In the years leading up to the pandemic, the expansion of Fintech institutions led the Mexican financial market to a process of unprecedented growth.
The underlying business model of many Fintechs is to grant loans to people who are neglected by traditional financial institutions. That is, people who have little to no credit history or access to the paperwork needed to start a credit application at a bank.
Traditional credit scores are widely inaccurate when applied to this population of applicants. Given this new target market and the exponential increase in the number of Fintech institutions, my team and I anticipated the demand for a specialized credit score that would help lenders estimate the level of risk of potential Fintech clients.
Throughout 2020, I designed the Fintech Score from start to finish - I analyzed the marked and talked to the biggest Fintechs in the country to understand their needs. I designed a sample population, defined the target variable and carefully selected the most relevant features among thousands of plausible attributes to create a model that could accurately predict the level of risk of a Fintech applicant.
Designing the Fintech Score stands as the greatest challenge I have faced in my professional career. In the end, the product was a great success, with multiple Fintechs consuming the score in real time every time they receive an application. The published model can be found here.