The Commercial Potential of Science

Zenodo (CERN European Organization for Nuclear Research) · 2026-04-24

[Updated! This version contains commercial potential predictions for over 68 million scientific articles published worldwide between 1990 and 2026.] This dataset introduces a novel index designed to predict the commercial potential of scientific articles. The index captures the probability that an article will be used by firms for the development of marketable products or processes. In addition to commercial potential, the dataset also introduces an index to predict scientific potential—the likelihood that an article will be relevant for the advance of science, regardless its commercial application. The indices are crucial for researchers focused on understanding 1) the production of science with commercial potential and 2) the pathway from academic research to market innovations and the factors that influence the commercial viability of scientific discoveries. Citation Information: If you use this dataset, please cite the article: “Masclans, R., Hasan, S., & Cohen, W. M. (2025). Measuring the Commercial Potential of Science. Strategic Management Journal, 46(9), 2199-2236.” Components of the Dataset: The dataset encompasses indices for over 30 million articles that meet the following criteria: Publication year: 1990 to 2026 Published under universities worldwide Articles in the applied and natural sciences and engineering fields Data is delivered via a single csv file. Each row contains information for a scientific article, with the following variables: ‘doi’: Digital Object Identifier—unique article identifier that can be used to match to other data sources, such as OpenAlex, Dimensions, or Web of Science. ‘compot’: commercial potential index. ‘scipot‘: scientific potential index. To develop the commercial potential index, we employed SciBert (Beltagy et al., 2019), a Large Language Model for scientific understanding. We fine tune SciBert with deep neural networks to classify scientific articles based on their potential for commercial application. We trained one predictive model per year using the text of an academic article’s abstract to generate ex-ante, out-of-sample, and out-of-training-time-period predictions of any given scientific article’s commercial potential. Licensing and Contact Information: The dataset and its components are distributed under a Creative Commons Attribution Non-Commercial license. Acknowledgments: We thank Duke University and the Kauffman Foundation for funding the creation of this dataset.

More Versus Better: Artificial Intelligence, Incentives, and the Emerging Crisis in Peer Review

Organization Science · 2026-04-27 · 4 citations

As the AI Task Force for Organization Science, we provide an early account of artificial intelligence’s (AI) impact on both submissions and reviews at a major academic journal. Submission volume has risen 42% since the late 2022 release of ChatGPT, while writing quality has declined. The rise in AI-generated writing accounts for nearly all of these trends. AI-generated writing in reviews has also increased, and is characterized by lower writing quality and less topical diversity than human-generated writing. We are, to our knowledge, the first journal to report these early impacts of AI in the review process. Conversations with editors across scientific disciplines, however, suggest that what we observe is not limited to our journal or to the social sciences. At this early stage of AI adoption, we cannot make a normative assessment about appropriate or ideal levels of AI usage. We can, however, conclude that the current state of AI tools, amplified by existing publish-or-perish incentives, appears to be pushing the system toward an equilibrium of more rather than better research. Reaching an equilibrium in which AI serves as a critical engine of innovation will require that our institutions and the incentive structures they create adapt. Funding: S. Hasan used research funding from Duke University’s Fuqua School of Business. C. Gartenberg used research funding from University of Pennsylvania’s Wharton School. Supplemental Material: The online appendix is available at https://doi.org/10.1287/orsc.2026.ed.v37.n3 .

From Model Design to Organizational Design: Complexity Redistribution and Trade-Offs in Generative AI

ArXiv.org · 2025-06-10

This paper introduces the Generality-Accuracy-Simplicity (GAS) framework to analyze how large language models (LLMs) are reshaping organizations and competitive strategy. We argue that viewing AI as a simple reduction in input costs overlooks two critical dynamics: (a) the inherent trade-offs among generality, accuracy, and simplicity, and (b) the redistribution of complexity across stakeholders. While LLMs appear to defy the traditional trade-off by offering high generality and accuracy through simple interfaces, this user-facing simplicity masks a significant shift of complexity to infrastructure, compliance, and specialized personnel. The GAS trade-off, therefore, does not disappear but is relocated from the user to the organization, creating new managerial challenges, particularly around accuracy in high-stakes applications. We contend that competitive advantage no longer stems from mere AI adoption, but from mastering this redistributed complexity through the design of abstraction layers, workflow alignment, and complementary expertise. This study advances AI strategy by clarifying how scalable cognition relocates complexity and redefines the conditions for technology integration.

