Enjoying the podcast? Don’t miss out on future episodes! Please hit that subscribe button on Apple, Spotify, or your favorite podcast platform to stay updated with our latest content. Thank you for your support!
Today, I welcome back to the show Arad Levertov, CEO and co-founder of Sunbit (I last had him on the show in 2019), one of the leaders in offline buy-now-pay-later. From 2,000 retail locations in 2019 to nearly 30,000 across verticals like auto repair, dental, and veterinary services, Sunbit has achieved tremendous growth. But what I found most impressive is this: a 90% approval rate combined with low single-digit default rates.
Arad reveals the secret behind these remarkable metrics lies not just in sophisticated AI-powered underwriting, but in their holistic approach that includes strategic merchant selection, gamified associate training, personalized offers, and a customer-first philosophy with zero fees. Now serving over 4.5 million customers and approaching $400 million in revenue while reaching GAAP profitability, Sunbit demonstrates how the right blend of technology and human elements can revolutionize offline lending.
In this podcast you will learn:
- The highlights for Sunbit over the last six years.
- Why they have remained focused on the offline space.
- The different verticals they have targeted.
- How their lending process works.
- How they are able to approve 90% of their borrowers.
- Why the key to their success is the associates in the physical stores.
- Attributes of the 10% of the borrowers they decline.
- Their average default rate.
- How they decide which merchants to bring on their platform.
- How Sunbit makes money.
- How they are growing their merchant base.
- What their partnership with Stripe looks like.
- How they are partnering with Checkout.com.
- How Arad views the competitive moat he has built at Sunbit.
- What they are doing differently when it comes to collections.
- Their expansion plans for the future.
Read a transcription of our conversation below.
FINTECH ONE-ON-ONE PODCAST NO. 550 ARAD LEVERTOV
Peter Renton
This episode is brought to you by the AI Native Banking and Fintech Conference, happening in Salt Lake City on September 30. This event, hosted by Spring Labs and in its second year, is all about practical applications of AI that are in production at banks today. There’ll be no hype, no fluff at this event, just real-world use cases for banks and fintechs. Join the more than 400 people who will be learning about AI solutions they can purchase today. There’ll be big name keynotes, panels of bankers comparing notes, case studies, and demos from some of the cutting edge AI companies. Use the coupon code PODCAST20 to receive 20 % off your ticket at conference.springlabs.com.
Arad Levertov
The key that we learned in the offline, in the brick and mortar, in order to succeed, it’s not enough to have a great product. This is a prerequisite. But in order to succeed, you need a great product. You need, of course, the finance to make sure that the underwriting works well, but you also need a retail DNA in order to do it and to bring the technology into it.
In the last 20 minutes, I talked a lot about this associate that works for Toyota, but actually presents the offer for the customer for Sunbeam. This associate is a key in our ability to deliver the product. And we approach this associate with technology, with AI, with gamification, in addition to everything else.
PR: This is the Fintech One-on-One podcast, the show for Fintech enthusiasts looking to better understand the leaders shaping Fintech and banking today. My name is Peter Renton and since 2013, I’ve been conducting in-depth interviews with Fintech founders and banking executives. Today on the show, I am delighted to welcome back Arad Levitov, the CEO and co-founder of SunBit. Now, I last had Arad on the show back in 2019 and needless to say a lot has changed at Sunbit since then. We delve into their innovative approach to consumer finance, highlighting their success in offline BNPL, deeply penetrating verticals such as auto repair and dental services. We explore how Sunbit is able to approve 90 % of their borrower applications while enjoying a low single digit default rate. It seems impossible, but once you learn their unique approach, you begin to understand the method that results in such stunning metrics. Now let’s get on with the show.
Welcome back to the podcast Arad.
AL: Thank you, Peter. Great to be here.
PR: Okay, well it’s been a few years 2019 was was the last time I’ve had you on so within the six years a lot has happened I know but maybe you could just tell us some of the highlights of Sun bit’s journey over the last six years
AL: First of all, a lot has happened in the world, know, years ago, I guess.
