Rethinking Concentration Risk in B2B Fintech Infrastructure

Hey friends 👋,

A bit of a change today from the MO, but today’s piece is a deep-dive into a meta topic with a good friend of mine.

Hope you enjoy it! Check out Chia’s newsletter here.

As more fintech infrastructure companies IPO and grow in the public markets with significant customer concentration, Chia (Saison Capital) and I wanted to dive into the implications for rethinking concentration risk in the unique context of b2b fintech infrastructure companies.

We look into:

Conventional logic says that revenue concentration risk is something to be worried about. This makes sense - after all, when a customer represents a significant portion of your revenue, they hold a significant bargaining position and may drive your margins down in the long run. They may also leave suddenly, creating volatile revenue situations.

Revenue concentration risk is such a standardly accepted risk that it is one of the mandatory flags that must be discussed in a company’s SEC filing.

It has therefore been fascinating to see multiple B2B Fintech infrastructure plays go public with significant revenue concentration risk.

How do you know if your customer concentration is high or low? A rule of thumb holds that if any single customer accounts for 10% or more of your revenue or if your largest five customers account for 25% or more of revenue, you have high customer concentration.

We have seen tech companies with concentration risk biting them. Twilio losing Uber, which was 12% of their revenue source at the time, and suffering a 30% share drop at the time, is one of them. As Yann Ranchere poses the question - is customer concentration an overstated risk, or a source of alpha to be lauded?

We see B2B concentration risk across different segments of Fintech. As we can observe, concentration risk is applicable across multiple verticals of fintech.

Revenue Concentration should be seen as a positive feature, and not a bug, of B2B Fintech Infrastructure.

The main bear factor for b2b infrastructure plays is the churn-off problem. This basically describes the situation where smaller startups use your infrastructure solution as an interim to building their own tools, or represents a potential for volatile revenues.

The churn-off problem is uniquely mitigated in Fintech Infrastructure. Fintech infrastructure is far less turnkey than other types of infrastructure provider (e.g. Communication (Twilio), CRM (Zendesk), etc). This is especially the case for highly regulated industries like Fintech (as opposed to for e.g. a standard CRM tool). The requirements in fintech infrastructure encompass not only teams with a capacity for not just regulatory expertise, but also a capacity for specifically proprietary regulatory expertise (i.e. a regulatory moat through the creation of a new financial process/product).

This is especially true in fintech where the customers of fintech infrastructure are less familiar of the regulatory complexities of the financial product (for example in embedded finance plays), or for new financial services (Kalshi’s development of event-contracts -betting on events happening -are a good example of this).

As such, fintech infrastructure is far less commodifiable, given the degree to which it abstracts the complexity and cost of regulatory compliance. In tandem with the degree of customization for customers, fintech infra is generally less likely to be a target for in-housing.

Quantitatively, this has meant that b2b fintech infra are more likely to see higher net dollar retention (NDR).

Above are the NDR numbers for the top public SaaS companies, with an average of 120%.

Marqeta’s NDR was over 200% for FY20 and FY19. Adyen has a NDR of 115-120%. Affirm has a >100% NDR.

In the SMB SaaS sector, Shopify has a 100% NDR, while its comparable SMB fintech infrastructure play like Toast has a 110% NDR. We note that a lot of our points on fintech infrastructure relate mostly to non-SMB fintech infrastructure players. Notably, although no numbers are publicly available, Toast does not have a single customer above 10% of their revenues.

(For the purposes of this hypothesis, we refer to customer revenue concentration only, and not country or product concentration)

Marqeta’s TPV growth by cohort

Is it long-term sustainable? Probably not - as companies get larger, they want to prevent single points of failure, which will include their financial services or fintech infrastructure. Further, it is true that margin compression is likely to happen in the long-run. For example, Marqeta's take rate fell from about 0.7% in 2019 to 0.5% in 2020, and has dropped further since then. However, if you see whale clients as an investment in capex/capacity building to allow a platform to scale quickly from an early-stage company to one capable of large volumes of transactions, then leaning into revenue concentration as a private company is a legitimate strategy. Having built up capex/capacity to serve other larger enterprise clients positions the infrastructure platform to be able to win over other whale-type clients (which in Marqeta’s specific case, we can see in the multiple other whales they onboarded). Frank Rotman of QED pointed out that being able to “practice” with an “at-scale” client is extremely valuable and paves the way for additional contracts. Sheel Mohnot of Better Tomorrow Ventures makes a point, which I believe is especially important for infrastructure plays, that getting blue-chip logos is an obviously great way to shorten the sales cycle at other ‘whale’ clients.

