Riskified Ltd. (RSKD) Earnings Call Transcript & Summary

Good afternoon. I'm Reggie Smith. I cover fintech stocks here at JPMorgan, and I am happy to interview the CEO and Co-Founder of Riskified, Eido Gal. He's in town today. We're going to do a Q&A. I've got a couple of questions lined up. I'm going to open it up to you guys towards the end of the session for your own questions. Welcome.

Thank you. Thank you for having me.

For sure. For sure. So Riskified, a company that we did the IPO a few years ago. Still a fairly new company, not well known. And so I want to let you start by just kind of giving just a feel on what Riskified does, how they make money, what problems they're solving, and we can kind of go from there.

Sure. So when we started the company probably about 10 years ago, the thesis was that we can leverage machine learning to better identify e-commerce fraud. And we focus specifically on enterprise e-commerce merchants. And really, the market we operate in is if I were to steal your credit card while we're speaking, order a bunch of iPhones online, ship them to me, so I would check it out, I would pack it somewhere online. And then you would see it on your statement. You would ask around at home, did anyone use my card? Everyone would say no. You would call your bank, JPMorgan, I assume. You would say, someone stole my card. This is an unauthorized card usage. You're protected as a consumer, right? So the bank would refund you the money. The money gets taken out of the merchant's account. So merchants, once they accept card-not-present transactions, they are liable in the event of fraud, right? So that really starts our world. And when we started Riskified, we said it doesn't make sense that we have hundreds and thousands of enterprise e-commerce merchants. Each one is trying to look at incoming transactions and identify if they're fraudulent or not, right? And they're building these large complex teams that are using a lot of third-party tools and data enrichment and systems, and they're basically screening every incoming transaction, and they have a lot of costly mistakes. They accepted bad transactions that the business needs to pay us a chargeback. That's like over 30 basis points from their overall revenue rate. They turn away a lot of good customers because they fear they might be fraudulent, okay? And they -- it's not a great experience. Sometimes, you need to send transactions to manual review and you can't offer same-day shipping. So we said, hey, how do we take this problem and we solve it end-to-end for merchants instead of being a small tool that they use? And then we came up with what was then very unique to financial guarantees. Number one, we told merchants, look, we'll take over the order screening process and tell you to accept the transaction or not. If we make a mistake, we'll pay you back. We'll provide you a chargeback guarantee. And we'll also provide you an approval rate guarantee. And that's really our way of telling merchants, look, you don't need to focus on this. Riskified can do this end to end. You're going to be guaranteed the cost, okay? And we're going to guarantee a higher acceptance rate. So that was really the original thesis and how we started the company.

Got it. Now I remember during the IPO, you guys talked about, I guess, your AI. And AI is very topical these days. Can you talk a little bit about your AI and how you screen for fraud without giving away your secret sauce? Because it was a really cool concept and approach, I remember from then. Maybe you can share that with the audience.

Sure, sure. Let me unpack it in a few minutes. So the type of machine learning we have done from day 1, we call it supervised machine learning, where we work a lot with domain experts to create feature engineering, okay? So to give a few concrete examples to make it a bit more realistic, when we work with an OTA on the flight volume, a feature is something that looks -- that holds analytical value. So we would look at the flight route between London and New York, and we would look at the nationality of the passport making that purchase. And for all the different permutations here, we would assign a risk score. And the risk score would be based on the historical occurrences that we've seen. So we're basically creating a lot of features that have risk value, okay? And our models later on are trained on those transactions. They know the good transactions from the ones that have resulted in a chargeback, and that's somewhat unique to us. Because we provide the chargeback guarantee, merchants need to give us the chargebacks in order to be reimbursed. So in the machine learning scenario, we have source-of-truth tagging that our modeling can train on. So over time, that's created a great environment for us, where we have thousands of features or hundreds of features in each model, and we have a lot of different models. We have different models based on geography. We have different models based on the vertical. So flights might be different than general retail, and they use different features. The models we use in APAC might be different from the ones in Americas or specifically, the model in Brazil holds slightly different data features. So that's the type of environment we run.

Yes. That's good. So competitively, right, I think about there's an in-house solution that a merchant can use, but then there are also some risk-scoring solutions. And just kind of talk about how your solution stacks up, why it's better, where it may be weaker. Anything you can share there in terms of why Riskified -- why I should use Riskified rather than do it internally or use a scoring system.

