The venture capitalist Marc Andreessen has pointed out that every failed idea from the 2000 tech bubble—think online pet food or grocery delivery—has now worked, 20 years later. And therefore, his job as a venture capitalist is to be open minded and fund every new idea, because whether ideas will work or not is just a matter of timing.
Buy now, pay later (BNPL) is a very old idea whose time has finally come.
One of the largest BNPL companies, Affirm, became publicly traded earlier this year, and we became shareholders. Affirm has a very clear vision: build honest financial products that improve lives. Its founder, Max Levchin, is an immigrant and computer scientist from Stanford University who was taking a nap in a classroom during the Palo Alto summer where an investor, Peter Thiel, was giving a lecture. The year was 1998. Levchin introduced himself to Thiel and the pair formed a company that pivoted into what later became PayPal. In 2012, Levchin founded Affirm.
Affirm fits the mold of what we look for in investments: mission-driven, founder-led companies riding one or more technology adoption curves, building a durable moat, and with ample room for reinvestment and growth.
In Brazil, where I grew up, we had a retailer called Casas Bahia that used to do installment financing. Customers would walk into the physical store and buy an otherwise unaffordable item—like a household appliance—in installments over several months. A Michigan Business School case study from 2003 explains:
“In order to serve the poor population, Casas Bahia developed an innovative approach. Part of the solution is the now famous "carnê," or passbook, that allows its customers to make small installment payments for the merchandise. Payment schedules range from one to 15 months. The passbook is payable only at Casas Bahia stores, and every month consumers must enter a store to pay their bill. This method also helps to maintain relationships with clients. The financed sale is responsible for 90% of all sales volume: 6% are cash payments and 4% are credit card.
All customers who wish to finance their purchase must submit to an SPC Credit check [similar to a FICO score in the US]. If the customer has a negative SPC score, Casas Bahia will not be able to complete the transaction until the customer resolves his credit problem. If the customer has a positive score, there are two alternatives. If the merchandise costs less than R$600, no proof of income is required; a valid permanent address will suffice. If the merchandise costs more than R$600, Casas Bahia has developed a proprietary system to evaluate the prospective client. They receive a credit limit based on total income, both formal and informal, occupation and presumed expenses. This "scoring" process takes less than one minute. If the system approves the prospect, the salesperson can continue with the sale. If the client is rejected by the system, he or she is directed to a credit analyst for further evaluation. This is where the importance of building a relationship is prominent. Based on training, the credit analyst will ask a series of questions to determine a client's credit worthiness. The entire process typically is finished in 10 minutes or less.”
Every step of this transaction has now been replaced by software:
The “carnê,” or passbook, is now the Affirm app on your phone
Rather than visiting the store to make payments—giving the store the opportunity to sell you more things—you instead open the Affirm app, giving merchants that have partnered with Affirm prime location for more sales
Payment schedules aren’t limited to 15 months; they go as long as 60 months on Affirm, and loans range from $50 to over $17,000
The manual credit scoring system taking “10 minutes or less” is now a machine learning system that “takes five top-of-mind data inputs and turns them into a total of over 200 data points” and delivers a result in seconds
Rather than being a captive system only usable at Casas Bahia, Affirm is a global system that can be used anywhere. Instead of simply offering loans with high interest rates, Affirm offers the traditional 10-30% APR financing, but also 0% APR. And rather than being a solution for the poor, Affirm can be a solution for anyone. Even wealthy customers appreciate the benefit of paying over time at a cost of 0% rather than upfront.
As an example, a $1,495 Peloton bike gets sold for a headline $1,495 to the customer, but behind the scenes, Peloton is taking a discounted price and Affirm is closing the sale by offering a $39 per month payment at 0% APR. This discounted price is called the merchant discount rate, or MDR. The budget for the 0% APR comes from the merchants’ marketing budgets: it’s a new form of customer acquisition cost.
MDRs can range from 4-8% on short-term financing up to 15%+ on longer term commitments. On its face, this sounds like a bad deal for merchants: why would they go from paying 2-3% to a credit card processor, to paying 4-15%+ to Affirm?
Merchants do it because they gain several benefits:
Increased conversion: this means a bigger proportion of customers that want to buy an item end up actually clicking that buy button
More sales: for many customers, a one-time payment of $1,495 is a non-starter, but $39 per month fits their budget. The merchant makes sales it wouldn’t otherwise make
New customer acquisition: acquiring customers is often the largest expense for a merchant; by helping win the customer, Affirm can earn budgets previously dedicated to other forms of marketing
Higher average order value (AOV): by splitting payments, customers can afford larger shopping carts. Affirm’s merchants report about 85 percent higher AOVs when using Affirm compared to other payment methods
Affirm is built on a foundation of trust: it never charges late fees for missed payments and doesn’t make money from customer mistakes or deferred interest, tactics used by incumbents.
Conceptually, the existing credit card supply chain involves three main players: the bank that offers the consumer his credit card (for example, Chase), the credit card network (Visa or Mastercard), and the bank that offers merchant services to the merchant (it’s more complicated than this, but for our purposes, we’ll just need to focus on these players).
