Order 1079275: Entrepreneurship - opportunity analysis video game industry in Australia

For further information contact: Richard T. Gretz, Associate Professor of Marketing, University of Texas at San Antonio ([email protected]).

2017 Summer AMA Proceedings F-35

Research Question Bundling in hardware/software industries is increasingly common. Especially bundling hardware with digital content. However managerial questions regarding bundle formation have not been adequately explored. For example, how do managers on both the hardware and software side evaluate each other as potential bundle partners? We examine this question using a structural two-sided matching model to account for the simultaneous decisions being made by both the hardware and software providers to bundle (see Yang, Shi, and Goldfarb 2009).

Method and Data We employ a modification of the two-sided matching frame- work with endogenous transferable utility developed in Fox (2010). This approach is useful in contexts where we observe bundles but we do not observe any endogenous transfers (i.e. payments) between hardware and software makers.

Using Fox’s (2010) terminology, each possible match has a “production value” which is the excess value the match pro- vides to the agents beyond being unmatched. The “production value” is based on the interaction of agent characteristics (e.g. the interaction of relevant hardware characteristics and rele- vant software characteristics). Observed matches are a “local production maximum” based on the game theoretic equilib- rium concept of pairwise stability (Fox 2010) where all matched agents prefer their observed match to exchanging

partners to form new match. For example, if we observe 4 agents, say A, B, C, and D, and 2 matches, say A & B; C & D, all agents must prefer their match rather than exchanging part- ners to form 2 different matches, or A & B; C & D is preferred to A & C; B & D; observed matches are a local production maximum in that the joint production value A & B + C & D ≥ A & C + B & D. In other words, the combined “production value” of A & B (based on the interaction of the relevant char- acteristics of A with the relevant characteristics of B) and C & D has to at least as high, if not higher, than the combined hypothetical “production value” of A & C and B & D.

We use this framework to generate “production value” inequalities that must hold in order for the observed bundles in our dataset to be preferred to potential bundles that are not observed. Our main dataset comes from the NPD group (a market research firm) covering monthly industry sales in the U.S. from January 1995 to September 2010. In total, we observe 117 bundles with 98 distinct bundle introductions. We generate 40908 inequalities from comparing actual bun- dle introductions to potential bundle introductions for all periods where we observe at least one bundle introduction. We then find the parameter values that satisfy the greatest number of inequalities.

Summary of Findings The interplay of console and game characteristics in our matching model gives insights into bundling arrangements

A Matching Model for Hardware and Software Bundles and an Application to the U.S. Home Video Game Industry Richard T. Gretz, University of Texas at San Antonio B.J. Allen, University of Texas at San Antonio Suman Basuroy, University of Texas at San Antonio

Keywords: bundling, hardware and software, matching model, video games Description: We develop a framework for the simultaneous decisions made by both the hardware and software providers to bundle using a structural two-sided matching model.

EXTENDED ABSTRACT

F-36 2017 Summer AMA Proceedings

most favored by console and game makers. High quality games are valued most by older consoles and market lag- gards. However, game makers with the better history of pro- ducing killer applications are valued most by older consoles and market leaders. With respect to the latter, game makers with a history of producing hits likely have more options and prefer to bundle with experienced market leaders. With respect to the former, managers of lagging consoles likely seek high quality games in an effort to keep their console relevant. High quality game makers will have a stronger negotiating position in this case, especially when the lagging console is trying to attract additional consumers later in the console lifecycle. Additionally, we find that first-party games are preferred by older consoles and laggards. For managers of these consoles, first-party games are likely suit-

able alternatives to dealing with third-party game makers who may be in a dominant bargaining position.

Key Contributions Our paper makes important contributions to the marketing field by providing a framework to examine the supply side of bundle formation in the hardware\software context. We then apply this framework to examine hardware/software bundling decisions in the U.S. home video game industry. Our matching model gives insight into the interplay between products characteristics that potential bundle partners tend to prefer. We believe that our framework can meaningfully be extended to other industries and other high-tech markets.

References are available on request.

Copyright of AMA Summer Educators' Conference Proceedings is the property of American Marketing Association and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.