Will Davis co-founded Able, an Austin-based peer-to-peer lending company, to provide loans at lower rates than traditional lenders. Plaid’s CEO Zach Perret had a chance to sit down with Davis recently and chat about startup learnings, backer-based lending, and how to achieve growth in a competitive market.

Zach Perret: Were there any big learnings that came out of your first startup, Outbox, which digitized U.S. postal mail, that you could translate into the fintech space?

Will Davis: One thing we learned is that you’re not going to win against regulation. What we have to do is figure out where the water can flow instead of trying to figure out how to bust through a dam. What we did with Able, for instance, was to say, “Hey, we’re not going to change lending law, we’re not going to change the SEC, we’re not going to change the nature of payment flows. So let’s figure out what’s possible.”

Unfortunately, the riskiest loans in the whole marketplace of nearly all lending is in commercial businesses. There’s sort of one idea of the American dream in starting a business, and we want to promote that. But at the same time there’s a regulatory infrastructure that makes it harder for small businesses to get loans.

So when we looked at the marketplace, we said, “Well, small businesses are going to find it harder and harder to get really low-cost sources of capital, so how can we be that low-cost source?”

ZP: So you guys founded the company in December 2013. How long did it take you to go from founding the company to issuing that first loan?

WD: When we started, we had a lot of theories about how the wrong lending models were in place. So we started doing some crazy loans to test our theory. First we tried a kind of a peer-to-peer pawn marketplace, but that didn’t really go well. Then we shifted to peer-to-peer car titles and car loans at ridiculously low rates. But we didn’t really like that market, especially once I found out that there could be a possibility that we’d have to repossess a single mother’s car or something.

Then a couple of businesses came to us about asset-based funding, because there’s a code that lets you lay claim to assets without having physical possession of them in small businesses. That’s where we actually got our name: In the banking world it’s called ABL, asset-based lending. And then we started understanding those people, and the aha moment came after that fifth month of testing over five different lending models. So starting in the fall of 2014 is when we started actually making real loans, with a legal structure and with real backers, and we were off to the races from there.

ZP: So let’s fast-forward to fall of 2014. You figured out the model, you figured out the target segment, and you decided to do backer-supported loans. What was the infrastructure from the banking side that you had to put in place along the way in order to actually set that up?

WD: We first tried to sort of pack it together through lots of different products, including balance payments, which I was a huge fan of. We were certainly stretching, but what we were trying to do by getting some of the access and ACH rails from a third party was actually really, really helpful to do early on.

But over time it became sort of untenable because we really needed control over the movement and flow of money from our borrowers to backers, to our senior lender, etc. So we integrated with two banks to do that: Square One Bank and Silicon Valley Bank. Then the ACH access to the rails was the most important thing that we built in that first true year of lending. We turned that on sometime in Q2 or Q3 of 2015.

ZP: And then during that time were you doing all of the servicing in-house, or did you use external servicers? How do you think about collections?

WD: We had looked at a lot of different third-party service providers, but our model is so different that none of them really had off-the-shelf tools. So we built our own servicing tool from scratch, including the administrative back-end to originate the loan, price the loan, build a risk model to do that, transfer funds, link accounts, provide ongoing collections and servicing of that loan, and then report that. Not only was it better long-term, but it was actually easier to build our product around our own back-end site than trying to use an off-the-shelf site or tool and bending it to our unique model.

ZP: There seems to be a number of lending models emerging that are either pure peer-to-peer, balance sheet, or raising an external warehouse line. You have somewhat of a hybrid model. Can you explain a bit more about what backer-supported loans mean and how they’re typically structured?

WD: Think about a backer model as a paradigm. Sometimes companies come onto Able, and we give them a loan offer that is 100 percent from an Able senior lender and 0 percent from backers, because they’re great companies and fairly low-risk on their own. But we give them the option to turn up the dial on backer support to 10, 20, 30 percent and as a result reduce their rate and save money.

For riskier companies, we simply dial the backer requirement up to, say, 10 or 20 percent or as much as 50 or 60 percent. We want all of our businesses to fall in the mid-teens interest rate range, because we know that small businesses thrive when they have lower cost in capital. That may require zero backers or 50 percent backers, but the more you raise the more you’ll save.

Borrowers are given the agency and power to say, “I’d prefer just to get the loan tomorrow and pay no backer and raise no backers,” or, “No, I’m very cost-conscious, so I’m going to go ask five people for $20,000 and unlock $200,000 from Able. But in doing so I’m going to save $30,000 over the next 30 months.” That’s really, really powerful for that small business.

