Max Levchin claims any single data point that seems to dramatically improve underwriting accuracy is a red herring. He argues these 'magic bullets' are brittle and fail when market conditions shift. A robust risk model instead relies on aggregating small lifts from many subtle factors.
Related Insights
Identifying flawed investments, especially in opaque markets like private credit, is rarely about one decisive discovery. It involves assembling a 'mosaic' from many small pieces of information and red flags. This gradual build-up of evidence is what allows for an early, profitable exit before negatives become obvious to all.
Effective due diligence isn't a checklist, but the collection of many small data points—revenue, team retention, customer love, CVC interest. A strong investment is a "beam" where all points align positively. Any misalignment creates doubt and likely signals a "no," adhering to the "if it's not a hell yes, it's a no" rule.
A key operational use of AI at Affirm is for regulatory compliance. The company deploys models to automatically scan thousands of merchant websites and ads, flagging incorrect or misleading claims about its financing products for which Affirm itself is legally responsible.
Affirm's CEO suggests competitors don't report payment data to credit bureaus as a business strategy. By keeping delinquencies off the 'permanent record,' they can implicitly encourage late payments, from which they profit via fees. Affirm, having no late fees, advocates for full reporting.
Max Levchin's firsthand struggle with hidden fees and the long-term impact of a credit card mistake—even after his PayPal success—was the direct catalyst for founding Affirm. The goal was to build a transparent lending model born from personal pain.
Merchants pay BNPL providers like Affirm more than credit card processors for three key benefits: converting hesitant buyers ('incremental sales'), ensuring high approval rates so the option is useful, and protecting their brand from association with lenders who charge punitive fees.
Affirm's CEO argues the core flaw of credit cards is not high APRs, but a business model that profits from consumer mistakes. Lenders are incentivized by compounding interest and late fees, meaning they benefit when customers take longer to pay and stumble.
Afeyan distinguishes risk (known probabilities) from uncertainty (unknown probabilities). Since breakthrough innovation deals with the unknown, traditional risk/reward models fail. The correct strategy is not to mitigate risk but to pursue multiple, diverse options to navigate uncertainty.
By eliminating late fees and compounding interest, Affirm removes any financial upside from borrower mistakes. This forces the company's business model to depend solely on successful repayment, demanding superior, transaction-by-transaction underwriting to survive.