You should use us because we're going to guarantee you a lower cost than your current cost structure, and we're going to provide you higher acceptance rate at the same time, right? That's consistent across all of our merchants, right? Merchants using Riskified, their bill ends up being lower than their current cost structure for higher acceptance rate. Now what you're saying is true. Usually, when we come to a merchant, day 1, they have a team whether it's 5 to 50 people that are building engines to identify transactions. And this internal team is the one using those scoring tools, those third-party vendors. These are tools that help the team build -- in the case of legacy solutions, to build very hard-coded rules. In the case of the modern scoring solutions, maybe a bit more machine learning in nature. And they score a transaction and decide what to accept. And they have case management, so they can manually review the transactions. Sometimes, merchants -- if you're selling to that team, the team would want to continue to managing that part of the business, right? And that would be one advantage that they have. And the Riskified pitch would be, look, this isn't mission critical for you. To succeed as an online retailer, you need to be doing 1 million things well. Why do you feel that you need to be the best at identifying if an e-commerce transaction is fraudulent or not, right? We can clearly do that better than anyone else because of our network effect, because of our ML platform. We have 500 people in Tel Aviv. Our biggest teams are the data science, analytics and domain teams. They can obviously outperform the smaller, even at the largest enterprises, teams that are trying to solve this.

That was very compelling. Why would somebody say no? Like let's talk about your go-to-market strategy and how -- like that resonates with me at least. Like what's the pushback you get? Walk me through how you sign a client.

And we find that when we talk to the finance folks and some head of payments and risk, that definitely resonates with them as well. I think it's important to remember that when we started the company 10 years ago, no one was doing this and we had 0 volume. We would ask someone, would you ever consider outsourcing this or letting someone else manage this process? And they would say, no, I've never seen this done before, right? Now fast forward to today in just some of the referenceable clients we have, whether it's Wayfair, some of the OTAs, some of the largest fashion brands that we work with, I think it's becoming much more common, okay? People are getting much more comfortable with the value that we provide and the proof points. And realistically, like what are some of the possible challenges that we have? It can sometimes be perceived that we're selling into the risk organization. Even though potentially, we would be replacing them later on, right? And that's more around politics and positioning. How do we make sure that we become -- that we showcase to the fraud manager that a modern fraud manager leverages Riskified to drive better financials for the organization? And it's less about building a large domain with a lot of people.

Got it. Perfect. So let's talk about kind of the addressable market opportunity. And so where you guys play enterprise-level merchants, how big is that market today? How penetrated is it today? And how do you think about kind of longer-term growth outside of what we're seeing macro-wise today? So obviously, it's a tough macro environment. But those things, if you could kind of talk us through.

Look, I view e-commerce, as a whole, and I put travel and services in there as well as probably like above $5 trillion-ish. I think most people would agree with that. I think that what we consider enterprise, which is above $75 million in annual GMV, is 70-plus percent of that market. There might be certain segments within that, that are a bit less relevant to Riskified. But as a generalization, okay, I think that's the market opportunity that we're going after. When you compare it to our current size, our current size is tiny relative to that, right? So I think that there's a lot of opportunity to grow and to capture more GMV. And the way we think about the business, we do think about how do we expand our GMV. And the second thing is we believe that we have a very deep and strategic relationship with enterprise e-commerce clients. When we talk about some of our annual dollar retention rates, they've been at 99%, 98%. So literally, almost no client decides to leave us. And we're consistently asking, hey, we have this great network effect of data and machine learning platform. And we're helping merchants predict, will this result in a fraud chargeback or not? What other predictions can we help them with, right? And that's what's been driving our product platform strategy that we've expanded into, and I'm sure we'll touch on some of that in a bit. But it's also informing how we think about potential M&A opportunities as well.

Yes. We'll hit on all of that in a second. Thinking, I guess, about verticals. So I know last quarter -- the last several quarters, your ticketing, your travel has been super strong. What's driving the growth there? And are there other verticals that maybe you have not gotten into yet that lend themselves to what you're doing that are maybe high fraud or high chargeback? Like what are some of the other verticals that are really attractive to Riskified?

So we are really happy with the traction in tickets and travel, right? It went from being slightly below 20% kind of pre-COVID to almost 1/3 of the business today. That's a combination of, obviously, the massive growth that the industry has had, just the tailwinds from the post-COVID recoveries, but also a lot of new client wins. And we do see that in these industries, the more successful we are, the easier it is to add that incremental merchant, right? You're already working with over 50% of the peer set. At some point, it becomes a disadvantage not to use Riskified and that helps us accelerate the growth there. . At the same time, we're really happy with some of the diversification that we have, right? So we mentioned that in the previous quarter, 8 out of the top 10 merchants that we added were from categories outside of tickets and travel, right? Our newer regions, APAC, EMEA, they're seeing over 40% growth, right? So we're seeing a lot of diversification and growth in some of these categories. And when I think about what's most interesting to me, I feel that food, which includes both groceries and food delivery, when I think about some of the traction within the remittance category, those are all very, very exciting. And even though they're traditionally might be lower fraud or chargeback risk, they actually have a lot of optimization, especially around their policy product and around refunds and returns and appeasements. So I'm really excited about that.

