The Dynamic Effects of Bundling as a Product Strategy

Several key questions in bundling have not been empirically examined: Is mixed bundling more effective than pure bundling or pure components? Does correlation in consumer valuations make bundling more or less effective? Does bundling serve as a complement or substitute to network effects? To address these questions, we develop a consumer-choice model from micro-foundations to capture the essentials of our setting, the handheld video game market. We provide a framework to understand the dynamic, long-term impacts of bundling on demand. The primary explanation for the profitability of bundling relies on homogenization of consumer valuations for the bundle, allowing the firm to extract more surplus. We find bundling can be effective through a novel and previously unexamined mechanism of dynamic consumer segmentation, which operates independent of the homogenization effect, and can in fact be stronger when the homogenization effect is weaker. We also find that bundles are treated as separate products (distinct from component products) by consumers. Sales of both hardware and software components decrease in the absence of bundling, and consumers who had previously purchased bundles might delay purchases, resulting in lower revenues. We also find that mixed bundling dominates pure bundling and pure components in terms of both hardware and software revenues. Investigating the link between bundling and indirect network effects, we find that they act as substitute strategies, with a lower relative effectiveness for bundling when network effects are stronger. ⇤Both authors contributed equally to this paper, and are listed in alphabetical order. Tim Derdenger is Assistant Professor in Marketing & Strategy, Tepper School of Business, Carnegie Mellon University. e-mail:derdenge@andrew.cmu.edu. Vineet Kumar is Assistant Professor of Business Administration, Marketing Unit, Harvard Business School, Harvard University. e-mail:vkumar@hbs.edu. The authors would like to thank Sunil Gupta, Brett Gordon and Minjung Park and for comments on an earlier draft of the paper. They gratefully acknowledge feedback from seminar participants at Yale University, Catholic University of Leuven, and the University of Zurich as well as conference participants at Marketing Science, Northeastern Marketing Conference and the UT Dallas FORMS conference. All errors remain their own. 1