Affirm — Could it make a difference?
Mission: Building honest financial products that improve lives.
Vision: To be as ubiquitous, secure, and convenient as legacy networks, yet far more transparent, honest, and both consumer and merchant-centric.
One of the biggest religion that functions our economy is capitalism and its core enabler is the financial system. For decades, the financial system has been dominated by a small group of large financial institutions and despite the existential threat felt by other industries (such as retail for instance) from the technological boom, the financial industry has largely (and comparatively) remained unscathed. There is, however, an increasing amount of force from the ‘fintech’ industry being built that may permanently restructure the financial system and potentially threat the financial institutions as well. One of such fintech company is Affirm, founded by Max Levchin, with a mission to build an honest financial product and attempting to widen its force through a niche in how consumers pay for goods and services online.
How is Affirm different from other fintech companies and could the company be the one that will move the needle in the fintech industry? Could it one day replace the conventional financial institutions such as Citi and Visa? They have made some impressive moves so far but how should they strategise from here on and what are some risks they should address?
The Company
Affirm chose the payments sector as their niche. I think this is partly because of its founder’s experience of successfully building PayPal but more importantly because it recognises the sector’s opaqueness as a source of inefficiency that limits the financial industry from realising a higher potential (this is reflected well in their Mission and Vision statement). Their solution is to provide a more efficient and transparent payment option to consumers which will benefit merchants and build this positive feedback cycle via a more robust loan underwriting system built through AI and machine learning.
Essentially, this is similar to the P2P lending companies (such as Lending Club) in the sense that the core value arises from a more efficient underwriting process that allows the industry’s pie to grow bigger by capturing sections that were previously ignored. This ‘growth by capturing’ is depicted in Figure 1 which I used in my other essay ‘Centralised VS decentralised validation’ to argue how the centralised credit validation system seems anachronic in this modern world (naturally, this is why I took an interest in Affirm ) and how a tech-enabled decentralised credit validation should be the norm in the future.
The company achieves this by offering three main pillars:
- Point-of-sale which allows consumers to pay a fixed instalment amounts without deferred interest, hidden fees, or penalties (commonly dubbed as ‘buy now pay later or BNPL’).
- Merchant solution by offering an enhanced demand generation, reduction in customer acquisition cost (CAC), reduction in checkout page abandon rate, and an increase in average order volume (AOV).
- Affirm app (or Affirm’s virtual cards) that offer consumers an easier way to experience Affirm’s benefit.
The first pillar is the main value proposition to its consumer target market because it allows them to pay for its shopping items in the desired instalment period, boosting his/her purchasing power at the time of checkout which would reduce checkout page abandonment and increase AOV. In turn, the benefits form as the main value of the company’s second pillar and justify Affirm’s fee charged on the merchants (which is individually negotiated) for installing its API on the merchant’s checkout page.
This is not a new concept and in other parts of the world such as EMEA and parts of Asia, credit card companies often use such interest-free instalments as a seasonal promotion (for example in Korea, 11 credit card companies are offering an interest-free instalment option for January in 2021. (Link here) and such promotions happen throughout the year). However, it is worthwhile to think whether such a payment option is fully in line with the company’s mission. For instance, could this encourage people to shop more than they could afford? Isn’t it creating another ‘class-division based’ on credit rating because the 0% APR is partly afforded by interests income earned from those of lower credit score? While I have not found any evidence that the company is operating in such a manner, this is something that critics could bring up to drag down the company’s brand value.
The final pillar of virtual cards account only about 6% of Affirm’s total revenue (its share of revenue was only 3% in 2019) but in my opinion, this strategic presence in the consumer sector indicates the company’s long-term intention to become the dominant payments platform in the entire online transaction market and is using BNPL merely as an entry point.
The online payment network is a marketplace with supply-side (the merchants) and demand-side (the consumers). Such a marketplace is hard to scale but it has been an accepted norm (refer to this Marketplace Guide from VersionOne Ventures) to build the supply-side first and this is a strategy that a potential competitor such as Stripe is executing on (refer to my previous writing ‘Stripe making conventional banks replaceable’). To prepare for such future (and imminent) competition on ‘who will be the dominant payments platform?’, an effective strategy for Affirm is to either 1) build a stronger and denser supply-side, 2) beat the competition in owning the demand-side, or 3) build both sides and be the first to create an effective network value chain.
