Affirm: The Internet’s American Express - [Business Breakdowns, EP.249]

Affirm Holdings

• Fraser Christie · Sandstone Global InvestmentBullish

  Affirm has run a disciplined cost base, with headcount, G&A and stock comp flat for years.

  Early in its life as a listed company, Affirm was high on the list of stock-based compensation offenders. Since then management has made hard decisions: headcount has been flat for several years, G&A has been flat for half a decade, and stock-based compensation has been flat for half a decade. Fraser presents this as evidence that the company is willing to do the hard things needed to strengthen the business for the long term.

  “Headcount has been flat for a number of years now. G and A has been flat for half a decade. Stock based comp has been flat for half a decade.”at 8:35

• Fraser Christie · Sandstone Global InvestmentBullish

  Affirm's transaction-by-transaction underwriting and AI models beat credit card underwriting outcomes.

  Credit cards approve a general-purpose line and leave the consumer to manage it. Affirm underwrites every transaction separately, feeding in merchant, basket and SKU data along with current rates and the economy, and adjusts who it lends to, how much, and at what rate. Affirm also releases new transformer-based underwriting models on a six-to-seven-month cadence and says each one materially improves outcomes as more data is loaded. Affirm's book is about 50% prime versus roughly 75% at Synchrony, skews toward ages around 30, and serves consumers who carry revolving card balances, so the better outcomes come from the underwriting method rather than an easier customer pool.

  “The clearest and biggest differentiator is that they're underwriting transaction by transaction... A firm is doing every single transaction uniquely, whether that is interest rates are changing, the economy is changing, the type of purchase you're making.”at 13:26

• Fraser Christie · Sandstone Global InvestmentBullish

  Affirm earns about a 4% gross profit margin on GMV after losses and funding costs, which is industry-leading.

  Affirm's revenue comes from two sources. Short pay-in-four loans earn mostly merchant fees, and their gross profit is about 1-3% of GMV. Larger interest-bearing loans, about 70% of GMV, earn most of their revenue from consumer interest, and their gross profit is about 4-6% of GMV. Mix favours the interest-bearing loans, and loss rates on monthly instalment cohorts run around 3.5% against about 1% for short-term loans. After losses and funding costs, blended gross profit is about 4% of GMV, which Fraser calls industry-leading and healthy compared with others in the space.

  “That number is after all of the losses, all of the funding costs, all in blended, Affirm generates about a 4% gross profit margin on GMV, which is industry leading.”at 22:35

• Fraser Christie · Sandstone Global InvestmentBullish

  Affirm's merchant network effects make it an internet American Express and justify a premium multiple.

  Affirm integrates directly with merchants at checkout and holds the credit itself, so it works like American Express, which also owns both the network and the credit, rather than like Visa or Mastercard, which only supply the rails. More merchants add the button, which puts it in front of more consumers, and merchants prefer the provider with the largest consumer base. Usage shows the effect: transactions per user rose from two to seven a year since IPO, and about 95% of quarterly transactions come from existing customers. GMV above $50 billion is still growing over 30% a year, faster than a few years ago despite the business being two to three times larger. Sales and marketing cash spend is low and has barely moved since 2021. Fraser contrasts this with app-based lenders that load credit onto an account and lack the network, and says Affirm should grow larger, faster and for longer.

  “we think Affirm will grow larger, faster, for longer because of the network effects and deserves to trade at a premium multiple because of that flywheel.”at 27:09

• Fraser Christie · Sandstone Global InvestmentBullish

  US BNPL can roughly triple its ecommerce share, from about 6% toward Australia and Sweden's 15-25%.

  BNPL is about 6% of US ecommerce, but in Australia, Sweden and parts of Europe, where Afterpay and Klarna were born, the share is around 15-25%. Ecommerce is also about 20% of US retail and keeps taking share from physical retail. Fraser calls these double tailwinds: penetration of ecommerce by BNPL and ecommerce's share of retail. Together they could drive US merchant-integrated BNPL volumes to triple, which benefits Affirm as the market share leader.

