Passion first, profit second: machine learning evidence for mission-driven selection and market-driven execution

Kustudić, Mijat and Kojić, Milena and Ljumović, Isidora (2026) Passion first, profit second: machine learning evidence for mission-driven selection and market-driven execution. Expert Systems with Applications, 331 (B). ISSN 1873-6793

Full text not available from this repository. (Request a copy)

Abstract

Do sustainability entrepreneurs select venture domains based on personal values or in response to market opportunities? This study applies machine learning as a theory-testing framework rather than a prediction tool, examining sequential entrepreneurial decision-making using 2,042 Kickstarter sustainability projects. Through Random Forest models and SHAP analysis, we test whether patterns in category selection differ systematically from those in execution-stage outcomes. Results reveal systematic differences between entrepreneur-controlled category selection and market-determined execution success. Economic variables show undifferentiated predictive patterns during category selection, but substantively meaningful and systematic importance during execution evaluation. This contrast aligns with theoretical accounts positing a shift from values-based to market-evaluated decision stages. Further, entrepreneurs persistently participate in underperforming sustainability categories despite clear performance hierarchies − a pattern that contradicts rational choice predictions and suggests initial domain choices resist market signals. Findings are consistent with sequential decision-making theories predicting entrepreneurs begin with values-based selection before market forces impose selection pressures. Results suggest successful sustainability entrepreneurship may require embracing this two-stage process: values-based domain selection establishing the constraint space within which market evaluation operates. This sequential perspective offers insights for platform design, entrepreneurship education, and support strategies for sustainability ventures.

Item Type: Article
Additional Information: COBISS.ID=196978185
Uncontrolled Keywords: Entrepreneurship; Effectuation; Sustainability; Crowdfunding; Machine learning; Sequential decision-making; Value-driven ventures
Research Department: Macroeconomics
Depositing User: Jelena Banovic
Date Deposited: 03 Jul 2026 12:43
Last Modified: 03 Jul 2026 12:43
URI: http://ebooks.ien.bg.ac.rs/id/eprint/2326

Actions (login required)

View Item View Item