales. We consider a game-theoretic model inwhich the platform hardware may offer different standalone utilities to users who have differentpreferences over the two platforms. We find that incentives to establish one-way compatibility—the platform with smaller standalone value allows users of the competing platform to access itscontent—can arise from the difference in their profit foci. As the difference in the standaloneutilities increases, royalties from content sales become less important to the platform with greaterstandalone value but becomes more important for the other platform. Compatibility increasesasymmetry between the platforms’ profit foci and, when the difference in the standalone utilitiesis sufficiently large, yields greater profits for both platforms. We further show that social welfareis greater under one-way compatibility than under incompatibility, and there exist no incentivesfor either platform to establish one-way compatibility the other way round. We investigate aswell how factors such as different platform production costs, exclusive content, and endogenizedroyalty rates affect compatibility incentives.Keywords: compatibility; two-sided markets; platform competition; e-reader market1
1IntroductionAn increasing number of markets today are organized around platforms that enable consumers toaccess and/or purchase complementary goods and services. These platforms are two-sided becauseboth sides—consumers and complementors—need to gain access to the same platform in order tointeract or conduct transactions. An operating system (OS) like Windows, Macintosh, or Linux, forexample, serves as a platform connecting consumers who need an OS to access software applicationswith independent software developers who need access to the OS’s programming interface to developsoftware applications that can be sold to consumers. Other examples of two-sided markets includevideo game consoles, newspapers, smartphones, e-books, credit cards, shopping malls, and socialnetworking sites.The literature on platform-based markets has examined strategies a platform can use to growits business such as two-sided pricing (e.g., Rochet and Tirole 2003; Parker and Van Alstyne 2005;Armstrong 2006; Hagiu 2006; Seamans and Zhu 2014; Cennamo and Panico 2015), quality investment (e.g., Casadesus-Masanell and Llanes forthcoming; Zhu and Iansiti 2012), adopting innovativebusiness models (e.g., Economides and Katsamakas 2006; Casadesus-Masanell and Zhu 2010), enveloping adjacent platform markets (e.g., Eisenmann et al. 2011), and managing relationships withcomplementors (e.g., Carrillo and Tan 2008; Hagiu and Spulber 2013; Huang et al. 2013; Kapoor2013). Our work complements these studies by examining competing platform providers’ compatibility decisions.The study is motivated by empirical observations in the e-reader market, in which two majorplatforms, Apple’s iPad and Amazon’s Kindle, compete aggressively against each other. Thesedevices enable consumers to read e-books through their respective proprietary e-book readers,iBooks and Kindle Reader. The Kindle device was introduced in 2007, the iPad in 2010. AfterApple’s entry into the market, Amazon decided to make its Kindle Reader available on the Appledevice, thereby enabling consumers to read e-books purchased from Amazon on the iPad. 1 Apple,well known for rejecting third-party applications that compete directly with its own offerings,nevertheless approved Amazon’s Kindle Reader for iPad, effectively rendering the two platformsfrenemies (friends and enemies). Apple has not, however, made iBooks available for the Kindle.1As Apple takes 30% of in-app purchases, Amazon allows iPad users to use its Kindle Reader to read e-books intheir Kindle libraries. To purchase e-books, iPad users need to purchase directly from Amazon.com via web browser.2
What motivates competing platforms to choose this asymmetric equilibrium of one-way compatibility? How does such compatibility affect their profits and social welfare? How do such factors asproduction costs and exclusive content affect compatibility incentives? To answer these questions,we develop a game-theoretic model in which two competing platforms with different standalonevalue to users generate profits from both hardware sales and royalties from content sales. Bothplatforms make compatibility decisions first and then set their hardware prices, and finally userspurchase hardware and content. Compatibility is achieved when one decides to make its proprietarycontent reader available on the competing platform, and the competitor agrees.We find that incentives to establish one-way compatibility—the platform with smaller standalone value allows users of the competing platform to access its content—can arise from thedifference in profit foci. As the difference in the standalone utilities increases, royalties from content sales become less important to the platform with greater standalone value but more importantfor the other platform. Compatibility increases asymmetry between the platforms’ profit foci and,when the difference is sufficiently large, also yields greater profits for both platforms. Compatibilityincentives can also arise when the royalty from content sales is large. In such cases compatibilityreduces incentives of platforms to compete for users and thus enables the two platforms to chargehigh hardware prices. We further show that there exist no incentives for either platform to establishcompatibility the other way round, and that social welfare is greater under one-way compatibilitythan under incompatibility.Our findings relate to the e-reader market. Apple’s iPad provides many features beyond readinge-books, while Amazon’s Kindle is almost exclusively an e-book reader. As a result, in equilibrium,compared to Amazon, Apple’s hardware profits are more important to its total profits. In contrast,for Amazon, royalties from from e-book sales are more important to its total profits relative toApple.2 When this difference in profit foci is large enough, having the Kindle Reader available oniPad is agreeable to both Apple and Amazon: Amazon’s e-book sales increase because iPad userscan now purchase e-books from Amazon and read them via Kindle Reader, and Apple’s hardwaresales increase because greater value accrues to the iPad with access to Kindle Reader than in thecase of incompatibility. The additional profits Apple generates from hardware sales more than2The result is consistent with reports that Apple profits from every iPad sale, but Amazon earns no profitson Kindle sales. Source: mazon-confirms-it-makesno-profit-on-kindles/, accessed March 2015.3