  “So today, BNPL is about 6% of ecommerce in The US. That number is about 15 to 25% depending on if you picked Australia or Sweden or other European countries... Afterpay, Klarna, Affirm, that in America can triple.”at 29:07

• Fraser Christie · Sandstone Global InvestmentBullish

  Affirm holds about 50% of US BNPL, is gaining share, and grows faster than Klarna and Afterpay.

  Fraser says Affirm has about 50% of US BNPL and is gaining share. It is growing faster than Klarna and Afterpay even though it is now much larger. Klarna is roughly half Affirm's size in the US and is mostly pay-in-four, while Affirm has a longer history of larger, longer-dated interest-bearing loans, including the Peloton financing that was its original business. Affirm therefore covers a wider range of merchants and products, and once a player gains escape velocity in a network-effect business it is hard to catch.

  “It's got about 50% market share in The US, and it's gaining share. And it's actually growing faster at its larger scale than both Klarna and Afterpay.”at 29:49

• Fraser Christie · Sandstone Global InvestmentBullish

  Agentic commerce favours Affirm, since AI shopping agents will pick its merchant-subsidised 0% offers.

  Fraser concedes that financial services sold off on the fear that agents will compare offers better. He argues Affirm is built into the checkout and personalises offers by transaction and user, and merchants often subsidise its 0% loans, which no credit card can match. An agent optimising for the best financing option would therefore tend to choose Affirm, and an agent removes the friction and emotion of a human signing up for the offer. Affirm already supplies white-label credit to digital wallets such as Shop Pay Instalments and expects to do the same for agent wallets, sharing economics as it has with Shopify.

  “The Affirm product is built for ecommerce. It is ecommerce native, and it's only going to be better in a world of agents or more rational, less emotional decisions on what the best credit option is.”at 33:11

• Fraser Christie · Sandstone Global InvestmentBullish

  Only 75 of the top 250 US merchants have Affirm integrated, leaving a long enterprise sales runway.

  Only 75 of the top 250 US merchants are on the Affirm platform or have its button in checkout, so much of the growth can come from enterprise sales to merchants not yet integrated. Adoption is slow because many merchants treat card acceptance as a cost to minimise rather than a growth lever, so Affirm has to educate merchants business by business. Once a major player like Amazon adopts a BNPL product, many other e-commerce players follow quickly, which speeds up the process.

  “So interestingly, only 75 of the top 250 US merchants are on the Affirm platform or have the Affirm button integrated into their checkout. So there's still a lot of growth”at 36:49

• Fraser Christie · Sandstone Global InvestmentBullish

  Affirm Card is about 20% of GMV and growing over 100% a year, reaching in-store and non-integrated merchants.

  The Affirm card uses the Visa network to toggle between debit and credit, so Affirm can reach in-store purchases and online merchants without a direct integration. The card is about 20% of GMV and growing over 100% year on year. Affirm still underwrites every transaction, and the fees it earns are roughly the same as on interest-bearing loans with integrated merchants, so unit economics hold. Affirm also gets data on where card users shop elsewhere, which it uses as a sales pitch to merchants.

  “And that product has been very popular. So about 20% of users in GMV now comes from card, and it's still growing over a 100% year on year.”at 37:48

• Fraser Christie · Sandstone Global InvestmentBullish

  Affirm's users now transact about seven times a year, up from about twice, showing rising engagement.

  Consumers typically start by buying through an Affirm offer about twice a year, and their usage rises over time. A few years ago the average was a couple of transactions a year, and today it is about seven. Fraser cites this cohort expansion as clear evidence that the consumer value proposition works and that churn is very low.

  “A few years ago, the average was a couple of times per year, and today, the average transaction count is about seven times per year.”at 42:27

• Fraser Christie · Sandstone Global InvestmentBullish

  Walmart's switch to Klarna cost Affirm little, since Walmart was unprofitable and many shoppers kept using Affirm.

