We helped Santander quantify the impact of their growth capital lending on firm KPIs including revenue, profit and employment. Our study used a two-stage causal evaluation design. The first step involved first matching firms that received Santander lending to a counterfactual control group selected from a large-scale dataset of over 2,000,000 firms from the FAME database, using a machine learning based propensity score model and distance-based methods. The second step involved undertaking regression analysis to estimate the response curve from lending to different firm KPI’s. Our approach was consistent with academically accepted causal evaluation design standards (e.g. WWC Standards) and was reviewed by internal and external experts.