compensate its loss in royalties from e-book sales through its iBooks. Similarly, the additionalprofits Amazon generate from e-book sales are greater than its loss in Kindle device sales. Inparticular, when Amazon subsidizes Kindle sales, it is always in Amazon’s interest to have KindleReader on Apple’s iPad. We also show that it is never in Apple’s or Amazon’s interest to haveiBooks available on the Kindle device.Extending our baseline model to examine a variety of factors that may affect the platforms’compatibility incentives, we find that factors that reduce (increase) asymmetry in profit foci tendto reduce (increase) incentives to become compatible. Higher production cost for the platform withgreater standalone value, for example, increases the importance of content sales to the platform.As a result, the profit foci become more similar between the two platforms, thereby reducingtheir incentives to be compatible. Exclusive content on the platform with smaller standalonevalue, on one hand, increases its reliance on content sales, and thus heterogeneity in profit fociand the likelihood of compatibility. On the other hand, exclusive content increases the platform’svalue to users, thereby reducing the difference in utilities from the two platforms. This reductionin heterogeneity reduces the likelihood of compatibility. In the end, whether exclusive contentincreases the likelihood of compatibility depends on its relative impact on utility difference andextra profits from additional content sales to the platform. We also examine the case in whichroyalty rates of e-book sales are endogenized such that they depend on market shares of e-bookreaders. We find that when market shares of e-book readers have significant influence on theroyalty rates platforms can negotiate with book publishers, platforms’ compatibility incentivesmay decrease with the difference in the standalone utilities. This is because as the difference inthe standalone utilities increases, the platform with greater standalone value may gain more profitsfrom e-book sales than from hardware sales when a greater hardware market share leads to a higherroyalty rate. Thus a greater difference in the standalone utilities may decrease the asymmetry inprofit foci of the two platforms and their incentives to be compatible.The rest of the paper is organized as follows. We discuss the related literature in Section 2. InSection 3, we present the setup for our baseline model. Equilibrium results under incompatibilityand compatibility are reported in Section 4. In Section 5, we compare the two cases and derivethe conditions under which platforms prefer compatibility over incompatibility. Extensions of ourmodel are presented in Section 7. We discuss the generalizability of our results and conclude in4
Section 8.2Literature ReviewOur model shares features elaborated in the theoretical literature on two-sided markets (e.g., Rochet and Tirole 2003; Caillaud and Jullien 2003; Hao et al. 2015; Bhargava and Choudhary 2004).Theoretical models in this literature are often industry-specific to incorporate the unique featuresof different industries. Rochet and Tirole (2003), for example, model the credit card market;Armstrong (2006) models shopping malls and newspapers; and Zhu and Iansiti (2012) model competition between video game consoles. We follow this tradition in building a model concerned withcompetition between e-reader providers.Many of the extant theoretical models focus on competition between symmetric platforms. Thefew papers examining competition between asymmetric platforms tend to focus on platforms withvery different business models. Casadesus-Masanell and Ghemawat (2006) and Economides andKatsamakas (2006), for example, investigate competition between proprietary and open sourceplatforms; Casadesus-Masanell and Zhu (2010) investigate competition between a platform that isboth subscription-based and ad-sponsored, and a platform entirely ad-sponsored. Niculescu and Wu(2014) study different business models in selling software products such as freemium and uniformseeding models. Our baseline model, in contrast, examines two platforms with similar businessmodels distinguished only by the amount of standalone value they create for users. We show thatthis difference alone yields opportunities to become frenemies.A subset of the literature on two-sided markets addresses the issue of compatibility. Doganogluand Wright (2006), examining the difference between multi-homing and compatibility, find thelatter to reduce incentives to pursue the former. Maruyama and Zennyo (2013) find compatibilityto depend on product life cycles: when most users have purchased hardware, platform profitsaccrue largely from content purchases, and then competing platforms have incentives to becomecompatible. The few studies that examine asymmetric platforms typically find weak platforms toseek compatibility in order to steal market share from stronger platforms, and stronger platformsto have no incentive to establish compatibility. Casadesus-Masanell and Ruiz-Aliseda (2009), forexample, explain large platforms’ preference for incompatibility in terms of the quest for market5
dominance, and Viecens (2011) shows that compatibility will always be preferred by a platform withsmaller standalone value and never by its competitor. Dou (2014) finds, in a model with verticallydifferentiated platforms and content, that when an inferior platform owns premium content, it isoptimal for the inferior platform to offer such content to a superior platform. The paper assumesthat one-way compatibility can be established without the permission of the rival. In contrast,in our model, content quality does not have to differ across the two platforms for compatibilityincentives to emerge. In addition, our model assumes compatibility to be a consensu