Analyzing Billions of Transactions to Understand Consumer Behavior

Babineau and co-founder Lillian brought experience building data systems for video games, where designers needed to answer behavioral questions without writing SQL. After a hedge fund contact asked how to load two terabytes of data into Excel, they saw an opportunity to give investors tools for exploring data themselves. They focused on credit card transactions because investors wanted the data but struggled with its messiness.

Second Measure uses observed purchases to help users investigate company performance, competitors, customer cohorts, and where customers shop. For example, the team found that Stitch Fix customers did not reduce their department store spending, and that active Peloton memberships had surpassed active SoulCycle riders. The data shows spending, not a company’s costs, so it cannot establish full unit economics.

Credit card descriptions were designed for people reading statements, not for machines, so one merchant can appear under many variations. Babineau said Second Measure had more than 50 billion transactions and one billion unique transaction descriptions, and built a pipeline to resolve merchants, estimate missing locations, and correct sample biases. He also looks for careful analysts who investigate data assumptions before drawing conclusions.

The company expanded from helping investors assess company performance to serving corporations that want competitive analysis. Babineau said all 150 clients at the time had arrived through inbound referrals, including when venture capitalists introduced the product to portfolio companies. He described the product as a way for customers to answer their own questions, with shared access to data making better questions a source of differentiation.