PR: For sure, for sure.
AL: A lot happened in the world and also a lot happened for Sunbit, which is great. You know, when we started Sunbit, we said that we’re to bring technology and machine learning and AI to build a financial services company that is focused on the consumer, with the consumer and the customer in the center. And we started with a binary product and with a vertical of car dealership. think when we talked in 2019, we were getting into the car dealership, we had about 2000 locations, maybe a few hundred of thousands of customers that used the product with with some bit and looking today we have multiple products. We have a binary letter that is active in nearly 30,000 locations in different verticals like car dealerships, repair, dental, eyewear, vet. have embedded finance. We also have a credit card product that is in the hands of hundreds of thousands of customers and that you can use Sunbit anywhere. With all this, while we were growing, serving more than 4.5 million customers, always remember that behind these huge numbers, there are great customers and we never charge any fees and remember our mission to do it. So keep growing, but not forgetting why we started actually back in 2016, what we want to do.
PR: And it’s still, you’re remaining purely in the offline space, is that correct? Like you’re focusing on some of the, the brick and mortar companies and healthcare and that sort of thing?
AL: Correct. So our go-to-market starts with B2B2C from the offline space. This is our DNA. But then when the customer gets the loan and work with us, then she can get a credit card and can buy anywhere with it. But yes, the approach is to go to the offline space. And the idea was that this is huge market, much harder to disrupt because you actually need to figure out a great technology to get to the own offline space and to be there and to work with stores and to leverage everything that you can do from the technology on the e-commerce into the offline space. But once you do it, it sticks, it works well and customers appreciate it and stick with it.
PR: Right. And so what are the primary verticals you’re working in, in the offline space?
AL: So our buy now, pay later, started, as I mentioned, with car repair dealerships. Today we are in more than 60 % of this market.
PR: Wow.
AL: There are 17,000 locations where you can fix your car in the authorized dealership. And we are in nearly 10,000 of them. And we have partnered with all the top 25 groups, like AutoNation, Osbury, Group Run, all of them, we partner with them. We have endorsement by the 15 or even 17 OEMs. They actually push us into their dealerships. This is the first vertical. The second vertical is dental, which is interesting. Dental, we’re actually in more locations. We are in nearly 14,000 locations in dentistry. However, there are nearly 200,000 dental offices in the country. Like my son used to say, we have more teeth in the families than cars. He’s huge. And then we also have penetration into vet into eyeware. And we also work with companies like Stripe to get embedded finance into additional verticals like car services, which are small locations, Med Spa and others.
PR: Okay, interesting, interesting. I’d like you to run through for the listeners the process when someone’s at a dentist or an auto dealership and suddenly they realize that it’s a thousand dollars or more and what’s the process like at this offline location?
AL: I will say, but if it’s a query, I will actually start even before that. We, and what we did from day one is to use data and machine learning and AI to decide which office and which vertical we’re going to walk. Okay. And also how we incentivize the, uh, the person in the office, uh, the one that offers the service to the customer, um, to offer something. And that’s even before the customer applied. Right. We would tell, let’s say Peter walks for a car dealership as a service advisor. And we will tell Peter, Peter, you have a great tool, which I’m going to describe so in a second, that within seconds, you can approve nine out of 10 people that apply to, and they will never get any fees and they will get personal rates. And by the way, Peter, as a service advisor, if you help a customer, you’re to get points and you’re going to have gamification and you’re going to, you know, to be a hero and to be number one in the country of helping customers. So this is even before the customer applied. then, so Peter was trained, we decided to choose the dealership because we know the customer will come back and we have some relationship with Peter. And now when the customer comes, the process is super, super fast. What we ask for the service advisor or for the people that work in the dental office, and we have about 150,000 associated, not work for somebody, but they’re in the system to show the customer the ability to pay over time. And we tell them, they tell the customer, you need to fix a car for $1,000, but don’t worry. We have ability to get the service now and pay over time. What they need from you is only the driver license. Scan the driver license. We basically scan the barcode at the back of the driver license. Then within seconds, you can see if you can get three, six or 12 months, or maybe up to actually 48 months of splitting your payments. Within seconds, we basically allow the customer to see how they can pay for the unexpected, unexpected repair. I would say that all the offers are personalized, which means one customer will get a different offer than others. We have many, many tiers, each customer will get a better offer compared to what she has in her pocket, which means if you pay there as a prime customer, you will definitely get a 0 % that you will not have to pay anything, which is better than what you pay for the credit card. If I am a sub-prime customer, will get an offer, which other traditional lender will not offer me. The offer will be much better than even a payday loan, of course. Within 30 seconds, we approve nine out of 10 people and each customer get the best offer. So the customer gets, scan the driver license, see the option, choose whatever he or she wants, complete the transaction, get back to your life.
