Part 1: An Intro to Online Lending

In conjunction with our podcast on Invest Like The Best w/ Patrick O’Shaugnessy, Brian Harwitt, Marc Porzecanski and I have decided to publish a series of blog posts to walk people through our investment process.

The content below (while posted through my medium account) was written by all three of us.

Our goal is to bring transparency to the often opaque world of asset-backed lending.

Below is an outline of the posts we have written:

(1) Part 1: An Intro to Online Lending (LINK) ← You are here

(2) Part 2: An Intro on How To Source Deals [LINK]

(3) Part 3: Initial Diligence [LINK]

(4) Part 4: Deeper Diligence [LINK]

(5) Part 5: Structuring The Deal [LINK]

(6) Part 6: Building a Credit Model [LINK]

(7) Part 7: Monitoring Your Investment [LINK]

(8) Part 8: Conclusion [LINK]

****************************

Online lending became popular in 2006, when LendingClub realized it could make loans online, and then sell those loans to Investors for a fee.

LendingClub would:

  • Use its technology to efficiently find borrowers who were looking for loans, and who could no longer borrow from banks that who had cut back their lending practices post financial crisis
  • Leverage software to quickly analyze data about these borrowers, and underwrite them to assess their risk profiles
  • Finally, sell the loans to individuals and institutional investors on its platform that wanted exposure to unsecured consumer credit

Investors were excited to invest in these loans, because it felt like they were getting direct access to a high yield in what was becoming a historically low yield environment. (Instead of holding their money with banks who were offering next to 0% returns on deposits, investors were able to use LendingClub to invest capital directly, which was a powerful concept).

And this kicked off the Online Lending movement, which spawned OnDeck, Prosper, SoFi, and many other companies who all hypothesized that they could:

  1. More efficiently find borrowers
  2. Use alternative data to underwrite borrowers in a cheaper, and often higher quality way
  3. Use their technology to access capital markets in unique and efficient ways

Efficiently Finding Borrowers

Online lenders had the advantage of not needing to operate brick-and-mortar retail locations. This lack of overhead allowed technology startups to originate loans that were smaller, and shorter in duration, without having to charge high rates or fees to make the economics work.

Traditional lenders, who operate brick-and-mortar locations, often had to charge high interest rates, together with extra fees (closing fees, servicing fees, etc.) to improve the unit economics of each loan. Compared to an online lender, this put them at a disadvantage.

To illustrate how difficult making certain loan products work while managing the costs of a brick-and-mortar store, take payday lending as an extreme example:

Originating payday loans profitably is difficult due to the duration, the size of each advance and the high probability of default. If a borrower wants a $500 loan, and the loan’s due date is the next pay cycle (say in 12 days), the lender is almost obligated to originate a very expensive loan. The reason is…

If the lender charged a 25% annualized interest rate (with no added fees), it would be very hard to earn a profit. A 25% annualized interest rate on a $500 loan over 12 days generates only $4.10 of interest. If the loan takes an employee 20 minutes to originate, and that employee is earning $12/hr, then the loan costs $4.00 of human labor just to originate. That’s before the cost of marketing to the customer, the real estate costs of leasing the storefront, the paper the loan is printed on, the overhead, the regulatory compliance, etc. — let alone covering the cost of borrowers who will default.

And therefore, payday lenders and other lenders of short-term loans must charge the high rates and fees that we all cringe at.

Payday lending is an exaggerated example, but it illustrates why brick-and-mortars have embedded cost structures that can make their loan products less competitive, or difficult to offer profitably.

Online lenders are able to strip out a lot of these costs by operating online, and offer smaller, and shorter duration loans at more reasonable rates than a brick and mortar business ever could.

Using Alternative Data

Online lenders promised the use of alternative data to their investors. The idea that FICO was still the primary driver of whether or not someone could get a loan felt outdated. After all, FICO was founded in 1956, before the world was flooded with 2.5 Quintillian (← that’s a real word) bytes of data per day.

