Reining in the BaaS Euphoria
TLDR
The initial conditions that made Banking as a Service (BaaS) a ripe category for asymmetric venture bets have shifted. BaaS startups now need to brace for unforgiving market pressures from across the digital banking ecosystem. Nevertheless, one subset of BaaS — Lending as a Service (LaaS) — appears to hold immense promise.
A Minimalist Definition
Banking as a Service (BaaS) platforms enable both fintech companies and other third parties (potentially non-financial brands) to create and distribute financial products to their customers by making it easier for them to integrate with the core systems of banks (See: Garcia et al. for more).
In the ordinary course, offering customers banking products such as debit cards, loans, or payments services requires a banking charter and comes with a host of onerous red tape. By leveraging a BaaS platform, however, a non-bank entity can seamlessly build financial products on top of the regulated infrastructure of a sponsor bank, thereby circumventing the need to acquire a license of its own.
Setting the Stage
Synapse — the original poster child for the modern BaaS revolution — was founded in 2014 when consumer fintech mania in the US was yet to reach any visible apex. By 2018 however, every fourth pitch to a fintech-focused VC was a slight variation on mobile-first digital banking that was meant to revolutionize the very idea of deposit accounts. Synapse, as one of the only BaaS providers on the scene with something even close to a developer-focused solution, grew rapidly as the ‘enabler’ in this newly proliferating ecosystem. Across offices on Sandhill Rd, two things were becoming unmistakably clear:
- Many newly forming fintech startups would comfortably get off the ground, but quickly incur growing pains from less-than-ideal competitive pressures leading to subpar unit economics.
- There was a wide-open opportunity for a pick-and-shovel play nonetheless.
And so we were off to the races with BaaS euphoria. By a preliminary count there are now close to twenty BaaS companies serving the US market including Bankable, BBVA Open Platform, Bond, Cambr, Cross River, Galileo, Green Dot, Hydrogen, Marqeta, Moov, Privacy, Q2, Rapyd, Rize, Stripe, Synapse, Treasury Prime, and Unit. Several regional banks are also making a go at it (See: Sterling National Bank and UMB Bank.), though it is less clear how many of these are doing so in partnership with independent BaaS providers, and how many are doing so directly by staffing up dedicated in-house teams. Outside of the US there are several more still: Setu in India; Flutterwave in the pan-African market; Solaris bank in Germany; and Nium and Railsbank are expanding rapidly in SE Asia.
While there are some non-trivial differences in terms of primary product functionalities, GTM strategies, and technical build, what they all enable (or are building towards enabling) is very similar:
- Card issuer-processing APIs
- Creation of deposit accounts
- Digital onboarding and compliance monitoring
- ACH payments, wires, automated transfers
- Loans
A Growth Ceiling?
Early BaaS platforms grew on the backs of thousands of newly formed fintech startups. But should we expect there to be as many new fintech startups this decade as there were in the 2010–2020 period? Shouldn’t we instead expect the massive rebundling efforts of the scaled fintech incumbents (Chime, MoneyLion, Robinhood, Square, SoFi, Wealthfront, etc.) to crowd out the formation of new startups in this arena? For instance, in the SMB banking space alone, we now have competition from Brex Cash, Kabbage, QuickBooks Cash, Shopify Balance, and Square Card. Not to mention companies like Autobooks, which are taking a white-label approach to helping banks revamp their SMB offerings from the inside out (more on this dynamic in the next section). There are still a handful of pure-play SMB-focused neobanking startups like Azlo, Novo, and Rho, doing good work. But going forward, we shouldn’t be surprised if the massive distributional advantages of incumbent banks and fintech giants prove to be sufficiently potent barriers to entry for new startups going after the same business banking market.
