Recent comments by Federal Reserve Governor Lael Brainard might be a shot across the bow to many upstart fintechs who are making their mark by creating new ways to underwrite. In comments at a conference in December, Brainard cautioned about the risks of using alternative data like social media information to determine creditworthiness. She also warned that innovation gone awry might result in “the digital equivalent of redlining.” My point of view, however, is that lenders big and small should not shy away from exploring new sources and uses of insightful data, but instead leverage this commentary as guidance to help focus their efforts and de-risk any solutions that they bring to market. In the business of lending money, the determination of credit risk is the key factor for success. Lenders have always invested in research & analysis to find new data that can better “split risk” and give them an edge over competitors. Historically, the most powerful sources of risk data were credit bureaus, application forms, and information from existing relationships between lender and applicant. With the advent of digital methods of applying for credit (online, mobile), the availability of new sources of underwriting data has expanded considerably. Insights are continually being uncovered with information that is native to digital channels or made easier to acquire through new technology advancements in the digital experience. New data sources might include social media information (who are your friends), online behavior information (where have you been browsing), device information (what kind of mobile phone are you using), or bank-account information sourced via new technology (what is your checking account balance). Some lenders operate exclusively in the digital space, and even traditional lenders have seen an increasing share of new applications coming from digital channels – to the point where digital applicants are now the norm. With this growth and innovation, the value of this new underwriting data to the overall ecosystem has increased exponentially. Beyond a general warning to be fair, take care of customers, and be cognizant that regulators are and will continue to scrutinize this space, Governor Brainard’s speech provides some very helpful guidance on concerns the regulators have or standards that they will emphasize. Specifically, she highlights several principles which will be useful for lenders to keep in mind for new credit data: proving the ability of the data to predict credit risk, ensuring transparency (likely for regulators and end consumers), and ensuring underwriting data is not inappropriately correlated with characteristics protected by fair lending laws. Even with significant uncertainty on how & when regulations will change, it is clear to me that a few things will be true: 1) All of the winning fintech lenders will be more strictly regulated (or bought by a regulated entity) and therefore need to heed the warnings on inappropriate use of data, 2) The risk insights and economic leverage from use of digital data will continue to grow and be important for all lenders, 3) Limitations on use of digital data will be less restrictive for activities like marketing, fighting fraud, and lending to businesses, and 4) Advanced machine learning statistical techniques will become more important to get the most power out of the ever-increasing amount of data. So the challenges remain -- Can fintechs safely build a model while navigating yet-to-be-written rules? Can traditional lenders continue to innovate and learn new tricks, or will legacy systems and risk aversion force then to hold onto underwriting strategies from the 90s? I look forward to seeing how the space evolves and being an active participant in helping lenders continue to improve their strategies.