  Walmart turned off Affirm and switched to its own Klarna-powered BNPL offer. The economic impact on Affirm was small for two reasons. Walmart was not a very profitable merchant for Affirm because it squeezes suppliers hard and so earns thin economics despite high volume. Affirm loses the customer acquisition cost, but a white-labelled Walmart product is less costly to acquire. The company also disclosed that most customers still shop at Walmart through Affirm by using the Affirm card, so loyalty to the Affirm brand persisted.

  “What is interesting is the economic impact to Affirm was actually quite minimal for two reasons. The first is that wasn't a very profitable merchant.”at 43:04

• Fraser Christie · Sandstone Global InvestmentBullish

  70-80% of Affirm's code is already deployed by AI, with G&A kept flat for five years.

  Affirm's engineering-led leadership, including founder Max Levchin, is extremely technical. About 70-80% of code is already deployed by AI, and AI is also embedded in underwriting models. G&A has been essentially flat in dollar terms for five years and SBC for six, so the company is growing without adding corporate cost. Fraser argues this culture makes it an unfair comparison with other credit card businesses.

  “70 or 80% of code is already being deployed by AI. You compare that culture and that leadership team with every other credit card business on earth, it's just not a fair fight or a fair comparison.”at 49:40

• Fraser Christie · Sandstone Global InvestmentBullish

  Regional banks will white-label Affirm's credit on their debit cards, making Affirm a payments network.

  Regional banks have debit cards but largely avoid credit cards because the space is too competitive. Through Affirm's Edge offering, a bank can turn on debit-to-credit functionality, with Affirm handling underwriting, servicing and the repayment functionality in the bank's app. This extends Affirm's reach beyond its merchant integrations, and the company describes itself as a payments network. The same approach already appears in Shop Pay installments and co-branded airline cards.

  “We haven't talked about it, but Edge is another example where banks are going to start white labeling Affirm's offering.”at 51:59

• Fraser Christie · Sandstone Global InvestmentBullish

  Affirm's diversified, redundant funding base should keep it resilient if private credit or banks wobble.

  Affirm funds loans through many sources: banks, insurers and institutions, securitised and unsecuritised, on and off balance sheet, static and revolving, and fixed and floating rate. Because no single source dominates, a problem in private credit, banks or insurers would not stop funding unless the whole system dried up. About 50% of loans sit off balance sheet in forward-flow and static ABS facilities and about 50% on balance sheet in warehouses and revolving ABS. The split is consistent and actively managed. The company is also getting a banking license and will add consumer deposits to its funding mix.

  “The key principle to think about on the funding side of the business is a firm is trying to build out a diversified range of funding sources with lots of redundancy.”at 53:48

• Fraser Christie · Sandstone Global InvestmentBullish

  Affirm's network effects make it more like American Express than Capital One or Synchrony, and a better business.

  Affirm's network effects, with merchants and consumers both on the platform, look more like Amex's model than the balance-sheet lending of Capital One or Synchrony. Credit also hits Affirm's P&L differently. Its fixed-rate loans avoid the revolving-balance credit losses that hit card issuers in downturns, so it is a healthier and lower-risk business. The fair counterpoint is that Amex serves an affluent customer base while Affirm serves the average American. Affirm also benefits from a faster-growing segment, since BNPL is taking share of e-commerce and retail from traditional card networks.

  “the network effects alone make the business more like an American Express than a Capital One or a Synchrony, but also just how the credit hits the p and l and what happens in these downturns also makes Affirm a better business than the Capital Ones and the Synchrony's.”at 59:08

• Fraser Christie · Sandstone Global InvestmentBullish

  Affirm can reach $100B GMV with 25%+ growth, a small slice of the huge US card market.

  Affirm delivered about $50B of GMV to June and targets $100B with greater than 25% annual growth, which would get it there within a few years in theory. The target covers core US, card and international, while Edge and agentic commerce would add upside on top. The runway is large: US credit card balances are about $1.3T against Affirm's roughly $20B loan book, and Visa, Mastercard and Amex together process about $13T of volume a year, versus Affirm's $50B. Capital One processes over 15 times Affirm's GMV. Management frames the $100B figure as a stepping stone, not an end point.

