Many organizations complement their search for innovation by outsourcing innovation-related tasks to external “crowds”. Set up as an innovation contest in which participants can win a prize, individuals self-select into working on and submitting their solution to the problem independently from each other. In many contests, individuals can also cooperate, i.e. they can see others’ ideas and comment on them after which they can decide whether or not to incorporate the knowledge into their submission. This creates a situation of individual-level coopetition, fundamentally changing the nature of the contest with unclear consequences for the resulting quality of the submitted ideas. In this paper, we investigate individual-level coopetition in innovation contests and study processes and effects within this emerging organizational form of innovation search. In Study 1, we report evidence from a lab experiment with 294 participants. Coopetition exhibits no effect on the novelty of an idea while it has a positive effect on its user benefit and a negative effect on its feasibility. In Study 2, we explore the underlying coopetition processes and identify patterns of knowledge exchange based on data from a qualitative content analysis. We find marked differences in the processes driving the three dimensions of idea quality, suggesting that cooperation between the contestants involves trade-offs.

Many platforms hosting innovation contests allow their participants to share ideas and knowledge with each other and many participants take advantage of this opportunity, questioning common beliefs that competitive rewards lead individuals to withhold knowledge and impair the progress of others while cooperative rewards promote knowledge exchange. In this paper, we build on social interdependence theory to analyze how and when competitive reward structures influence the likelihood that contestants cooperate. Using an experimental vignette study, we first test the effects of competition on cooperative behavior (Study 1) and find that competitive reward structures negatively impact an individual’s likelihood to cooperate, with the effect being more pronounced for sharing compared to absorbing knowledge. In Study 2, we apply a multi-stage configurational approach to first identify, categorize and specify attributes that influence individuals’ decision to cooperate (or not) and then explore their importance weights and inter-relatedness in an adaptive choice-based conjoint analysis combined with an ensemble cluster analysis. We find that it is not a single, but bundles of attributes dominated by either the fear of misappropriation in all competitive reward structures or by extrinsic motivation in the non- and lowly competitive conditions and intrinsic motivation in the highly competitive condition. Our findings hold important implications for the design of innovation platforms and contests. 

Creative production increasingly happens in teams. However, despite the importance and long tradition of team research we lack a good understanding of core aspects: How do successful teams form in the first place? What role does intellectual distance between individuals play for team formation and subsequent team output and performance? We theorize that collaborators face a complicated trade-off regarding the intellectual distance of their team partner: intellectually closer (distant) team partners face lower (higher) coordination costs but less (more) diverse knowledge recombination. Using data data from Threadless, a large online platform for T-shirt designs, on which participants can form teams and submit design ideas for evaluation. To analyze creative production, we collected images of all designs submitted to the platform. We use deep neural networks to situate designs (and thus the individuals who produced them) in a conceptual “design space” by computing image embeddings in multidimensional vector space. We then measure intellectual distance as the cosine similarity between individuals' prior designs. We find that individuals are more likely to choose individuals as team partners that are either very similar or very distant. At the same time, intellectual distance between individuals relates positively to the novelty of the team output. The novelty of the team output then has a u-shaped relationship with performance (i.e., the vote average of the design). Our findings contribute to the literature on team formation, knowledge diversity, and team performance.

Crowdsourcing is viewed as a solution to distant search by attracting a large number of diverse and distant ideas. While previous literature has focused mainly on the generation of these crowdsourcing outcomes, we know little about firms' translation of these outcomes into innovation performance. Moreover, evidence suggests that crowdsourcing is of low importance to firms and frequently fails. Hence, we need a better understanding of the boundary conditions that specify under which conditions the ideas generated through crowdsourcing translate into firms’ innovation performance. I argue that the benefits of crowdsourcing also form its drawbacks, as the evaluation and exploitation of such a vast number of solutions posit significant information-processing challenges to the firm. Drawing on the literature on information processing and digitalization in organizations, I theorize that digital capabilities can be such a boundary condition that enables the translation of crowdsourced outcomes into firm innovation performance. Using a sample of 1,657 German firms, I show that crowdsourcing relates to higher innovation performance, but the extent depends strongly on the firm's digital capabilities. These findings advance our understanding of the role of digital capabilities in distant search, the relationship between crowdsourcing as an innovation method and innovation performance, as well as the benefits of digital capabilities in organizations to overcome information-processing challenges in distant search.

The determination of an organizational design is a major milestone in the evolution of successful startups. In particular, the introduction of middle management frees up scarce founder attention for processing important information and make strategic decisions. However, digitalization, i.e. the use digital tools and technologies, changes the volume and nature of information for processing in startups. We theorize based on information-processing theory that startups respond to these new, digital information challenges with an increased proclivity to professionalize and introduce middle management, especially when founder attention is scarce and other startup activities demand attention. We test our hypotheses using a sample of 1,441 startups founded in Germany between 2010 and 2015 and find that increasingly digitalized startups rely relatively early on middle management in their lifecycle. This effect is stronger when founders lack prior management experience and the startup is comparatively large. Our findings have important implications for theory on the emergence of organizational designs in startups and the consequences of digitalization for the allocation of entrepreneurial attention. 

The delegation of decision-making authority from founders to dedicated managers is an important step in the professionalization of startups. However, such delegation changes the distinct nature of startups as workplaces with frequent interaction between founders and their employees. We build on theory of relational disruption in workplaces and theorize how delegation of decision-making authority affects employee retention. We reason that through delegation, employees lose opportunities to learn from founders and influence startup decisions. As a result, relational advantages of startup employment decrease, leading to employee mobility. Moreover, employees are particularly likely to leave after decision-making has been delegated when they are early in their careers and when founder teams are small. Using a sample of 13,737 employees in 1,797 German startups, we find support for our conjectures.