Historically, financial institutions have relied on FICO credit scores to determine whether or not to make loans to individuals. SoFi uses customer data to personalize lending, enabling them to offer lower rates and savings for customers.
SoFi was initially focused solely on student loan refinancing. In 2016, it was estimated that students owe over $1.4 trillion in outstanding student loans. SoFi recognized an opportunity to refinance student loans at lower rates by looking at information other than credit scores to assess risk, such as career, educational background, and projected cash flow. This enabled SoFi to provide value to customers in the form of lower interest rates and significant savings. SoFi captures value in the form of interest payments from a lower risk set of customers that other institutions may have deemed not creditworthy. Sofi has since diversified and moved into other lending categories including personal loans, mortgages, and wealth management.
In 2016, Sofi had extended $8 billion in loans and achieved revenues of $650 million and is considering an IPO in the next year, with an estimated valuation of more than $4 billion. Despite this success, SoFi’s business model raises some questions. For one, SoFi maintains a highly exclusive customer base. Based on data collected by Nerd Wallet, the average approved borrower has an annual income of $130K and a credit score of 766, putting these individuals in the top 6.9% of all working professionals. The vast majority of individuals seeking refinancing do not quality for SoFi’s product, or would not benefit from the savings. This exclusivity puts SoFi at risk of attracting negative attention from regulators. Second, SoFi faces a tightening interest rate gap between private lending rates and the federal student loan rate, which they were previously able to leverage to give customers better rates.
It is not surprising that credit scores may understate credit worthiness of certain customer segments, notably millennials without credit card history and low income customers without access to traditional financial institutions. According to Forbes, an estimated 25% of Americans don’t even have a FICO credit score. These customers offer significant opportunity for lenders willing to rely on other data sources to assess risk.
Other companies are also leveraging customer data sets to make financing decisions. One interesting company in the space is Tala. Tala uses mobile phone data to make micro-loans in developing countries. Through an app, Tala collects data on users’ routine habits including the number of people someone contacts, where a user goes, and whether or not they pay bills on time. The company has found that “a person’s routine habits are more meaningful than traditional credit scoring” and is able to quickly approve borrowers and transfer money. In August of 2016, Tala had already loaned $20 million to 150,000 people and had achieved a repayment rate of 90%, notably high for emerging markets.
Tala enables lending access like SoFi, but without the exclusivity to the upper echelons of society. As customer data becomes commonplace in personalizing risk assessment, it will be interesting to see how customers and regulators react and respond.