The challenge for BaaS platforms (created by the trend of rebundling) is that many of the rebundlers will not be reliant on the BaaS platforms for the heavy-lifting of building integrations with sponsor banks. For a fledgling fintech startup, working with a BaaS platform is often a no-brainer, since the alternative is to make a massive upfront investment in building direct integrations with the legacy core systems of banks. Stomaching this sunk cost can be prohibitively challenging without having enough proof points about the actual product-market fit or viability of the end product. But the sheer size of a Shopify or a Square flips that calculus. That upfront cost in the case of these giants can effectively be passed through over hundreds of thousands of end customers. What’s more, these giants understand the importance of owning an integral piece of technology like this, which could otherwise become a critical chokepoint; platform dependence can become an existential vulnerability, something that you pay a lot to avoid at this scale (See: Varo plans to migrate its 2 million accounts away from The Bancorp Bank now that it has been granted its own banking charter).
That said, there will inevitably be a few fintech rebundlers, which do decide to partner with the BaaS players to roll out their expanded suite of banking products. However, we will still have to wonder where the balance of power will lie — a point we will return to shortly.
No, a Momentous Vertical Shift in the Growth Function!
You might argue that regardless of what happens in the pure-play consumer fintech market, the more exciting growth opportunity for BaaS platforms actually lies in partnering with non-financial digital brands, and helping them create wholly new revenue lines through the distribution of digital banking products. By late 2019, the dominant narrative in fintech had indeed shifted, and all anyone could talk about was embedded financial services: Matt Harris declared fintech the fourth platform, and we got Angela Strange’s famous, “Every company will be a fintech company.”
But here too, the hand seems to have been overplayed: First, how much sense does it make for every brand to get into the embedded finance game (See: Uber’s Departure From Financial Services: A Speed Bump On The Path To Embedded Finance)? For Google, for Gusto, for Quickbooks, right on. But beyond the behemoths, it seldom makes for a coherent and unified product experience to just start throwing in a deposit account and a debit card in the mix. From the consumer’s perspective, there are only so many disparate financial relationships one can tolerate before fatiguing. There’s a reason why the traditional relationship banking model worked well for so long, and it’s not just that technological limitations allowed it to; there’s a reason Amazon’s one-stop shop model resonates still.
If we accept that there is a power law at play here and that the top non-financial brands will be responsible for driving the bulk of user adoption in embedded financial services, then the critical question becomes: where does the balance of power lie? When dealing with a massive account like Google or Uber, how much negotiating leverage does a small BaaS platform have to maintain healthy economics? Especially when the competition is over a dozen other BaaS players with deep-pocketed VCs to bankroll some good ol’ fashioned price-undercutting. The market dynamics of digital banking simply don’t afford the same luxuries enjoyed by, say, Stripe in payments: even if payment facilitation enablers (PayFac as a Service) startups like Finix eventually succeed at picking off some of Stripe’s mid-large customers by helping them turn their payment-processing from a cost center to a revenue generator, Stripe will largely remain unscathed. Partly because of its scale, but more importantly because Stripe is a true bet on the growth of internet commerce in the aggregate — every single up-and-coming startup, which needs to process a payment over the internet at some point is a potential Stripe customer, regardless of how small or large it is. The top of the funnel for Stripe is thus unfathomably large. It benefits from a similar sort of antifragility that has served Shopify and Facebook well (See: Ben Thompson on Shopify; Facebook). In the case of BaaS platforms however, digital brands with massive distribution make for ideal customers while smaller brands really aren’t logical participants in the embedded financial services trend at all. So we should expect the price competition to win the top 5% of digital brands is all the more intense.
This brings us finally to the last big consideration: are banks going to stand idly by? BaaS players who think they can build connectivity to multiple banks and enable non-banks to ‘shop’ around easily for a sponsor in some sort of marketplace fantasy are in for a rude awakening. Banks aren’t going to bend the knee and allow themselves to be relegated to the function of serving as dumb pipes. Owning the customer relationship is the absolute key to success in financial services and the banks know this all too well. We can be darn sure that they’re going to put up a fight. And there is after all an entire category of fintech startups whose success is pegged to helping banks modernize and fill the CX gaps, which have allowed fintechs and big tech to encroach on their turf in the first place. Companies like Amount, Blend, Mantl, and Personetics are helping banks rapidly crack the code to digitization such that they don’t lose their all-essential end-customer relationship to third parties. Further still, there’s yet another competitive pressure on the BaaS players — created this time by an entirely separate category of fintechs: the cloud-native core banking software providers. Companies like Mambu, Thought Machine, Neocova, and Finxact are helping regional and community banks bite the bullet early on legacy core transformations. These new cloud-native cores are designed from the ground up with third-party integrations in mind. While they won’t obviate the need for a BaaS provider altogether, they certainly lessen the load, and consequently also the value being delivered by a BaaS provider acting as a middleman between a bank and a third party (See: Finxact’s Marketplace).