So these online lending platforms promised to use new sets of data to answer questions such as:

  • Why should a student at Stanford have to pay the same for a student loan when that individual was likely a lower default risk?
  • Why should a business that had a ton of foot traffic each day have to pay the same rate as another business on a less busy street, with less foot traffic?

Access Capital In a New Way

Finally, Online Lenders promised the ability to access capital in new ways. How could any Adam Smith loving individual dispute the idea that a “marketplace lending model” or a “peer-to-peer lending model” would ultimately become superior than another model rife with brokers and middle-men/women?

The Result of These Three Advantages?

Ultimately — some of these “Advantages” ended up playing out, and others have yet to be proven.

For most of these online lenders, FICO is still their primary driver for underwriting. We often hear pitches from founders who tell us they “collect 150 data points per borrower.” Our immediate next question is: “how many of those data points provide real signal and act as a leading indicator of default rates?” The answer is almost always “not many.” It’s simply harder than most people think to use “alternative data” to materially improve the performance of loans.

And most of these lenders have scaled back their use of “peer-to-peer fundraising methods,” because they found that the quantity of investor dollars was much greater with traditional, institutional, financing.

OnDeck uses multiple credit facilities to finance its loans, and so does Avant, and so do most of the other lenders that have scaled — mostly from well-known Wall Street banks. This is both to diversify their lender base and because different types of loans — each originated from a unique “credit box” — require different types of capital.

And finally — while in many cases the costs of originating loans has come down, the cost of acquiring customers has gone up due to competition. Following the success of LendingClub, many new entrepreneurs realized they could inexpensively build their own lending businesses that would one day compete to originate and underwrite unsecured consumer loans.

The more people bidding for ad space on Google and Facebook to sell the same commoditized product, the higher the cost of marketing became, and the more difficult it became for these companies to grow at their promised pace.

All three of these supposed advantages have had mixed results. Returns for investors across all platforms, for the most part, have come down, and the equity valuations of many “marketplace lenders” began to reflect that, at the end of the day, these were specialty lending businesses using their technology to create operational efficiencies. The public multiples of these businesses began to reflect that as well.

While a number of CoVenture partners were early investors in these businesses, some of which will sustain enduring advantages and remain valuable over time, we have since focused our attention towards the next wave of lenders: “The Lending 2.0 Companies”, which we believe will return higher yields for us as lenders, and for us as equity investors.

Lending 2.0 → Where We (Think We) Are Now

We define Lending 2.0 Companies as companies that are using their technology to:

  1. Observe a previously unobservable data point to invent a new type of credit — where the yield is high not due to greater risk, but rather due to a lack of capital in the space
  2. Build a barrier to entry that will allow their yields to stay high, despite an increase in competition

*The companies that characterize these two behaviors are our “unicorns.”

Observing a New Data Point to Invent a New Credit Product

Technology companies can build software that imbeds themselves in the workflow of a borrower, and use the borrower’s day-to-day interactions to underwrite credit risk, and provide a new type of loan never previously available.

For example — if the barista of a coffee shop is owed $1,000 by her employer, and wants to take an advance against the amount of money already owed, a lender could integrate into the coffee shop’s scheduling system. It could then confirm that the money is owed (by knowing how much the employee has worked), and lend to the barista based on the creditworthiness of her employer.

Barriers to Entry

Barriers to Entry are often the hardest thing to build in a lending business. Often you’ll hear an investor say something like this:

“I am getting a great yield because the banks pulled out of [x] space, and so I am lending at a higher rate because there is a lack of capital in the space.”

This is a great way to get access to outsized returns for some limited period of time. However, as soon as other lenders realize there are attractive rates to be earned by financing these underserved borrowers, more capital will come into the space.