20 Field Experiments in Entrepreneurship and Innovation

2025-04-07

From Model Design to Organizational Design: Complexity Redistribution and Trade-Offs in Generative AI

SSRN Electronic Journal · 2025-01-01 · 5 citations

Measuring the commercial potential of science

Strategic Management Journal · 2025-05-05 · 5 citations

Abstract Research Summary We develop an ex ante measure of commercial potential of science, an otherwise unobservable variable driving the performance of innovation‐intensive firms. To do so, we rely on large language models and neural networks to predict whether scientific articles will influence firms' use of science. Incorporating time‐varying models and the quantification of uncertainty, the measure is validated through both traditional methods and out‐of‐sample exercises, leveraging a major university's technology transfer data. To illustrate the methodological contributions of our measure, we apply it to examining the impact of university reputation and university privatization of science, finding that firms' reliance on reputation may lead to foregone opportunities, and privatization (i.e., patenting) appears to increase firms' use of the science of one university. We make our measure and method available to researchers. Managerial Summary Using machine learning, we develop a measure that estimates the probability that a scientific discovery will contribute to a commercially valuable innovation. This work addresses a key challenge: the inability to observe what scientific discoveries are worth pushing forward into commercial application. We illustrate the usefulness of this measure with two examples: 1.) firms’ use of research from prestigious universities over equally promising work from less prominent ones; and 2.) how patenting affects the diffusion of commercially relevant science across firms. For practitioners, this measure can inform R&D, licensing, and other innovation related decisions by guiding firms’ search for commercially relevant scientific research. The measure and the associated code are publicly available.

If You Had One Shot: Scale and Herding in Innovation Experiments

National Bureau of Economic Research · 2025-04-01

Solving complex problems-in medicine, engineering, and other technological domains-often requires exploring multiple approaches, particularly when significant uncertainty exists about which one will lead to success.Conventional wisdom assumes that having many experimenters independently decide which approaches to pursue increases diversity and, thus, also the chances of finding a solution.However, if experimenters herd toward the most promising approach, this convergence may reduce diversity and thus the likelihood of solving the problem.In this paper, we develop a simple model to show that, holding the total number of experiments constant, markets dominated by a few large-scale experimenters-firms conducting multiple experiments-explore more diverse approaches than markets with many single-shot experimenters.Single-shot experimenters tend to converge on the most promising approach, while multi-experimenters are more likely to diversify to avoid the correlation inherent in pursuing multiple experiments within the same approach.We test our model's predictions using data from pharmaceutical R&D.Our analysis shows that increasing the average number of experiments per firm by one unit raises target diversity by over three standard deviations.In turn, a one-standard deviation increase from the mean in target diversity boosts the likelihood of at least one experiment reaching Phase 1 clinical trials by 25.9 percentage points.Our findings inform policies for the optimal allocation of experiments across firms to maximize approach diversity and market-level success.

If You Had Only One Shot: Scale and Herding in Innovation Experiments

Academy of Management Proceedings · 2025-07-01

Solving complex problems often requires exploring multiple approaches, especially when a solution is unknown. However, the collective risk of failure likely increases when solvers concentrate on a narrow set of approaches. In this paper, we develop a simple model to argue that markets dominated by a few large-scale experimenters---firms that launch multiple experiments---produce greater diversity in approaches than markets with many small-scale experimenters. This finding challenges the conventional wisdom that having many distinct experimenters leads to many distinct approaches. We test our model's predictions using data from pharmaceutical R&D. We define an experiment as a pre-clinical trial, a market as a therapeutic class--year, and an approach as the choice of target. Our estimates suggest that a unit increase in the average number of experiments conducted per firm in a market results in an increase of over three standard deviations in the diversity of targets explored, even while controlling for the total number of firms and experiments. Furthermore, a one-standard-deviation increase in target diversity corresponds to a 25.9 percentage point increase in the likelihood that at least one experiment progresses to Phase 1 clinical trials. Our findings have implications for technology policy, highlighting the importance of optimizing the allocation of experiments across firms to maximize diversity and the likelihood of success at the market level.

The aging firm

Research in Organizational Behavior · 2025-09-11

Opinion Trumps Fact: Entrepreneurship and Competition in Political News

SSRN Electronic Journal · 2025-01-01