Process Mining-Driven Optimization of a Consumer Loan Approvals Process

Abstract

A real life event log of the loan and overdraft approvals process from a bank in the Netherlands is analyzed using process mining and other analytical techniques. The log consists of 262,200 events and 13,087 cases. Using a combination of traditional spreadsheet-based approaches, process-mining capabilities available in Disco and exploratory analytics using Classification and Regression Trees (CART). We examined the data in great detail and at multiple levels of granularity. In this report, we present our findings on how we developed a deep understanding of the process using the event log data, assessed potential areas of efficiency improvement within the institution’s operations and identified opportunities to use knowledge gathered during process execution to make predictions about likely eventual outcome of a loan application. We also discuss unique challenges of working with such data, and opportunities for enhancing the impact of such analyses by incorporating additional data elements that should be available internally to the bank.