A chat with Charlie Deutsch, GM of Financial Services at TrueAccord

Historical evolution of lending products & capital markets, UX vs product innovation, disruption of FICO, valuations & strategy of lending & banking companies, and more

November 09, 2020

Sar : You began your career analyzing and valuing structured credit and later went on to work at OnDeck Capital, Acorns, Wave, and now at TrueAccord. Can you walk us through your career and how what you’ve learned in these experiences has led you to where you are today?

Charlie : Absolutely! My goal after graduating was to get a job - any job - in financial services. I didn’t go to a top school, so it was an uphill battle for me. I ultimately landed at a lowly insurance company underwriting these esoteric AAA CDO bonds that were *supposedly* impervious to losses. Little did I know we’d underwritten many of the bonds that ultimately failed in the financial crisis.

This was a formative experience in a few ways; the first is I came to realize FICO was not the end-all, be-all, in terms of predicting risk - after this I became obsessed with alternative credit scoring (more on that later). The other learning was personal; the people who caused financial destruction made a killing, while a lot of innocent bystanders lost their homes. I felt like I was on the wrong side - I wanted to be a value creator, not a value extractor.

This is what moved me toward trying to help the little guy. I get excited by serving segments that are avoided by traditional financial services. I joined OnDeck early and built their direct origination muscle and partnerships business. I’ve since helped consumers get financial services at Acorns and micro-enterprises get access to 1-click invoice financing with Wave.

Now I’m leading an initiative at TrueAccord to build products for our consumers. It’s a unique opportunity to help a challenged segment and build a great business at the same time. We have a unique position in the market where we’re paid to help these consumers which means room for real financial product innovation, not just UX innovation. Sometimes, these things are confused.

Sar : Can you expand on what you mean by conflating product innovation and UX innovation? Any historical examples come to mind?

Charlie : Sure. The first phase of financial innovation is usually about making a bad experience more user friendly - say an instant loan or digital investment account. While it may represent a better experience, it often requires novel distribution - let’s say Facebook. At first, distribution is inexpensive, but as popularity grows competition drives up acquisition costs. Ultimately, the company must either shrink margins or these pass costs to the customer. Since venture-backed companies often have negative margins to begin with - and raise money on the hope of margin expansion - it’s typically the latter outcome.

There are two ways to solve this challenge. The first is to introduce a novel business model - like generating revenue off your customer base, but not your customers. Payment for order flow (Robinhood) or unregulated debit (Chime) are two examples of this. The second is structural - if you can eliminate your acquisition costs, it gives you a competitive advantage to offer the same product at a lower price because the customer has already bought something from you (Square Capital) or what we’re working on at TrueAccord fits into this model.

Sar : Tell us a bit about what you’re building at TrueAccord? You guys are working on very unsexy but important problems. What do we have to know to appreciate what you are doing matters?

Charlie : TrueAccord is a great opportunity to help a segment that’s been left behind. Most new credit products fall into the category of “new to credit” or “credit building” for consumers looking to open their first line and establish a history. The consumers we’re serving come through our core business - this means they’ve used credit before, and it hasn’t gone well for them.

While this is a challenging customer to serve, we have a few advantages. The first is consumers love working with us - our CEO likes to say our reviews are “shockingly good” for the business we’re in and it’s true. The second dynamic is we have what amounts to a negative acquisition cost - we get paid to help these customers. These factors allow us to design products that most closely serve the needs of our customers, knowing we don’t start in the red.

Sar : How has the landscape of lending products from non banking companies evolved over the past decade?

Charlie : It’s evolved a lot. In 2008 banks were in balance sheet repair mode - consumers and SMEs were starved for capital, and a lot of mortgage and ABS buyers were in search of yield. The resulting vacuum yielded the “tech-enabled lenders” like LendingClub, Prosper, OnDeck and Kabbage. The model was simple - rapid digital fulfillment for a premium cost product. It flourished for a few years but in 2015-ish the market realized capital is a commodity and there are no barriers to entry. The subsequent venture goldrush drove acquisition costs through the rough.

While these lenders were slugging it out, payment platforms like Square and PayPal realized their customers wanted capital, and the embedded nature of their products solved the retention challenge. They also moved the money, which they used as an additional risk management tool. Since lending wasn’t their core business and they didn’t have acquisition costs, they didn’t need to lend to everybody - only their best customers - and it strengthened these relationships. The model has proliferated and every relevant payment provider offers it today.

