Can Venture Capitalists make better decisions by utilizing Machine Learning to predict the success of potential investments?

Advisor: Martin Diessner (martin.diessner@tum.de)

Start: ASAP

Core Issues

• Venture Capital has grown significantly in relevance in the last 20-30 years. Significant funds have been invested and top funds have achieved returns significantly exceeding the stock market. Key driver of fund performance are obviously the decisions which company to invest

• Previous research highlights, that Venture Capitalists use a large number of hard data points as well as soft factors to decide which venture to invest into, primarily with the objective to invest into ventures that have the highest expected return. It also shows, that Venture Capitalists are often biased and have difficulty integrating all data points available during their decision making process and therefore often use less-than optimal heuristics to decide whether to invest into a startup or not

• In other areas, companies and individuals have started to utilize data science and machine learning approaches to augment human decision making (e.g. manufacturing, medicine and also public equity finance)
• There are initial use cases of including such decision support systems into VC as well (e.g. EQT using their “Motherbrain” or the machine learning help system of Correlation Ventures)
• This thesis is therefore aiming to help better understand how Machine Learning could support the decision making of Venture Capitalists and which approaches are potentially fruitful to apply in the real world. This will include, among others, the following tasks:
  • Brief literature overview of known approaches and previous research
  • Review and integration of multiple large datasets provided by the chair
  • Extension of this dataset by measures chosen by the student (multiple avenues are possible here and can be discussed with the supervisor)
  • Conducting of multiple Machine Learning approaches and comparison against Random Chance model
• The thesis will be embedded into an ongoing research project which aims to identify certain design standards for a potential “Machine Learning Decision Assistant” for Venture Capital investors
• The thesis will rely heavily on quantitative analysis, previous knowledge in statistics and programming in Python (or R) is highly encouraged

Literature

• Arroyo, J., Corea, F., Jimenez-Diaz, G., & Recio-Garcia, J. A. (2019). Assessment of machine learning performance for decision support in venture capital investments. Ieee Access, 7, 124233-124243.
• Blohm, I., Antretter, T., Sirén, C., Grichnik, D., & Wincent, J. (2020). It’sa peoples game, isn’t it?! A comparison between the investment returns of business angels and machine learning algorithms. Entrepreneurship Theory and Practice, 1042258720945206.
• Calafiore, G. C., Morales, M. H., Tiozzo, V., & Marquie, S. (2020, May). A classifiers voting model for exit prediction of privately held companies. In 2020 European Control Conference (ECC) (pp. 615-620). IEEE.
• Corea, F., Bertinetti, G., & Cervellati, E. M. (2021). Hacking the venture industry: An Early-stage Startups Investment framework for data-driven investors. Machine Learning with Applications, 5, 100062.
• Ferrati, F., & Muffatto, M. (2021). Entrepreneurial Finance: Emerging Approaches Using Machine Learning and Big Data. Foundations and Trends (R) in Entrepreneurship, 17(3), 232-329.
• Krishna, A., Agrawal, A., & Choudhary, A. (2016, December). Predicting the outcome of startups: less failure, more success. In 2016 IEEE 16th International Conference on Data Mining Workshops (ICDMW) (pp. 798-805). IEEE.
• Ross, G., Das, S., Sciro, D., & Raza, H. (2021). CapitalVX: A machine learning model for startup selection and exit prediction. The Journal of Finance and Data Science, 7, 94-114.