We do all of our lending off–balance sheet to senior lenders, who essentially purchase loans from Able and hold them in a portfolio of their own, which is cross-collateralized with everything else in that portfolio. That allows them to diversify across a group of companies with varying degrees of backer support based on their risk.

ZP: When you think about the backers themselves, who are they? Are they customers of the small business, friends, family?

WD: That’s a great question. We’re averaging a little bit more than three backers per deal. And they’re contributing approximately 18 percent on average to the loans. So for a $100,000 loan, about $18,000 or $20,000 is coming from backers, and $80,000 is coming from Able. On average, about 25 percent of the money from the backers is coming from family, and the other 75 percent runs the gamut from advisors to pre-existing investors, angels in the community, and customers.

I expected the majority of backers’ funds to come from family, and the opposite is actually true. You can imagine a hazard if grandma says, “Oh, you don’t have to pay my loan back, I forgive you.” We wanted something like real, social peer pressure to exist, so what’s neat is that a supermajority of non-family members are backers in our loans.

The other side of the business is the cost of acquiring a customer, but we’re finding that backers actually help us acquire customers more efficiently. Unlike every other lender that has a one-to-one relationship, we have a one-to-many, and not everyone in that backer group is going to be looking for another business to back, but many are. So for every two loans we push out, we receive one referral for a new loan from that backer network. These are advisors and investors who are looking for opportunities, are fans of a business in the community, and are socializing our product among their entrepreneur friends.

ZP: How do you think about the credit model? What are the inputs you use for that model, and how has that evolved over time?

WD: The credit model is really interesting because that was one of the first things that we built. Obviously, we didn’t have any of our own data at first, so any model we built had to be a proxy model. The first thing we needed were inputs from the market, and we found three publicly available databases: Lending Club, Prosper, and SBA. We took all three of these databases and whittled them down to about 30,000 businesses that had the loan size and stage of company that were in line with what our model was going to be lending to. We back-tested that at Experian and timestamped each business owner’s individual credit across the life of the loan at each quarter to see how those loans performed. That built the baseline of our current model.

Once we figured out how to build it on our own, we discovered the need for a novel model of pricing a loan, because whether you are Able, Funding Circle, Bond Street, Lending Club, Prosper, OnDeck, or Square, you’re basically getting the same credit information from the same source, and therefore everyone is pricing the same. So you either need a different way to price based on different and new information you get on a proprietary basis, which is what Square does so well, or you extend a different way of pricing the risk that’s different from everyone else’s. And that’s where we come in, because if we can share risk among tiers, then we can always undercut the price of our competitors.

ZP: Do you see your customers shopping around for loans when they’re thinking about it, or do they usually see you as being in the right place at the right time, talking to them on the right day?

WD: Yeah, they absolutely shop. The scary thing, though, for lenders, is that they shop and apply to multiple lenders at a time and get two or three approvals, and sometimes they take all three at the same time. And you can’t know what’s happening. So we use Plaid, for instance, to monitor bank account activity, not just when a customer applies, but even throughout the application process.

ZP: So the lending industry broadly, especially alternative lending, has had a really interesting 12 months. I think in the past 12 months we’ve seen something like $6 billion of equity investments come into the space, but in the past two to three months we’ve seen a drastic reduction in those investment numbers. It seems like you are having no trouble continuing to grow and raise money. How do you explain that? Is it a flight to quality?

WD: I think one reason is that we have a very different shade of model. We’re not a me-too platform, and that helps to attract really high-quality investors.

The second thing is that we have a raw material that is very hard to come by, which is capital. People don’t like to part with their money. So we’re not just having to raise equity capital to fund operations and grow, but we’re also having to raise debt capital and lending capital. For me, it felt 10 times harder raising lending capital than equity capital, because we had to ensure that our model was correct and that we had good data integrity.

But once that lending capital is secured, there are two different types of investors: an investor from the equity side that’s highly incentivized for growth and an investor on the debt side that’s actually highly incentivized for slow and methodical learning and growth. So the whole time there is this tension that filters into the company itself, so it’s just a very interesting business to build. And because of that I go into detail. You see other startups that end up failing, because it’s very hard to live through those life cycles and hold everything together. But even companies that get there can unravel because they haven’t built proper risk metrics or proper internal tools to ensure that they’re lending to the people and to credit profiles they claim they’re lending to.

And so one thing that we’re highly focused on now is having disciplined growth. Obviously, we don't want to let in everyone that comes in, but to have the disciplined growth to do that, the discipline to grow wisely. And it’s hard for any startup to do that profitably.