PR: Okay, well there’s quite a bit to unpack there, but the first thing I want to focus on is this nine out of 10 approval rate. mean, 90%, that’s unheard of in the unsecured consumer lending space. How are you able to have such a high approval rate? I presume you’re working with different lenders with different credit risk appetites on the backend, but tell us a little bit about that.
AL: I will say that we take all the risk. will say before, before I would say that how we do it, will say that we never charge any fees. So no late fees. We do not go above the user rate, which means we don’t charge more than the consumer lenders are charging. And we, we are the lenders. I mean, we work with the bank, but we take all the risk and how are we doing it? It actually starts, and that’s a little bit counterintuitive to explain, it starts before the customer even applies. We use artificial intelligence across the system. So if the associate that works in dealership offer knows that there is a great product that will be approving every customer, this associate will offer it to more customers. If the offer is more personal, which means if a prime customer will get a good offer and the subprime customer will get a good offer, which by the way, takes 30 seconds, more customer we will willing to apply. And then this without again, before even I did the underwriting gets some bit more customer that actually is more like a normal distribution. So you get more good customers that join you. And then it helps you to approve more that riskier customer. And that’s how we do it. And we use AI across like the system. We personalize the offer, we adjusted down payment. We adjust the amount. We just, of course, the APR, we just a risk to give and it works. We did it more than 4.5 million times. We have great credit facilities. We just did a great securitization deal with a AA rating, which means the great performance are there. And when we started, people ask us how we do it. I always try to explain that it’s not only the underwriting, it’s actually looking at the entire process and using data from choosing your partner, incentivize your associates, treating your customers well, because once you treat the customer well at the end, they will come back and use this again. And actually almost 40 % of our loans come in from repeat customers. And all these get us into a great position that allows us not only to keep growing at the 35 % over here, but also to be profitable, GAAP profitable without any adjustment.
PR: The 10 % that you’re declining, they deep subprime? Are there problems with identity verification? I what is the reason that you would decline someone?
AL: So these are really, really the edge of, as you mentioned, sometimes we are really the rare, right? There is some problem with identification, some risk of fraud and we use a lot of technology to identify it. And this is kind of really the edge cases. In some cases, by the way, we have programs that we call it approve all, which we approve actually 97%. So only when there is a really problem on identification, we don’t do it. But the idea is really to try to give every customer a great offer, a great rate, and we take the risk on us. Like if the customer doesn’t pay, we don’t make any money from it. We actually blame ourselves and do better job to make sure that we make money.
PR: Right. Well, that’s going to happen obviously when you’re in doing these loans. Can you share what default rate is your portfolio is receiving?
AL: Yeah, I think, mean, you see that the rating that we got and the raise that we got into securitization that we just did a few months ago, and you see that we have great performance. The default rate varies, of course, on the vertical, but it’s low single digit, and this is great. We adjust the amount, we adjust the down payment, we adjust everything, right? Of course, a separate customer, if two customers, one is a subprime and one is a prime customer. We want to fix the car, each one of them for $1,000. One will get a lower down payment than the other. And we’ll make sure that every customer will be able to pay. And usually the loan amount is small. So a subprime customer can make a $50 monthly payment. You just want to make sure that this customer would prioritize the payment to you to Sunbit and we use data to predict this customer behavior. We call it, it’s not a willingness to pay, it’s not an ability to pay all the time. Sometimes it, what we call willingness to pay and then we try to predict the behavior of the customer by using the artificial intelligence.