In essence, referring to the three figures below, while competitors such as Stripe is moving from the current status to Sept 1 and Step 2, Affirm can go straight from the current status to Step 2.
Competitive Advantage
There are two competitive advantages that Affirm is currently pursuing and which could act as a strong moat once they reach a certain threshold.
Network Effect
As with most marketplace, it is effectively a race of ‘who can build a positive feedback network cycle faster’. It is hard to build and the threshold to reach is high but once the threshold is reached, benefits immense and will act as a strong moat.
As mentioned above, Affirm is executing a strategic but very hard and expansive strategy of targeting both the supply-side and the demand-side at the same time. So far, the metrics show that Affirm has been successful in the execution. On the supply-side, merchant retention rate (in dollar terms) is above 100% and the number of the listed merchant is growing (currently at 6,500+). Also, its revenue from merchant network fee almost doubled in 2020, growing from $132M in 2019 to $256M in 2020. On the demand-side, it is maintaining a high NPS score of 78 while its GMV grew from $2.6B to $4.6B, a 77% increase.
It has, however, a room for an improvement in both sides and I will bring up the demand-side’s shortfall first in this section and bring up the supply-side’s shortfall in the Risk section. The average number of transaction per customer is still low at 2 times a year and only about 22% of consumers repeat purchase from the same merchant within a year. Given the fact that an average Gen-Z and Millennials shop online 2+ times a month, and a need for Affirm to build a stickier relationship to create a denser network loop, these metrics definitely should be one of the core focus to improve on. Also, the average value per transaction decreased from $640.6 in 2019 to $610.3 in 2020, which could impact overall revenue and GMV and a one-year LTV per customer was $62.2 in 2019 but it fell by 8.4% to $62.2 in 2020. Since the CAC per customer is roughly around $5.3 the margin is still healthy (thanks to the organic growth via word of mouth) but the number could definitely trend higher in Affirm wants to keep its edge in building the network loop before others.
*These numbers are calculated based on the financial numbers provided in the company’s recent S-1 form and hence are my estimation. It may be different from the actual number.
Proprietary risk model enabled by technology
As mentioned earlier, a large part of Affirm’s value arises from its robust underwriting model which can be summarised by the graph below. From this graph, it is clear that their risk model is identifying consumer credit risk more efficiently than that of FICO standard. The company’s weighted-average quarterly delinquency rate is around 1.1% (until 30 September 2020) while an average delinquency rate for the credit card industry is 1.81% (according to creditcard.com). The difference of 71bps may look small but when it is magnified by a growing GMV that is in billions of dollars, it makes a meaningful difference to the bottom line and ultimately to the robustness of a business model (especially when compared to the conventional financial institutions).
In addition, the difference allows the company to offer more competitive pricing to consumers which leads to an increased value to the merchants and ultimately completes a positive loop in the network.
Reflecting such importance of its risk model, about 47% of the company’s workforce is in the technical team and the technology is also protected by 1 listed patent and 6 pending patent applications (while reliance on Intellectual Property is never a good idea in a consumer tech company, it could nevertheless act as a deterrent).
Business model & Risks
Revenue
As can be seen from the table below, the company’s revenue is growing strongly across all sources but two things are worth mentioning: concentration and revenue capturing. Both of them could be of less relevance had the company stayed in the private market but since it opted to go public, a radical fluctuation in revenue will hurt shareholder’s equity value which ultimately could negatively impact its brand image and its negotiating power in the consumer market.
Concentration
For the fiscal year ended June 30, 2020, Peloton, Affirm’s top merchant partner, accounted for 28% of revenue and that increased to 30% by Sep 30, 2020. During the same periods, the revenue represented by the top 10 merchants, including Peloton, accounted for 35% and 37%, respectively.
