Detection of Hidden Venture Risks: Analysis of Contradictory Data for Investors

2025;
: pp. 29 - 39
1
Lviv Polytechnic National University, Department of Marketing and Logistics
2
Technical Professional College, Lviv Polytechnic National University

The venture capital and startup market is a critical driver of innovation and technological progress in the modern economy. However, evaluating early-stage startups (Seed, Series A) faces significant challenges due to high uncertainty, a lack of reliable historical financial data, and the presence of conflicting qualitative evidence. Traditional methods, such as SWOT analysis and discounted cash flow (DCF) models, are often descriptive, static, or reliant on unverified assumptions, making them inadequate for quantitative risk assessment in such environments. This article proposes and substantiates the FrameD methodology, a novel structured approach based on Dempster-Shafer theory (DST), designed to transform subjective, qualitative, and often contradictory evidence into objective, quantitative, risk- oriented metrics for venture capital investment decision-making. The methodology is built around four integrated analytical frames: Strategic, Financial, Operational, and Market. Each frame assesses critical success factors, from team competence and product-market fit to financial discipline and regulatory risks. The core innovation of FrameD lies in its ability to formally model uncertainty through belief (Bel)
and plausibility (Pl) measures and to quantitatively assess the degree of conflict (k) between different information sources. This provides investors with a dynamic, holistic tool that avoids the ‘blind spots’ of isolated analysis and adapts to a startup’s lifecycle stage. The practical application of FrameD is demonstrated through three hypothetical case studies (DCM AeroTech Innovations, DCM DataSphere Analytics, DCM FinTech Nova), simulating typical investment scenarios. The results show how the methodology successfully integrates diverse and conflicting data from various sources (market reports, expert opinions, media analysis, legal assessments), aggregates it using DST’s combination rules, and outputs clear, quantified risk assessments. The study concludes that the FrameD methodology offers a robust framework for enhancing the due diligence process, enabling venture investors to make more informed and justified investment decisions, minimize risks, and develop value-creation strategies for portfolio companies in an increasingly competitive and uncertain market.

  1. Arundale K (2020)/ Venture Capital Performance: A Com.parative Study of Investment Practices in Europe and the USA. London : Routledge
  2. Bain & Company. Global Venture Capital Outlook: The Latest Trends. Retrieved from: https://www.bain.com/insights/global-venture-capital-outlook-latest-trends-snap-chart/      .
  3. Barone, M., Bussoli, C., & Fattobene, L. (2025). The influence of financial literacy and financial self- efficacy in equity crowdfunding investment decisions. Venture Capital, 1–41. Retrieved from: https://doi.org/10.1080/ 13691066.2025.2535957
  4. Wang, X., Ihlamur, Y. (2024). An automated startup evaluation pipeline: startup success forecasting framework (SSFF). University of Oxford; Vela Partners. Retrieved from:: https://arxiv.org/abs/2405.19456.
  5. OpenAI for Startups. (2025). OpenAI. Retrieved from: https://openai.com/startups
  6. Dempster, A. P. (1967). Upper and Lower Probabilities Induced by a Multivalued Mapping // Annals of Mathematical Statistics. Vol. 38, No2. P. 325–339.Retrieved from: https://doi.org/10.1214/aoms/1177698950
  7. Shafer, G. A (1976). Mathematical Theory of Evidence. Princeton: Princeton University Press. Retrieved from:https://doi.org/10.2307/j.ctv10vm1qb
  8. Wang Y., Liu L., Zhang Y. (2022). An Adaptive Dempster-Shafer Theory of Evidence Based Trust Model in Multiagent Systems. ResearchGate. Retrieved from: https://www.researchgate.net/publication/362340029.
  9. Shenoy C., Shenoy P.P. (2002). Modeling Financial Portfolios Using Belief Functions // Studies in Fuzziness and Soft Computing. 287–315. DOI: 10.1007/978-3-7908-1798-0_11.