PR: Right, right, okay.
AL: I can give you a simple one example, as you example, right? So let’s say there is actually a prime customer with a 700 FICO score. And you look at the data, you look at the file and it looks that this customer pays for the mortgage and pays for the auto loan. But this customer is late for mobile phone, the AT &T phone for like $50. You see, is something that doesn’t make sense for this customer. is like something irrational with this customer. Why would you risk your credit for $50? It may not hurt the FICO score so much because they have some other stuff, but this kind of means there is something irrational. So we would approve this customer, but maybe charge a higher down payment to make sure that these customers really want to take the loan with us. So we look at these things and we try to make the best decision. And as I mentioned, the decision actually starts even before the customer actually applied. With the vertical we went and with the associate and actually worked on it.
PR: Right. And so that brings up my next question. Like how are you approving the, say there’s a, you know, in the auto, in the auto repair, there’s a lot of mom & pop operations, right? Small companies like, you know, maybe three to five employees and operating in a relatively small town. How do you decide who to approve to come onto your platform?
AL: That’s a great question. We use data and I’ll give you an example. In the auto dealership, we usually work with the authorized dealership. actually when you need the authorized dealership, Toyota, Honda, that means it’s more than this. But I give you an example of the dentist, right? As I mentioned, there are 200,000 dental offices in the country. some of them is a dentist office that actually the dentist, the doctor comes every day five times to the office. In some cases, the doctor will come only two times to the office and the rest of the five days will be maybe only smaller service or different service. We take all this into account when we kind of analyze how much to invest in each merchant. So, for example, for the dentist that comes five times, we estimate there are many customers. We may actually fly a person to train the staff over there to make sure they’re all equipped to help all their patients. the dentist that maybe use only, you know, with the office that actually the dentist comes once a week and the rest is something that is small procedures, we may do a Zoom call or maybe do online training or AI training. We’ll send them AI video and they will be trained by AI. So we manage the cost of how much we invest in each merchant. And then this is only the beginning. Then we continue to work with the merchant based on the tiers and continue the relationship. So that’s how we use the data to make sure that, you know, help us to what we call the CAC to LTV, how much to invest in each merchant and how much this merchant can produce to themselves and also to us.
PR: What is your business model, how are you taking an origination fee on the loan or are you making a spread on the interest? I how are you making money?
AL: So I said, no fees.
PR: Yeah, I know you said no fees. I knew you’re say that.
AL: No late fees, 35 to 40 % of our loans actually 0 % doesn’t add any cost to the customer. So how do we make money? And we are making money. I said, are probably approaching $400 million in revenue run rate and we are profitable, we got profitable. So we make about close to half of our money from the merchant. So for the merchant is paying us a fee for helping them to close the transaction. so for $1,000 transaction for a dental for, repair car repair, we let’s say we’ll, we’ll transfer the merchant the next day. Cause we take the risk. Let’s call it 950 and this number is very slight vertical. we, we, we keep paying 950. And even if the customer doesn’t pay anything, but just give us the money back. We basically made $50 on this example. The other thing is the merchants are paying us also monthly fee that allows us to invest in the training and work with them, just subscription fee to do it. And for the customers who actually take the loan and pay interest, we would charge interest and make money from the spread. But no fees, no late fees, no origination fees, annual nothing. That’s what we do. And this is with the BNPL product. Then we have the credit card product, which we offer to our customers as a second product or with partners of corporate card when we make the interchange and also the interest spread. Okay. But again, no fees.
PR: Yep, that makes sense and understood. Okay. So as you’re growing, is this a one by one type deal when you’re adding merchants? It seems like that would be slow and expensive and you’ve grown quite fast since we’ve last spoken. What’s your approach to, I guess, go to market on the merchant side.