This means that marketing to these borrowers will get more expensive. Other lenders will begin bidding on ad space and will pursue competitive acquisition techniques. Further, rates will get compressed (as borrowers have more options of who they’d like to take a loan from).

At CoVenture we’ve found three strategies that build barriers to entry that make sure a lender is able to not only begin investing in an underserved space, but keep doing so at a similar yield even when other lenders begin to compete.

(1) Switching Costs | Some lenders have built technology, that integrates with their “Point of Origination.” This gives them defensibility via “switching costs.”

Example, Imagine a lender who provides Point-of-Sale software to a jewelry store. The PoS software handles orders, cash, inventory, etc. But even better than that, the PoS has the ability to offer loans to the customers of the jewelry store.

This allows the jewelry store to sell to more customers, because instead of only selling to people who could afford to buy a necklace for one lump sum, it allows customers to purchase that same necklace by making monthly payments over time.

The lender who is providing the PoS software, and originating the loans through it, has a defensible moat. It is much easier for this lender to get access to its target borrower (the jewelry store customer) than another competitor not integrated in the sales process. And if the lender started offering 18% rates to the jewelry store’s customers, it’ll likely be able to maintain that yield, even if others attempt to compete.

If a new a competitor tells the jewelry store it’ll build PoS software too, and offer 16% rates to its customers, the jewelry store likely will not switch — because the cost of learning a whole new PoS system is greater than the burden of offering a 2% higher rate to jewelry store’s customers.

If a competitor offers 15%, the answer would probably be the same, and the same at 14%. (At some point there is a level that is just too great, just like eventually there will be a much better PoS system, but this example is more defensible than traditional lending).

(2) A data point others cannot observe

In some cases, there are lenders who have built businesses where they can see horizontally across an entire industry. They therefore know what is going to happen, before it is reported, based on the first party data the lender has been able to collect.

Companies that can observe borrowers across multiple platforms are powerful. For example, the type of lender who can observe a user’s propensity to cancel Uber rides last minute is likely to be able to observe the user’s likelihood of cancelling Lyft rides last minute.

There are lenders who can observe analogous behavior on one platform to predict behavior on another platform, and underwrite against this. This gives those lenders data to use, to offer a loan other lenders wouldn’t be able to underwrite.

(3) Ability to affect outcome

Lenders who can actually help create a repaying loan, are some of our favorites. For example: imagine if Amazon made a loan to one of its vendors who was late on its payments. Amazon could then take that vendor, put them on the company’s home page, drive a lot of traffic to their site so the defaulting borrower could do more revenue, and make them current again.

More practically, marketplaces could lend to the suppliers on their website. If the suppliers start to fall behind, they could promote those suppliers to the buyers on the platform.

— —

Online lending has spawned a new level of creativity that we believe will be enduring, and a part of our DNA forever.

We believe that the companies using technology to invent new types of credit, and who have built barriers to entry, will end up becoming the most exciting for us to invest in, from both the debt and the equity.

And while many capital allocators have been given mandates to high returns that feel impossible to meet in today’s environment, creativity may provide an opportunity to increase returns without increasing default risk.

As a lender to these companies, we try to partner early on in a company’s life. This is when we feel like we can have the greatest impact, and be a part of a company’s “story.” It also allows us to invest when the yields are still high, and maintain high yields with the company as it continues to grow and scale.

Over the following blog posts we’ll discuss:

  1. Our underwriting processes — how do we decide on what yields make sense, how we determine our downside, and how we model out expected performance
  2. The types of financing we provide — and analyze the structure of each type of lending (we will go through legal terms of each type of structure, the points of negotiation, and the advantages/disadvantages of each type of facility)
  3. An analysis of key takeaways from having lent to a variety of companies across consumer and small business loans, and specifically niches in the 1099-worker receivable space, auto loans, farm lending, and beyond.

The past few years have been fun for us, and we’re excited to share as much as we can about what we’ve learned.

See below for a link to our next post on [How to Find the Best Deals].