Which Factors Matter Most? Can Startup Valuation be Micro-Targeted?

TL;DR

This paper investigates how categorical factors like geography, sector, and urban location influence startup valuations, using hierarchical models and machine learning to identify key determinants and build a valuation scorecard.

Contribution

It introduces a method combining econometric and machine learning techniques to analyze categorical variables' impact on startup valuation, enabling micro-targeted valuation insights.

Findings

Categorical variables significantly influence startup valuation.

Hierarchical models identify key factors affecting valuation.

Machine learning techniques reveal complex interactions among variables.

Abstract

While startup valuations are influenced by revenues, risks, age, and macroeconomic conditions, specific causality is traditionally a black box. Because valuations are not disclosed, roles played by other factors (industry, geography, and intellectual property) can often only be guessed at. VC valuation research indicates the importance of establishing a factor-hierarchy to better understand startup valuations and their dynamics, suggesting the wisdom of hiring data-scientists for this purpose. Bespoke understanding can be established via construction of hierarchical prediction models based on decision trees and random forests. These have the advantage of understanding which factors matter most. In combination with OLS, the also tell us the circumstances of when specific causalities apply. This study explores the deterministic role of categorical variables on the valuation of start-ups (i.e. the joint-combination geographic, urban, and sectoral denomination-variables), in order to be able to build a generalized valuation scorecard approach. Using a dataset of 1,091 venture-capital investments, containing 1,044 unique EU and EEA, this study examines microeconomic, sectoral, and local-level impacts on startup valuation. In principle, the study relies on Fixedeffects and Joint-fixed-effects regressions as well as the analysis and exploration of divergent micropopulations and fault-lines by means of non-parametric approaches combining econometric and machinelearning techniques.