On the consumer side, Buy Now Pay Later and Earned Wage Access have been the areas of explosive growth. Affirm, Klarna and Afterpay realized merchants would subsidize rates in exchange for increased conversion and basket size, making the product a viable alternative to traditional credit cards. As a byproduct, these companies also solved the distribution model that plagued the original tech-enabled lenders in a defensible, embedded way. EWA companies like Even and EarnIn give employees instant access to their income, but they are ultimately taking risk to corporates, so it’s not rocket science from an underwriting perspective.

Now businesses like Clearbanc and Pipe are taking a novel approach to what’s old-fashioned factoring. The innovation is about directing funds, using scalable electronic data, and being further up the waterfall with a shorter feedback loop. While they have a tech flavor, these companies will ultimately be viewed as specialty finance lenders in my opinion, which makes them less desirable from a valuation perspective. Don’t get me wrong, these can be highly profitable businesses, they just won’t ever trade at 100x revenues!

Sar : That was a great explainer of the multiple waves of lending products we have seen over the past decade. You mentioned there was a reckoning of sorts in 2015 when the tech zeitgeist collectively realized the commodified nature, rising acquisition costs, and lower barriers to entry in the market of pure lending plays. LendingClub and OnDeck went public in 2014. In a lot of ways, the idea of embedded lending has been popularized by people who have seen the distribution challenges of pure lending startups play out in 2010-2015. Do you believe we are repeating the market cycle of lending in the challenger banks market over the past 2 years? Especially right now in a low interest environment where both high savings yield consumer value proposition and the core interchange based biz model are challenged. There are a handful of strong growing players like Chime, Varo, Current in the US.

Charlie : It’s at least a reasonable question to ask. Neo-banks exist due to an exclusion in Dodd-Frank called “Unregulated Interchange” created to protect small businesses from large banks who were charging credit interchange for debit transactions (the former has chargeback risk, the latter does not.) The loop-hole allows sub $10BN banks to charge ~10x on debit. But once deposits exceed $10BN, you’re technically no longer exempt. To further protect this advantage, many neo-banks use distributed deposits. What that means is the $5,000 you have on deposit is actually spread across many small banks, allowing neo-bank to claim a much smaller deposit base than if they were concentrated on their own balance sheet (or really their partner bank’s).

It will be interesting to see how public markets view these models. I’d expect them to ask two questions; the first is how durable is the exemption because if the loop-hole is closed, they will lose 80%+ of their revenue. You have to wonder how the banking lobby will feel about it - and it won’t just be the big banks. The second question is whether investors will be comfortable with more and more of the neo-banks revenue mix coming from lending. The purpose of having a long-term banking relationship is to access other (credit) products so customers will eventually want to borrow. Chime seems likely to go first so we’ll learn from their experience, but their CEO is definitely working to establish the narrative they are a payments company, not a bank.

Sar : You mentioned startups taking a differentiated approach to old fashioned factoring earlier. There’s always a visceral reaction from management teams at next gen lenders for ecommerce brands (Clearbanc et al) or SaaS companies (Pipe, Capchase) to being branded as factoring lenders. How much do you believe that flows from a regulatory angle?

Charlie : This is mostly about crafting a narrative they are a tech company, not a specialty finance company, deserving of a tech multiple. The irony is from a regulatory perspective, these companies likely work hard to be classified as a factoring business, not a lender, as it’s quite beneficial in easing regulatory and compliance burdens vs lending. Unless you classify yourself, a regulator will do it for you, so you can’t ignore the issue.

Sar : How did online lending companies like OnDeck go about building their underwriting models and raising money in their first 2-3 years? My understanding is companies used their equity funds to prove out their models and then try to find capital providers for raising debt funds to scale up.

Charlie : Yes - for pure lenders like OnDeck, the model required an on balance sheet “test and learn” approach. Generate 1-2 performance vintages to show a debt provider you could execute the lending mechanic with stable performance, and you could get an advance rate structure.

For established companies looking to add lending as a complimentary business, even though most can raise debt up front, it’s advantageous for them to fund the first few vintages on balance sheet, as it gives them the flexibility to try whatever it is they want from a product perspective.

This is how we’re approaching new product development at TrueAccord - while we have access to debt, we have some novel ideas for how to best design for our customers’ needs, and we want to iterate quickly without any unnecessary constraints created by third party structures.

Sar : How has the landscape of fundraising options for lending products evolved from 10 years ago?

Charlie : The capital side of the market evolved quite a bit. After the financial crisis, there were a slew of mortgage, ABS and whole loan funds who were looking for yield. They were open to financing nascent products as long as they were priced to multiples of expected defaults and they offered low advance rates (the term for how much equity the originator needed to contribute per loan.)