AL: So it’s again, this is where data, AI tells us where to go. There are some verticals that is actually worse to add one by one. That’s okay. Mostly in the car dealerships. And we started with just knocking on doors. And then once you knock on doors, you go to Honda and you tell them Honda helped me to get to all your 1000 dealerships in the nation. And then it tells. There are some merchants and some verticals that you need to work with partnerships in embedded finance to go faster. This is mostly, as you mentioned, smaller verticals like Med Spa, when the store is smaller, veterinarian, some home improvements that are all home services. Over there, we do partner with SAS platforms that provide services for these merchants. And then Sunbit becomes a payment option on their system. So in this case, we don’t have to send a salesperson. Of course, the margin is different because everybody needs to get a cut, but it doesn’t change the story of focusing on the customer, providing the customer with great solution and providing the merchants more sales. So, it depends on verticals. That’s what I’m saying and I said it a few times is that. We use the data and AI to look from all the way to reach out to get to the merchant, to how we serve the customers, which we do all the service in the in house internally.
PR: You mentioned Stripe before, and I want to dig into that a little bit. What’s your partnership with Stripe? What’s that look like? How are you adding new customers that way?
AL: So, Stripe is part of the embedded finance strategy that we’ve developed over the last probably three to four years. And the idea is exactly based on the previous question, how do we help more merchants without sending person after person after person? then what Stripe enables us is, Stripe is a payment processing that works with a lot of SaaS platforms that services merchants across the nation. And our partnership with Stripe allows those SaaS platforms to offer actually buy now pay later when they offer Stripe. So with Stripe, we actually partnered with a company, for example, called Shop Monkey that service a lot of small mom and pop shop on the car repair. And with Sunbit and Stripe, they were able to just turn on the ability for these mom and pop shops to offer buy now pay later. This allows us to go to many locations that are probably smaller and we don’t have to send a salesperson in person to sell to them.
PR: Gotcha. And what about checkout.com? Because they’re, I think of checkout.com more as in the online space. What are you doing with them?
AL: So, checkout.com is a great partnership for us. Actually, it’s more on the hour operation. When we collect the money from customers, we use a lot of the debit card of the customer because remember what we try when I describe the process that takes place in the store, we want to make it smooth and fast and feels like the customer only swipe a card on the machine. we, a lot of time we…the end customer uses the debit card to pay the down payment and pay the monthly payment. So we use checkout.com to collect it and to optimize it, reduce our cost that we can pass the savings back to the consumer.
PR: Gotcha, gotcha. Okay, so I wanna talk about your competitive moat, you’ve built something that is not gonna be easy for a new competitor to come in head to head. You’ve got the distribution moat, and you’ve also talked about the data, and all of the AI that you’re doing, you’ve got the technological moat. So maybe you could just talk a little bit more about when you think about the competitive market, how do you view the moat you’ve built and how you like, how do you sort of view the overall, particularly in the BNPL space, which is it’s a big space, but it’s also competitive. How do you view the moat you’ve built?
AL: So first of all, it’s a big space, but the financial services and consumer finance is huge space. mean, the market is getting started. We in the FinTech space talk about buy and pay later, but when you look at credit card spend and just in general consumer spending, it’s just a fraction of this space. So we are just getting started. For many, many, many, years, lenders had kind of this approach of, if you want good performance, you need to increase APR. If you want great return, you need to decline more customers. What we were able to do is to do that away. We can get great performance. We can do it with lower APRs and we can do it without any fees. And the way we do it, it’s not only in the underwriting, but it’s the entire ecosystem that works with approaching the merchant, training the people, have a great capital markets behind us, and of course, amazing underwriting and personalized offers. And it started also with go-to-market. So the key that we learned in the offline, in the brick and mortar, in order to succeed, it’s not enough to add to a great product. This is a prerequisite. But in order to succeed. You need a great product. You need, of course, the finance to make sure that the underwriting works well, but you also need a retail DNA in order to do it and to bring the technology into it.
