a16z Podcast | Building Affirm, and Why Max Levchin Has Watched Seven Samurai 100-Plus Times

Affirm's Origin and Market Opportunity

• Max Levchin founded Affirm to address the stagnation of the FICO credit scoring model, which was designed in the 1980s and fails to account for the gig economy's fluid job markets.
• Post-2008 financial crisis data shows young consumers entering the workforce are increasingly distrustful of traditional banks, with two-thirds lacking credit cards and four of the top ten most "despised brands" being banks.
• Affirm positions itself as a transparent "purchase finance service" for "considered purchases" ($200–$3,000+), offering pre-calculated interest with no hard credit pulls or hidden fees.
• The company aims to build a trusted brand for millennials and Gen Z by avoiding late fees; Affirm takes responsibility for missed reminders, absorbing the cost of errors to maintain consumer trust.

Business Model and Merchant Impact

• Affirm integrates at online and offline points of sale, converting "window shoppers" into buyers; this has driven a 30% sales increase for participating merchants, with Casper cited as a primary partner.
• Customer acquisition relies on on-demand timing, where merchants introduce Affirm's installment plans at the moment of purchase need, contrasting with traditional bank branch models.
• User trust builds through a data-driven retention loop: 85% of subsequent purchases from a brand occur within 45 days of the largest purchase, and the average second Affirm transaction is 88% larger than the first.
• The long-term strategy involves transitioning from a point-of-sale lender to a comprehensive financial system for young people, including student loans and eventually mortgages.

Risk Underwriting and Data Strategy

• Levchin asserts that successful fraud and risk models rely exclusively on internal transaction data, noting that external social signals (e.g., Facebook friends, eating habits) are inferior predictors.
• Building a robust risk database requires "paying tuition" by accepting initial losses; Levchin noted PayPal lost millions weekly to fraud in 2000, a necessary step to train their underwriting algorithms.
• Affirm has reunited its original PayPal risk team, led by Nathan, to apply similar rigor to new loan books, acknowledging that new credit products typically lose 50% of dollar volume in the first several months.
• The company rejects "toxic" financing structures like point-of-sale cards that penalize a single late payment with retroactive maximum interest rates.

Future of Banking and Partnerships

• Levchin predicts traditional banks as currently structured will not exist in the future, replaced by digital entities that restore the "trusted advisor" relationship lost after the 2008 crisis.
• Affirm views the financial sector as evolving into two tracks: transactional, disaggregated services (e.g., debt refinancing via Credit Karma) and a primary "trusted bank" for daily deposits and bill pay.
• The company plans to partner with large institutions like JPMorgan Chase, leveraging their capital efficiency and loan portfolio purchasing power while maintaining their own underwriting standards.
• Levchin argues that legacy banks cannot successfully court the millennial demographic due to brand perception, making partnerships with fintechs like Affirm essential for accessing young consumers.

Personal Philosophy: "Seven Samurai"

• Levchin has viewed Akira Kurosawa's Seven Samurai over 110 times, using the film as a core metaphor for startup team building and crisis management.
• He identifies with the character Gambe, focusing on assembling a team from imperfect ingredients (e.g., a "joker" turned samurai, untrained peasants) and training them for high-stakes environments.
• The film is used in management training to illustrate the necessity of discipline and unity ("deserters will be killed") to prevent individual fear from compromising the collective mission.
• Levchin's obsession with the film led to an academic "A-plus" paper written entirely from memory after memorizing every scene to bypass a 25-page written requirement for his Asian literature certification.