Multihoming -the decision to design a complement to operate on multiple platforms -is becoming increasingly common in many platform markets. Perceived wisdom suggests that multihoming is beneficial for complement providers as they expand their market reach, but it reduces differentiation among competing platforms as the same complements become available on different platforms. We argue that complement providers face tradeoffs when designing their products for multiple platform architectures -they must decide how far to specialize the complement to each platform technological specifications. Because of these tradeoffs, multihoming complements can have different quality performance across platforms. In a study of the US video game industry, we find that multihoming games have lower quality performance on a technologically more complex console than on a less complex one. Also, games designed for and released on a focal platform have lower quality performance on platforms they are subsequently multihomed to. However, games that are released on the complex platform with a delay suffer a smaller drop in quality on complex platforms. This has important implications for platform competition, and for managers considering expanding their reach through multihoming. The logic is that multihoming offers potential additional revenues at limited costs for complementors, implicitly assuming that complements are the same on different platforms. But many platforms are not simple (two-sided) markets enabling transactions; they are also technology infrastructures whose features shape the development of third-party complementary products (Anderson et al. 2014;Gawer 2014, Tiwana et al.2010, Yoo et al. 2010. 2 Complements frequently must be tailored to a platform's core technological functions and interface specifications to take full advantage of its performance (Anderson et al. 2014, Claussen et al. 2015b, Tiwana 2015. In developing for multiple platforms, complementors must decide how far to specialize the complement to one platform (Schilling, 2000, Tiwana 2015, Yoo et al. 2010. We ask how the tradeoffs complementors face when designing products for multiple platforms affect the quality performance of multihoming complements across platforms. , Tiwana et al. 2010, Gawer 2014. As a system comprises the platform and its complements, the effect of platform architecture on both platform performance and complement supply matters for value creation at the system level (Cennamo 2016). the core technology components (e.g., processor cores), and the larger the number of specialized processors requiring a specific programming language for optimal utilization (i.e., no clean interface to handle the interdependencies), the more complex the platform. Complement developers must therefore choose to either use that language to optimally "conform" (Tiwana 2015) to the complex platform's specifications at the expense of integration and performance on other platforms, or to design the complement so it uses the lowest common denominator acr...