In the last 20 minutes, I talked a lot about this associate that works for Toyota, but actually presents the offer for the customer for Sunbit. This associate is a key in our ability to deliver the product. And we approach this associate with technology, with AI, with gamification, in addition to everything else. So I would say that the mode is not only in the location that we are, but also with the relationship that we have with the merchants, with the associate, with the customers that love Sunbit and takes Sunbit again and again and again. And overall, with the reputation we are building with the customers. We have net promoter score in the high 80s. As I mentioned, we have over 30 to 40 % of repeat customer users again. Once you fix your car with Sunbit, you’re going to continue fixing your car with Sunbit. And unfortunately, there are no shortcuts, but if you do it day after day after day and customer like you, and not only they will use you again, they would also prioritize the payment to you and they will not switch for you. So we see it the car dealership, we already 60 % of the market. We can see it in more and more verticals.
PR: So what we haven’t talked about yet is collections and to have a low single digit default rate while approving 90 % of customers coming in the door and obviously different terms for those customers. But still, the collections piece has to be critical. What are you doing there that’s different?
AL: So it start with the fact that, as I said, we do not charge late fee. It starts with the underwriting, it started with the way we choose the merchant, it starts with the amount that we get. And let me explain how this work. A lot of customers, especially on the subprime, the monthly payment is maybe 40 or $50. Because again, we reduced the risk, so we lowered the monthly payment. Nobody wants to get in trouble. And I say get in trouble means reporting to the credit bureau for 30 or $40. So what we do in the collection and we use data, we use technology, we send email, know, of course, the phone, but we basically help the customer not to get into problems with the credit bureau. So we use all the technology in the approach to do it. And we do it all in-house. And we have the data loop between the call center in Vegas, which is US-based to the sales team. If we see some more complaints or high collection from one merchant or another, we go back to them and train them. So we use this loop of data from the beginning to the end to make sure that the performance is good.
PR: Okay. That’s interesting. So last question then, as you’re looking forward here, there’s obviously lots and lots of verticals you could go into, what are your expansion plans for the next two to three years?
AL: I will start with the future. In many industries, technology already changed the way the industry works. mean, we’re doing right now a Zoom discussion, which couldn’t have been the case 20 years ago. We read media differently. But in financial services, especially consumer finance, still it stays the same. Like the statement of the credit card that you have in your pocket is the same statement that was like 30, 40 years ago. Which means it’s not personalized. It’s focused on the company instead of the customer, which has hidden fees and high fees. And this will not be the case, whatever, 20 years from now. It will be personalized. It will be focused on the consumer just because it will get there. And we want to be there, one of the leading companies that does it. The way we do it, we do it by focusing on the customer and we’ve taken different approach of go to market and definitely using AI across the entire operation. We started with the car repair, going to healthcare right now and healthcare is a huge opportunity doing embedded finance, but keep developing more products for consumers, not only buy and pay later to help them manage their life in a better way. So if you think in the coming few years, huge opportunity in healthcare, huge opportunity in car repair, unbelievable opportunity in embedded finance, in addition to more products. As I said, we have a credit card. Well, actually launching another credit card. We’re going to have more product for the consumer after the use of Sunbit to manage the life better.
PR: Okay, well we’ll have to leave it there, Arad. Really great to chat with you again. Congratulations on the success you’ve had over the last few years and best of luck in the future.
AL: Thank you, great talking to you, great seeing you again, listen to all your podcasts. Thank you, Peter.
PR: Thanks, all right. See ya.
It was interesting that Arad continued to mention the people on the front lines of his business. The person at the auto repair shop or dentist office who is helping their customer with financing. To achieve their touted 90 % approval rates and low single digit default rates, you need everything working in unison. The underwriting has to be spot on. Your data collection and identity verification needs to be flawless. And the person dealing with the customer has to understand what you’re trying to do so you can get a positive selection bias. I have a feeling that the human element might be the key to such great numbers. Anyway, that’s it for today’s show. If you enjoy these episodes, please go ahead and subscribe, tell a friend, or leave a review. And thanks so much for listening.