Master´s Thesis Evaluation by the Opponent
Title of the Master´s Thesis:
A Machine Learning Approach to Startup Success Prediction in the Context of Venture Capital Industry
Author of the Master´s Thesis:
Bc. Tereza Kalendová
Goals of the Master´s Thesis:
The goal of the thesis was to analyse factors that predict success of start-ups funded by the venture capitalists (VCs).
Evaluation:
Criteria Description Max.
points
Points
Content 70%
Output Quality Results are well presented, discussed - substantiated, relevant and original (i.e. novelty produced by the author). They are of high practical/theoretical
relevance. 20 16
Goals The goals of the thesis are evident and accomplished.
10 8
Methodology: Methods are adequate and used correctly in relation to pre-set goals.
20 18
Theory/
Conceptualization:
Demonstration of an in-depth understanding of the topic area (state-of-the- art) including key concepts, terminology, theories, definitions, etc. based on
a literature survey. Literature review. 20 16
Formal requirements 15% Structure: The thesis is a consistent, well-organised logical whole.
3 3
Terminology: Linguistic and terminological level.
4 4
Formalities: Formal layout and requirements, extent, abstract.
4 4
Citing: Quality of citations and reflection of Ephorus results.
4 2
Delivery 15 %
Presentation document:
Is the presentation itself structured in a clear way? Is it appealing and easy to
follow? Does it convey the message efficiently? 5
Presentation skills:
Are you conveying the message efficiently and timely? Do you use
appropriate words, speed, tone of voice, gestures, movement etc. to express
your thoughts in a clear manner? 5
Argumentation: Are you able to readily and briskly react to questions or comments? Are you able to explain unclear parts and connect comments to relevant places in your presentation or parts of particular analyses? How well are you able to defend to your ideas and recommendations?
5
100 0
Other comments:
The author has studied factors predicting the success of start-ups funded by the venture capitalists (VCs). The thesis is based on a solid review of academic literature. Data on start-ups were mined from the international Crunchbase database, and the author has worked with more than 900 ths. observations covering years 1970-2013. Using several statistical approaches (logistic regression, random forest, extreme gradient boosting, support vector machine) the author estimates the probability of start-up being acquired or brought to IPO within the 7-year window. The implications from the conducted analysis recommend using random forest model to predict success or failure of the start-up and document the importance of several predictors. In overall, this is a well-written thesis and correctly conducted empirical analysis. At the same time, I also see several sources of improvement. First, the authors should be more careful when reproducing content (i.e. figures and graphs) from already published works.
Posting screenshots without publisher´s approval is not allowed. Second, the focus of the thesis is rather on technological start-ups, thus I would welcome more information in this regard. Third, some variables are questionable, given the period of analysis. How could be internet presence important for firms founded in 1970 (i.e. before the 90s)? How about total amount of aquired VC funding? Discussion of the results in light of the previously published studies is missing. The implications for the VCs are not discussed deeply. I recommend the thesis for defence
Questions or comments to be discussed during the thesis defence:
1. How could information about acquired (future) funding (i.e. not about initial firm assets) help VCs in making a decision whether to finance start-up or not?
2. What kind of variables can really help VCs in making evidence-informed-decision about the funding of a particular start-up? Can such knowledge be provided by random forest approach?
The name of the Opponent:
doc. Ing. Ondřej Dvouletý, Ph.D., MSc.
The employer of the Opponent :
Department of Entrepreneurship, Faculty of Business Administration, University of Economics, Prague