Visa and Mastercard make a tiny amount from every transaction, typically 10 to 15 basis points (0.10% to 0.15%) on a transaction that costs the merchant 2-3%. Most of the remainder accrues to the bank (Chase), and on top of that, Chase makes money from late fees and revolving credit card interest, including the reviled deferred interest, which accelerates interest on the entire purchase price upon failure to make any repayment along the way.
Affirm took a firm stance against all of this; the only way Affirm makes money is if the customer pays back on time. They’ll remind you to repay, but they’ll never charge a late fee.
This principled stance has won over partners who value cultivating long-term, positive-sum relationships with their customers, such as Peloton (which was 20 percent of Affirm’s revenue last year), Shopify, and many others.
Most repayments on Affirm loans are made through direct deposit from the customer’s bank, bypassing credit cards altogether. Incredibly, a full 30 percent of Affirm’s transactions occur on Affirm’s own mobile app. That is, customers are finding deals on Affirm’s app without ever visiting a merchant. With over 7 million active consumers (nearly double the amount a year ago), Affirm is building a destination for deal discovery.
On top of this destination, Affirm now offers checking accounts and crypto trading. Engagement is the name of the game for digital wallets; as it grows, Affirm aims to offer ever more tailored deals to consumers.
If you squint, you can see the birth of an entirely new payment rail. Similar to Netflix, which pioneered over the top (OTT) streaming and eventually bypassed the existing value chain of movie studios and cable distributors, Affirm is bypassing the existing credit card value chain, and capturing the profits for itself.
Along the way, Affirm is building a valuable two-sided network and offering more services on both sides. We’ve discussed the wallet above. For merchants and consumers, Affirm acquired Returnly, a SaaS product that allows customers to return an item, receive credit immediately, and place a new order with the merchant even before sending back the item. This solves a big problem for both sides of the network and increases customer and merchant satisfaction.
For merchants, Affirm will also start offering low-cost cash advances (similar to Shopify and Square Capital programs). Like Square, Affirm looks like its aiming to eventually replace incumbent banks altogether.
Unlike the existing value chain, which is mostly message-passing back and forth, Affirm has SKU-level data. For the first time, it’ll allow merchants and manufacturers to offer product-level deals, and Affirm will have much richer data for credit underwriting.
This, to me, is the money slide on Affirm:
Notice how Affirm isn’t just going after the $115 billion in credit card processing fees paid by merchants, the $135 billion in late, interest, and overdraft fees paid by consumers, or the estimated $1 trillion in customer acquisition spend paid by merchants: it’s going after all of it.
Currently, the penetration of BNPL in American commerce is estimated to be quite low, around 1-2 percent of total ecommerce, while it’s much higher in other countries (25 percent in Sweden, 18 percent in Germany and 8 percent in Australia), but Levchin thinks Affirm can grow to address nearly every consumer spend use case.
To help make this happen, Affirm will soon roll out its Debit+ card. This is a card consumers can use for all their purchases. By default, it’ll act like a regular debit card, making payments from a linked bank account. But consumers can also open up the companion app and tap on any transaction and turn it into a BNPL transaction after the fact, and have the same Affirm promise: an upfront payment schedule with no gimmicks or late fees.
According to the US Census Bureau, retail sales in the US are around $6 trillion, with ecommerce at about 14 percent of that, or $840 billion. The line between what’s ecommerce or not is increasingly blurry, but let’s just run with these figures—the exact split is unimportant.
Shopify is doing about $150 billion of GMV; last fiscal year, Affirm did $8.3 billion of GMV. This gives us a sense of the size of the prize (and remember, Affirm is targeting adjacent revenue pools as well).
Affirm has guided to 20-30 percent operating margins at scale. We believe Affirm can grow both in the US and internationally to become a meaningful player in the $10 trillion global ecommerce industry.
Our financial model assumes Affirm can capture just 4 percent of global ecommerce in the next 12 years. Including share count dilution, from our cost basis of $99, we estimate an IRR of at least 19 percent over the long term.
We now own what could be the three most valuable BNPL providers in the world besides Klarna, the privately-held Swedish behemoth. That’s because we are investors in PayPal, which recently began offering BNPL, and of course, Square, our second largest position, recently announced a deal to acquire BNPL provider AfterPay for about $30 billion in stock.
In the proxy for the deal, Square had to show its internal projections for AfterPay, alongside Wall Street analyst consensus estimates for AfterPay.
Looking out a decade, what’s notable about Square’s own projections is that Square believes AfterPay will be about 2.5x more valuable than Street estimates, as measure by gross profits.
Secondly, by layering these AfterPay estimates onto our internal model for Square, including the share count dilution from the deal, we arrive at a long-term IRR that is three percentage points higher than standalone Square.
I believe this is a necessary, and transformational, deal for Square. If it’s true that BNPL has the potential to develop an entire new payment network, then Square must be a part of it, or risk disintermediation, since both its merchant services and Cash App currently rely on existing credit card and ACH rails.