The expected value of applying machine learning for predictive lead scoring based on customer contact-form input - a case in the B2b energy sector

Abstract

This thesis examines the performance of machine learning to predict lead conversion probability in an early lead maturity stage based on customer contact-form data. An empirical case study was conducted developing models at two maturity stages in the lead funnel using automated machine learning and their expected value for the business was calculated. The resulting models prove the suitability of machine learning to predict lead conversion and reveal that predictions are better in a later maturity stage. Furthermore, the findings suggest including cost and benefit calculations in the development is beneficial, as not all models are profitable despite good performance