https://ir.upstart.com/events/event-details/upstart-network-…
My notes from the UPST CEO fireside chat with a Piper Sandler analyst that took place on Sep 13.
On your last earnings call, one bank eliminated FICO requirements.
Some of the checks I have done on your partners is they have reduced FICO score minimums. You have a better underwriting model and better customer experience as 71% are automated. What more can we see over next couple years?
We have a lot of AI models underlying our system. But principally, first and foremost, they decide who and what bank should offer a loan to and at what price and very specifically pricing that loan to the risk profile of the individual, doing that based on training over a million loans and 50,000 repayments or delinquincies that happen every day. It automates every step of the process and happens almost instantaneously and provides a better experience for the customer.
[Me here: 50000 payment due dates a day equals 1,500,000 currently active loans, right?]
You haven’t focused on repeat customers. How are you going to go after repeat customers to monetize them?
We are still a relatively young platform so most applicants are first timers. We view this as a good thing, because there is enormous upside in the future to offer other products such as we recently launched auto refinance product. We think we can serve this customer again and again especially as we have extremely high NPS in the 80 range which is unprecedented in financial services. That quality of experience brings people back to us. Now of course we are not a bank so we will continue to refer them to the partner banks.
Can you size the auto opportunity and why buy prodigy and not go it alone?
Expanding from personal into auto made a lot of sense as we had to learn something new in secured loan versus unsecured loan. The opportunity is large, 6-7 times the size of personal loans. This is important to the banking world too and entering into auto let’s us step out into their limelight. It’s also an incredibly inefficient world and risk is not accurately modeled. The process of buying a car is also not pleasing for anyone. With a much better loan and automated process we can have a bette experience for the customer, better profitably for the dealership as Prodigy gives us a footprint in the dealership to help them sell more loans and better product for the bank, it might sound like magic to make all three parties happier but that’s the opportunity we see ahead of us.
I agree with you, no one ever says let me go get a root canal and then I’ll go buy a car. (Laugh)
Can you talk about what’s near and long term for auto refi and auto purchase?
Our intention is to participate in the whole market of auto finance. We already launched autorefi, started end of last year in one state, and since rolled out into 47 states and now we are just fine tuning and scaling it up. We believe most purchase loans happen at the car dealership. So prodigy is like a shopify for dealerships with a modernized car buying process and allows us to bring loans to the point of sale by end of this year. I call the auto industry as the world’s largest buy now pay later industry because pretty much everyone finances their cars.
Can you talk about the differences in pricing and margins in auto versus personal?
It’s still early stages, but we know auto loans are about twice the size of personal loans. We believe dollar contributions per auto loan will be similar to personal loans but it will work out over time as costs of funding reduce and more banks participate. We are not looking to maximize profit per loan, it’s not how we think, we are looking to make the best consumer product we can and at the same time make a reasonable gross margin for ourselves - I mean that’s how we built such a fast growing and profitable business in personal loans, sort of Amazon like, we aren’t looking to over monetize, we would rather make a better consumer product and just charge banks only what we need to have a strong and growing business.
[My reaction: what an incredible mindset!]
As we look at the next 12-18 months, what milestones should we look at in auto?
There’s a bunch of things we watch, our whole business is based on conversion rates and the funnel, if conversion rates get better then the volumes go up, so the most fundamental thing we do is refining and refining and automating the product so conversion goes up, exactly the way our personal product tracked, so that’s number one, the second thing we will share is getting more bank and credit unions support for auto lending which is the money/supply side, and the number of car dealerships using Prodigy across the US. We anticipate scaling from 2022 and beyond, starting from a negligible market share today but we are very excited for the next 3-5 years because we believe we have the better product for all parties involved.
How should we think of the TAM in 5-10 years long term? What are your ambitions?
There’s the notional TAM, which is many many trillions dollars of credit in every flavor around the world. That’s a half baked answer. The way I think of it is when we learn a new skill or trick, we get to leap into the next category and the TAM expands. So we learned unsecured and now we figured out how to deal with secured loans properly like liens titles and different credit model and different waterfall of payments. The next secured type could be home loans, or business loans, or international. We keep unlocking more TAM the more we do things. Unlocking secured loans was a big deal for us. It all comes down to how well we execute. We don’t need to do this serially, one at a time, I think you will see us as a company fan out a bit and execute against a few different opportunities simultaneously. We are looking to take the company across multiple products and geographies simultaneously.
We see Upstart as an AI company that happens to be in lending. Others see it as a fintech. How do you characterize Upstart?
We are first and foremost an AI tech company. We are deep vertical experts in the field of credit and credit risk. There are adjacent areas we can apply ourselves to like insurance. We are not a general-purpose AI company. We don’t aim to become a generic tool provider competing with Amazon or Google. We want the very best risk models related to credit and lending in the world.
What are you most excited about the next 12-18 months?
Having bought cars for my kids recently, the car buying experience is distinctly unpleasant and we are excited to really make a difference there. It’s going to be some of the most important work we do over the next 12-18 months.