This signals two things 1) Affirm’s reliance in Peloton is unusually high hence it poses as a risk and needs to be diversified and 2) revenue from the remaining merchants are flat or reducing indicating a need to refine and re-evaluate its value proposition, pricing model, or target industry.
The company currently qualify as an ‘emerging growth’ company (limited disclosure requirements) and hence shareholders and prospective investors could be more forgiving in such unusual concentration but it may change when the registration status changes to ‘large accelerated filer’ (the company would need to disclose more information than today, which may raise more questions).
Revenue recognition
Following the accounting standard, Affirm recognises merchant network revenue when the merchant receives payment AND confirms a transaction (internally called ‘merchant capture date’). Since the merchant capture date is out of Affirm’s control, this is a source of potential revenue fluctuation as well and could potentially paint an inaccurate picture in its revenue and GMV timing. For instance, Peloton has already made such change in December 2020 and hence Affirm’s quarterly revenue and GMV results will be negatively impacted in the quarters of 2021.
Therefore, it may be worthwhile to standardise the merchant/partnership agreement and reduce such adverse volatility in revenue and GMV.
Funding capability
Affirm originates its loan (or the advance payment it makes to the merchants on behalf of consumers) from the third party. While it also uses its own equity too (currently about 19% of its equity, higher than a regulatory requirement of 8%), a large portion of such origination relies on Cross River Bank, a bank with $10B assets (it sounds a lot but the largest banks in the county have assets in trillions and hence it is considered a small bank) from Fort Lee, New Jersey. Cross River Banks is an FDIC insured lender with BBB credit rating, lowest tier in an investment grade, for its bond issues. Affirm has a 3-year contract with the bank with automatic 1-year rollover and 90 days termination notice period. Termination in this relationship or a negative event in the bank could seriously damage Affirm’s ability to fund the cash for loans which will break the network loop, depicted in the Competitive Advantage section. This is an indication of two things: 1) Affirm needs a strong partnership with a conventional financial institution and its proprietary risk model (one of their core competitive advantage) is yet to be fully adopted and recognised by the mainstream financial industry and 2) Affirm identifies large banks (such as Citi) as their competitors which means it is unlikely that they will be able to secure a fund origination relationship with a larger and a stronger financial institution.
Luckily, Affirm’s fund velocity is faster than normal due to its short duration (less than 1 year) which provides a multiplicative effect to its funds (i.e. the same money can be re-used multiple times a year). As the business grows, more data and refinement to its risk model should enhance its efficiency, further increasing the velocity of their funds. Ultimately this will alleviate some of the concentration issues in the funding source.
Future opportunities and growth strategy
Grow tangential to current network chain
For marketplaces, it is best to grow tangential to its current networks rather than starting an entirely new network chain in a new market. For instance, the rise of E-Commerce has brought western market closer to the Asian consumers and size of such a market is growing fast. In Korea, GMV of a direct purchase from abroad (called jik-goo) was US$2.5B for the first three quarters of 2020 which is an 8% growth YoY. The 4th quarter of 2020 alone is expected to bring in about US$800M in GMV, making the total 2020 GMV of just under US$3.5B.
The largest E-Commerce companies in Korea such as G-Market and eBay Korea are responding to this trend by creating a new team dedicated to the jik-goo demand. Directing such demand to Affirm’s current merchants in North America would not only strengthen the network chain but also prove to be a valid entry point in its global expansion.
Diversify funding source
A continuous improvement to its risk model and an effort to have the model accepted in the broader financial institutes across the country should be a medium/long term solution to the problem. On the other hand, looking into a variety of other funding sources such as a loan from business-development companies (or BDCs such as Owl Rock) who lend to SMEs or other private lenders such as endowment funds could provide a short term alleviation to the problem.
Another solution is to continue fine-tuning the risk model and increase the velocity of the loan (or reduce its duration) making the fund work more diligently.
Other key opportunities and strategies include:
- Increase consumer purchase frequency and repeat purchase
- Diversify merchant indsutry to smooth revenue
- Omni-channel payments in commerce while increasing touchpoints with conventional commerce through innovative financial products
- Enhance value to merchants such as better consumer data, more consumer traffic…etc