Some of these forward thinking players also got creative and negotiated equity kickers in the form of warrants in the case these tech-enabled lenders became valuable.

In the mid 2010s these structures became more standard and there were funds who became focused players in this space, like Victory Park, but they weren’t typically the first capital in - you had to prove a model on your own balance sheet using equity, first. Now there are funds specifically designed for this. While I’ve not met them, I know CoVenture looks to invest both equity and debt into companies looking to unlock new asset classes. It’s an interesting model, but I think we need to be cautious about thinking every business should be a lending business.

Sar : There’s a running joke in fintech circles that everyone wants to be a payments company. You mentioned people should be careful about wanting every business to be lending business. Can you expand on why you think that?

Charlie : Sure. Software businesses and lending businesses are antithetical. Software businesses are valued based on growth and negative retention, while lending businesses are valued based on profitability. You must also grow a lending business in a controlled manner, which is somewhat opposite to the ideals of the high-growth venture world.

The outcomes of the first cohort of tech-enabled lenders demonstrate this risk. Ultimately, public markets viewed these companies as specialty finance businesses, not tech, and it had serious consequences for valuations. Overnight they had to change the way they did everything - no more R&D, no more investing in growth - they had to become profitable overnight. It slammed the brakes on these companies while other companies picked up market share.

This is a key learning for me - and I tell anyone who is thinking about lending - that it makes a phenomenal accelerent or complementary business, but you can’t lend to everyone. If you do, it makes you a lender, not a tech company. Square has done a great job of finding a balance here - they lend to their best 10-15% of their customers, and these customers pay this off by staying and growing their payments volume.

Sar : I always say lending is a very difficult primary business but a great secondary business if its growth is not directly linked to the growth of a company. I definitely agree with you on that. I think there is a wide gap between when software companies decide to spin up a lending business and when they actually are ready to scale it up. The first step is almost always doing a pilot with your own capital to test and learn. But, the next step is where it gets all murky. There’s cap raise, legal, logistical challenges.

Charlie : Assuming it makes sense to lend to your customers, I’d bifurcate the approach based on whether or not the company is already in the business of moving money. At Wave, we were a PayFac, so it was easy to simply “turn things on” from a lending perspective. We had all the functionality around funds flows, reconciliation, etc. that is necessary to spin up lending activities.

For non-payments companies, there are a growing number of lending-as-a-service offerings that are turnkey and prevent you from having to build a lot of complex software from scratch. At the moment, I’m looking at a few of these providers and I’m astonished how far things have since when I joined OnDeck - we had very limited options and had to start with a clunky core processor. Now there are API-based servicing solutions which you can spin up in weeks, which prevent you from having to make major investment in infrastructure to be a compliant lender.

Sar : You mentioned disrupting FICO - why haven’t we seen this happen?

Charlie : Great question - the direct reason is the creation of the Basel Accords. Unless you’ve spent a lot of time on bank liquidity management, you probably haven’t heard of them, but they govern how much capital a bank is required to hold for each asset. These risk-weightings are determined by rating agencies, and ratings are heavily influenced by commonly accepted metrics like FICO. In my opinion, this is what forces FICO usage downstream.

More importantly, it’s now my perspective that the type of Fintech organizations who would have the influence to drive this change don’t actually care to do so. They’ve realized what makes an embedded lending model shine is an orthogonal data set - unrelated to credit - that sheds more a more nuanced light on risk than the common credit score.

In the late 90s, merchant cash advance companies figured out processing data and transaction elasticity were much more important than a personal credit. FICO was useful for exclusions, but not for inclusions. Now companies like Clearbanc can advance funds, dictate your spend, and tap into your revenue stream before you’ve even fulfilled your order.

Sar : What skill sets should software companies be hiring for to go from an idea to implementation when thinking about using lending as a growth driver?

Charlie : Taking a traditional risk management approach to these sorts of products does not make sense. For the type of products we’re discussing, lending is much more of a data and design business, than a “credit risk” business. You create effective credit products by understanding your user motivations and behaviors, and aligning the offerings within this context. Square personified this concept, by creating a zero friction product - essentially, they said “we have your entire application on hand, so we’re just going to run it for you and tell you if you are approved.” This was a very novel concept, but one that made a lot of sense. Buy Now Pay Later companies also represent a great example of this - you create a win-win experience for both consumer and merchant by reducing credit friction - and really the traditional cost of underwriting.

Previous interviews :