Research
My research focuses on entrepreneurial strategy, particularly how startups scale and why there are international differences in scaling. Specifically, it assesses the role of strategic decision-making and digital technologies in driving convergence and divergence in entrepreneurial growth around the world.
Publications
Can accelerators pick the most promising startup ideas no matter their provenance? Using unique data from a global accelerator where judges are randomly assigned to evaluate startups headquartered across the globe, we show that judges are less likely to recommend startups headquartered outside their home region by 4 percentage points. Back-of-the-envelope calculations suggest this discount leads judges to pass over 1 in 20 promising startups. Despite this systematic discount, we find that—in contrast to many past studies—judges can discern startup quality and are no better at evaluating local firms. These differences emerge because the pool of startups accelerator judges evaluate is both broader and less “local,” suggesting that judging ability depends on the composition of the companies they are tasked with evaluating.
This is the first study to consider the relationship between open source software (OSS) and entrepreneurship around the globe. This study measures whether country-level participation on the GitHub OSS platform affects the founding of innovative ventures, and where it does so, for what types of ventures. We estimate these effects using cross-country variation in new venture founding and OSS participation. We propose an approach using instrumental variables, and cannot reject a causal interpretation. The study finds that an increase in GitHub participation in a given country generates an increase in the number of new technology ventures within that country in the subsequent year. The evidence suggests this relationship is complementary to a country’s endowments, and does not substitute for them. In addition to this positive change in the rate of entrepreneurship, we also find a change in direction—OSS contributions lead to new ventures that are more mission- and global-oriented and are of a higher quality. Together, the results suggest that OSS can boost entrepreneurial activity, albeit with a human capital prerequisite. We consider the implications for policies that encourage OSS as a lever for stimulating entrepreneurial growth.
Selected Working Papers
Can strategy help startups compensate for constraints in local endowments like venture capital, market size, and talent pools that prior work finds important for startup performance? This paper addresses this question by interviewing executives of 253 scaling software ventures from over 30 economies and scoring the alignment of their market and organizational choices to detect whether they have a strategy, developing the first dataset of its kind. The study finds that having a strategy negatively moderates the relationship between HQ city and economy endowments and startup performance. This effect is concentrated among startups in more irreversible circumstances. The results suggest that strategy is a substitute for local endowments, with the potential to help startups overcome resource scarcities that constrain growth around the world.
Does participating in open source software (OSS) communities spur entrepreneurial growth? More efficiently developing shared code, learning from what the OSS community has developed, and shaping the direction of massive projects, such as those linked to frameworks for AI algorithms, attract many participants. Yet, contributing valuable resources to OSS, such as time and code, might give away too much, making it more difficult for firms to appropriate value from innovations. To gain a deeper understanding of participation, we analyze novel data matching accounts from GitHub—the largest OSS hosting platform—to the universe of global software venture-backed firms identified by PitchBook. We find a robustly positive relationship between OSS contributions and entrepreneurial growth, driven by both selection and treatment effects. The treatment effects account for roughly one-third of the overall impact: firms that increase GitHub contributions—in terms of the number of lines of code and number of users contributing—see an increase in their valuation and funding. Human capital, OSS policies, and market size moderate the statistical relationship between contributing to OSS and valuations, suggesting that OSS complements supply-side and demand-side country endowments. This research reveals that contributing to OSS can lead to entrepreneurial growth worldwide, with implications for policy and entrepreneurial strategy.
Attracting early adopters in new markets is crucial for startup growth. But it is not clear where startups should test to learn about demand in these markets. While testing with existing users can offer them clearer and cheaper signals, testing with new market ones can offer more transferable ones. This begs the question: (When) Do startups that test in existing or new markets attract more new market early adopters? I test this question in the context of international decisions on a digital product platform where startups post their early-stage products to attract early adopters from around the world. Taking advantage of variation in feature timing, I find that startups that test with a higher share of existing (local country) users initially attract more new market (foreign country) early adopters after they feature on the platform. This effect magnifies among startups headquartered in countries where fewer languages are spoken, making local signals clearer, but only when they are in product categories with small differences in tastes between countries where these signals are more transferable. Consistent with the idea that testing with local users gives clearer signals that improve product development, a supplementary experiment reveals that evaluators give higher ratings to products of startups that tested locally. Together, this study reveals that testing in an existing market can attract new market demand in more homogeneous contexts where signals are clearer but still transferable.
International Exposure and Entrepreneurial Pivoting (w/ Stine Grodal and Laura Huang)
How does international exposure shape entrepreneurial pivots? Through a field study of 84 startups across 27 countries, we develop a model that uncovers how international exposure not only spurs ventures to update their understandings of the international market but also generates pivots in the addressed market. Structural differences between markets, as well as entrepreneurs’ cognitive openness, makes new information about the international market more salient. This new information opens ventures’ eyes to novel opportunities. After experiencing success in the international market, ventures realize that the same opportunities always existed in the addressed market, but that their initial taken-for-granted market assumptions blinded them. Together, this model shows how international exposure can be an important mechanism shaping entrepreneurial pivoting, with implications for entrepreneurial experimentation and strategy.
Does AI Cheapen Talk? Theory and Evidence From Global Entrepreneurship and Hiring (w/ Bo Cowgill and Pablo Hernandez-Lagos)
Screening human capital based on signals such as job applications or entrepreneurial pitches is crucial for organizations. Signals are informative insofar as they are costly. Generative AI (GAI) complicates screening by lowering the cost of producing impressive signals. We model the informational effects of GAI, showing that applicants' use of GAI can increase—but also decrease—an evaluator's screening mistakes. This result depends on how GAI affects experts' signals compared to non-experts'. Using experiments in hiring and startup investing, we estimate that the adoption of ChatGPT lowers employers' and startup investors' screening accuracy by 4-9%. Consistent with our model, we also find that the use of GAI in some settings improves screening accuracy (in our case, non-English-speaking countries). These results show that GAI can profoundly shape screening accuracy.
How startups scale into new markets: Large-scale evidence from digital language tools (w/ Ed Saiedi)
Technology startups often scale by entering new markets. Doing so nearly at once as a full commitment allows them to gain more users to spur network effects, while doing so experimentally by staging market entries enables more learning. (When) do startups expand into new markets as full commitments or experimentally? We assess this question in the context of international expansion decisions. As the first study to track digital startup internationalization worldwide, we use BuiltWith data of website language tool adoption by nearly 50,000 software firms from 2001-2022. Startups, on average, adopt foreign language tools gradually across their lifecycle, even from smaller markets and with platform business models (characterized by network effects). Experimentation by smaller market startups with greater expansion incentives predicts greater internationalization. These results suggest that startups pursue market expansion experimentally.
Strategy, team alignment, and differentiation: Evidence from a global startup field experiment (w/ Chris Eaglin)