A Greenfield
Despite what may have seemed like flat-out skepticism on the outlook for BaaS, I remain bullish on one important subset of the BaaS trend: Lending as a Service (LaaS).
Before delving further let’s first delineate three different lending models, which are easily conflated:
- There are buy-now-pay-later (BNPL) models like Affirm or GreenSky, in which a merchant will pay the provider a percentage of the total transaction (~3–7%) for increasing average order sizes and ensuring that the customer actually completes a purchase, similar to paying a merchant discount rate when a customer checks out using a credit card.
- There are software subscription securitization models like Pipe, in which the merchant (the SaaS business) will sell their receivables to the provider for some discount in order to get free cash upfront for reinvestment in the business.
- And then there are LaaS models like Wisetack, whereby a digital brand (typically a vertical SaaS business), empowered by the provider to underwrite and distribute loans to end customers, will actually generate revenue from facilitating the origination of a loan.
We are interested here only in the latter category. Traditionally, a vertical SaaS business like Mindbody or Housecall Pro would monetize borrower leads simply via pushing them out to fintech lenders like Fundbox or LendingClub and earning referral fees. The promise of embedded lending however, as A16Z argues, is that in place of just reselling leads, a SaaS business can hypothetically participate more in the underwriting and distribution of loans in order to improve its margins. Generally, a business like Mindbody would avoid involving itself in the lending game because of one simple question — can you generate software-like returns on capital from owning larger slices of risk on portfolios of loans? To which the answer is of course, no. But there is still a massive opportunity in embedded lending given the value created for all the relevant stakeholders:
- The SaaS business earns more margin for its role in providing relevant data for underwriting and for helping distribute loans to potential borrowers with high intent. It also benefits from stronger retention and engagement with the customer. Note: Payments companies are already knee-deep in the POS lending game (See: Lightspeed, Stripe To Offer SMB Loans; Square Capital) but while they are only able to leverage surface level transaction data, vertical SaaS businesses have access to proprietary datasets that are richer, more holistic, and ultimately more relevant in the underwriting of loans for their customers.
- The borrower gets a more seamless loan application, approval, and repayment experience.
- The capital markets are able to originate higher quality loans more cheaply.
The key ingredient is a LaaS platform — an independent, full-stack, white-label, scaled, loan facilitator, which could step in and intermediate the loan origination and servicing process on behalf of digital brands, their customers, and debt investors. Alex Hartz from SciFi has an excellent piece on embedded lending, in which he argues that lending is the ultimate scale business. He masterfully reasons that delivering a better lending product to the consumer is predicated upon having massive economies of scale in underwriting, distribution, access to capital markets, and process knowledge. In underwriting for instance, an independent embedded lender (what we have referred to here as LaaS) working with multiple brands could hypothetically see the same borrower across disparate transactions/contexts and use a greater number of data points on customer behavior than any one brand could easily gain access to.
More simply, the key insight behind embedded lending is that a brand sitting closest to the customer ought to be able to arbitrage its competitive advantages in distribution and in underwriting to deliver a loan for less. LaaS is an extension of the same idea; it merely abstracts away the various functions of a lending business and optimizes them at scale such that the relevant stakeholders in the loan lifecycle can have a piece of the pie without having to do much incremental work.
For any Founder or VC looking to get in on this action, I reserve the highest respect and admiration. The flywheel in a model like this doesn’t even begin to budge until you achieve a critical mass of buy-in from all the relevant stakeholder groups: brands, borrowers, debt investors, ratings agencies, credit bureaus, alternative data providers, loan servicers, data privacy vendors, regulators, data scientists, developers — the list goes on. This path couldn’t be any farther from the lean startup philosophy; and that’s precisely why the rewards, for those who successfully tread it, will be incomparably large.
