Innovation is a cumulative process in which past knowledge created by others can be both a source for predictable outcomes and also a barrier to significant change. The recent literature on digital innovation suggests that open platforms, which encourage their developers to build upon each other’s knowledge when innovating their add-on apps in the periphery, face a related paradox. Developers face the tension of either being coherent with the past, or flexible to adjust to the future. In this paper, we examine how the trade-off between coherent and flexible search mechanisms affects the individual developer’s choice of innovating a certain app as well as his or her cumulative impact, i.e., the degree of modifications to the app. We study an open platform in the multi-disciplinary field of nanotechnology, in which 480 developers perform more than 30,000 problem-solving actions over a period of 10 years. We use relational event modeling to differentially assess the effect of the coherent and flexible search strategies. We find that developers are significantly more likely to choose a certain app that is consistent with both a coherent and flexible strategy. However, a coherent strategy leads to greater cumulative impact on an app compared to a strategy of being mutually coherent and flexible. Thus, our findings indicate both a complementary and a contradictory logic in how the tension between coherence and flexibility unfolds. We make contributions to the recent literature on digital innovation as well as the innovation literature more broadly. Further, our results inform innovation policy and platform design.

Introduction

The established innovation literature has highlighted that innovation is best described as a path-dependent, cumulative problem-solving process. In this process, innovating actors identify new opportunities and solve problems through reuse and adaptation of past solutions including those developed in other domains (Carlile, 2004; Katila and Ahuja, 2002; Murray and O’Mahony, 2007; Nelson and Winter, 1977). The literature manifests in an inter-temporal trade-off: attention to the past can be a fruitful source for “assembling new trajectories into the future” (Walsh and Ungson, 1991, p. 72), e.g., because of more predictable outcomes. However, a focus on the past can also be a barrier because it slows down change (Carlile, 2004). In this paper, we revisit this paradox of change in the context of the emerging literature on digital innovation (Nambisan et al., 2017) which suggests that the rise of digital platforms (Yoo et al., 2010) transforms how such a paradox of change unfolds at the individual-level. On digital platforms,3 innovation is primarily carried out by a heterogeneous ecosystem of third-party software developers (Lyytinen et al., 2015; Parker and Van Alstyne, 2017; Yoo et al., 2010). A platform offers its developers a stable software-based system – an extensible code base with a core set of functionalities – which they are invited to extend with their own software-based applications, or apps for short, also referred to as extensions or modules (de Reuver et al., 2018; Tiwana, 2015). A digital platform relies on a modular architecture to coordinate the distributed ecosystem of developers (Baldwin and von Hippel, 2011), with the apps in the periphery decoupled from the platform and other apps through standardized interfaces (e.g., APIs); i.e., a change in the app does not affect the core of the platform or another app.