The underlying risk that an investor assumes when assuming such concentration risk is that the growth rate of the ‘whale’ client is an anomaly not reflective of broader industry growth rates. However, if the investor is already willing to make a broader macro bet on a specific industry (say card-issuing), then they are already underpinned by a rising tide that is only supported by the added growth from a concentrated client, to support infrastructure/capacity building. A short-term focus on concentration risk is therefore misleading, and sub-optimal from the lens of an investor.

To illustrate Marqeta’s ability to win over existing ’whale’ clients over time, and their ability to capture upcoming whales, we see that Marqeta has been able to capture both large enterprise clients as well as selective small clients that will become whales:

At point of onboarding:

The key here is to understand that even if fintech infrastructure plays exhibit high churn rates among a long-tail of startups, that they possess an ability to remain sticky among the core group of fast-growing companies. If they have a double-churn issue (i.e. churn on the long-tail of customers, and churn after a certain scale with their fastest-growing companies), then it is more unlikely to see a path to success.

Furthermore, one of the best arguments we can make that revenue concentration does not matter for fintech infrastructure startups is actually the very fact that so many fintech infrastructure companies IPO at large market capitalizations, with significant revenue concentration. This indicates that you can build a successful IPO-able startup over multi-year periods, with significant revenue concentration, in excess of at least 30%.

Lastly, capturing Alpha is a far greater growth driver than capturing Beta. Given the sunk cost of setting up the initial regulatory and operational moat, being able to scale with that particular whale ‘alpha’ captures a faster growth rate than ‘beta’/industry benchmark growth. Yann Ranchere, partner at Anthemis, adds nuance to this argument by pointing out that ‘riding the whale’ is less about one whale and capturing the top 5/segment of potential/current industry players.

Usage-based pricing is why fintech companies are better able to extract value from their whale clients while sticking to them. Fintech infra generally has usage-based pricing, as opposed to general software that is priced on a license/seat basis. As Insight Partners’ James Woods points out, usage-based pricing allows for a natural rate of “land-and-expand” growth higher than any other form of pricing model, noting that SaaS companies across all verticals that can justify a usage-based pricing all tend to naturally have much higher net dollar retention.

A bear case is the emergence of infrastructure-as-a-service providers like Finix, that empowers companies to be their own pay-fac’s, allowing them to more easily escape reliance on a single infrastructure provider. I foresee the main short-term bottleneck here to be regulatory and compliance capabilities by companies in utilizing such infrastructure-as-a-service providers.

Given this specific understanding of fintech infrastructure where we’ve learnt to accept concentration risk in its growth journey, what are the boundary conditions for this principle across sub-segments of fintech infrastructure?

Fintech infrastructure that will do particularly well would:

Revenue concentration risk is also highly dependent on the industry. Frank Rotman of QED pointed out that revenue concentration was, for example, not an overwhelming risk in the case of Credit Karma. Although there was a concentration of lenders to Credit Karma, the quality of Credit Karma’s leads (and in our opinion, the non-fungibility of capital sources - i.e. each lender runs similar processes), meant that there were substitute lenders that could replace any deficiency in capital sources.

An example of a B2B Fintech Infrastructure company that suffered (at least temporarily) concentration risk included Green Dot. Green Dot’s debit card program with Walmart constituted 66% of revenue and net earnings.

Green Dot shares dropped in 2012 as, right after the IPO, Walmart announced it would also additionally partner with a competing debit card issuer.

In mid-2016, the Walmart contract was extended until 2020, and in end-2019, the contract was extended until 2026. At the same time, Green Dot’s overall revenue CAGR was relatively low at ~11% between 2012 and 2020. Arguably, the difficulty in growing revenues by winning over fast-growing clients for its BaaS products violated a core boundary condition for fintech infrastructure plays - to leverage existing resources to acquire new fast-growing clients.

As an aside, this seems to have turned around in the past few years with a new CEO installed by an activist fund. Green Dot announced partnerships with Quickbooks (Intuit) and Apple in 2019, with Wealthfront in 2020. In 2021, Amazon launched a debit card product with Green Dot.

We argue that not all fintech infrastructure is equal. Below, we highlight some common features that fintech infrastructure in certain verticals may have that make the revenue concentration argument less applicable:

In conclusion, concentration risk is a feature, not a bug, of best-in-class B2B Fintech infrastructure. They are:

Thanks to Frank Rotman, David Lilienfeld, Yann Ranchere, Andrew Starker, Natalie Luu, Sheel Mohnot for comments/input! Input does not constitute endorsement.