  “They're targeting a 100,000,000,000 greater than 25% annual growth on that path, so it works out. They'll get there within a few years in theory.”at 1:01:57

• Fraser Christie · Sandstone Global InvestmentBullish

  BNPL can triple its share of US e-commerce, and Affirm already holds about half of the BNPL market.

  E-commerce is still taking share of US retail, at a little over 20%. Benchmarks from other countries suggest BNPL can still triple its share of US e-commerce. Affirm holds about 50% of the BNPL market and is expanding its share, so it captures much of that tailwind.

  “BNPL can still triple its share of US ecommerce if you look at benchmarks elsewhere in the world, and Affirm is about 50% market share of BNPL and expanding.”at 1:02:57

• Fraser Christie · Sandstone Global InvestmentBullish

  Affirm can win head-to-head global merchants from Klarna over time, though Klarna leads internationally today.

  Klarna has a much stronger international business, with a large merchant and consumer base, so global merchants such as Airbnb and eBay chose it for one global button. Affirm cannot yet offer that globally. Merchants usually choose the platform with the most users because it drives the most transactions, which currently favours Klarna abroad. As Affirm opens in more regions, head-to-head competition with Klarna should get stronger, and Affirm can use its existing US merchant relationships such as Shopify and Costco to enter international markets.

  “Over time, as [Affirm] opens in more regions, we think that head to head on Klarna for some of those merchants actually gets stronger so that there's no reason why anyone shouldn't go with [Affirm].”at 1:06:22

• Fraser Christie · Sandstone Global InvestmentBullish

  Losing Amazon or Shopify would hurt Affirm, but long-term contracts make that surprising and unlikely.

  Amazon and Shopify are among Affirm's largest partners. Losing either would hurt the customer acquisition flywheel, so concentration is a real risk. However, Affirm has very long-term agreements with both, and the partnerships have been very successful for all parties. Fraser says it would surprise him if those relationships ended, though the risk is real.

  “There are very long term agreements in place with both of those companies. ... So it would surprise us if those relationships were to end, but it obviously is a risk.”at 1:08:43

• Fraser Christie · Sandstone Global InvestmentBullish

  Rate hikes cause a one-to-two-quarter growth and margin dip, not an existential credit risk for Affirm.

  Affirm's loans are fixed-rate for their 11-12 month average term, and only a small part of its funding is floating. When the Fed raises rates quickly, the book reprices within one to two quarters, creating an air pocket. Affirm also stops writing loans to borderline customers at the higher rate, so growth slows, but loans already outstanding keep their fixed rate, so credit losses do not blow out as they would on revolving credit card balances. Management said a 100bp rise would hit P&L by about $90M against roughly $2.5B of gross profit. The 2022 quarter was the worst case, and the company recovered within two quarters. The share price can still be hurt by major rate hikes.

  “I think you can get comfort on the credit. Not to say the share price won't get hurt if there's major interest rate hikes, but it's not an existential risk, I think, the way some people would otherwise think of it.”at 1:10:22

• Fraser Christie · Sandstone Global InvestmentBullish

  Affirm's marriage of tech and underwriting talent is a competitive advantage that AI makes more valuable.

  Large platforms such as Google, Amazon, Apple and Shopify have no interest in extending credit to their customers and outsource it, because underwriting is hard, requires a lot of expertise, and distracts from their core businesses. That leaves room for a specialist. Affirm has built its advantage by combining strong tech talent with strong underwriting talent. This advantage existed before AI became a major factor, and Fraser's team thinks it is a competitive advantage in the new AI environment too.

  “And so a firm [Affirm] has been able to build its competitive advantage by marrying its tech talent and its underwriting talent, and that gives it a competitive advantage previously before AI became a huge factor and we think a competitive advantage in this new environment.”at 1:12:13