Yes. There are so many things you just covered I want to dig into. So just new products, you launched a chargeback kind of management service. Can you talk about that?

Dispute Resolve.

Dispute Resolve. I would imagine that, that conflicts with your -- or it's separate and apart from your guarantee product. How did that come to be? And what's the opportunity there with that? And any other products that you're excited about?

So Dispute Resolve just started for merchants asking for us to help them with not just their fraud chargebacks. They have chargebacks for other reason codes. The customer said, I never received my package, and they felt there was a false claim, so how can we help them? So there's a process where the merchant goes in front of the bank and submit some information and says, well, you know what, maybe this customer said, someone stole his card or he never received the package. But here's evidence that he did receive the package or that he should not win this case. And it's a form of arbitration. So we both help automate this process, and they tend to do it manually. But we also leverage some of our machine learning capabilities to do it in a smarter way, right? Depending on the reason code of the chargeback, you need to present specific evidence points and that can change based on the issuer that you're in front of. And it's all these different things. We're able to increase the win rates, and that helps drive more profitability for the merchant. Longer term, it also helps drive better margins for us, right, because especially on the fraud side, as we are better at disputing and winning those disputes, it helps improve our bottom line as well. So that's great. And the second product that I'm probably most excited about is the policy product. And actually, just this morning, there was a big Wall Street Journal article about the cost of returns and the inefficiencies there. Our policy product also came from merchant feedback where they said, hey, I have a lot of loss going on in these areas. Customers, they're not filing a chargeback, but they're filing a refund claim or request. And they're saying, hey, you know what, I never received my package, or I received the wrong color or the wrong size. And it turns out, not surprisingly, that even easier than stealing someone's credit card and defrauding him. Based on that, I can just call the retailer and say, I never received my package, and they would refund me, right? So a real instance that we recently had is that for a $10 billion-plus e-comm merchant, they were manually screening all these refund requests, and they were probably rejecting around 2% of them. Using the same -- basically the same technology and data platform that we already have, we started screening those transactions, and we were able to block over 10% of those claims, right? Find that they were linked to abusive entities and customers. And the amazing thing is that the insult rate, the way they measure it, the amount of people calling in and say, hey, why did you block this, the amount of people later on filing a chargeback, did not increase at all, right? So that flows straight to their bottom line. So we think the value there is very significant.

And that's a fairly new product, not well penetrated within your portfolio today. How do you think the adoption of that will kind of play out? How quickly can it become a significant...

So this is the first year where we're actively in market with this product. And we set a target internally at the beginning of the year that 10% of the new revenue for this year, the new revenue sales, will come from this product. So on an absolute basis, it might still be on the smaller side. But when you think about how it's scaling, it's scaling from 0% to 10% within a year, and we feel very good about being on track for this year.

Yes. I know your retention is really high. Two things I kind of wanted to talk about, one, up-selling. And so -- and then I would imagine that because retention is so high, that it's probably easy to up-sell things like you're talking about. So beyond just up-selling additional volume but like these additional services, customers seem to be receptive to that.

Yes. So we're doing 2 things with policy. Number one, we are introducing in newer pitches. It's not just that we're solving the problem for our chargebacks. We're solving chargebacks and abuse, all forms of chargebacks and all these different forms of abuse. And that's helping us both create differentiation within the sales cycle. It's leading to larger initial deal sizes. And we are seeing kind of great traction, especially on the new, the up-sell or the cross-sell motion. But to your point, we also have a great base of existing clients where we can cross-sell some of these capabilities. And the integration is pretty much all there.

Yes. No, that's good. I wanted to talk about, I guess, the balance sheet and M&A. How you're thinking about that? Are there any pockets of technologies and capabilities that you would like to add or are curious about?

Yes. So look, we think that the balance sheet is a strategic asset. It's about $485 million. As of last quarter, we were actually kind of free cash flow -- slightly free cash flow positive. So feel great about that. When we think about the best way to utilize the cash, it's, hey, how can we find products that we can cross-sell into our existing base, okay? We view our strategic e-commerce merchants as a strategic asset. They stay with us, they trust us, they have a great relationship with us. What other decisioning endpoints can we provide them? So that's what we have in mind. And we just need to find the right technology at the right price, something we haven't been able to do up until now. But we are actively looking, and I hope that we'll be able to find something, but we do have a high bar there. Prices have come down. I think there's still some disconnect between private and public expectations. And something that people do ask us a lot about, so it's worth mentioning, our opinion on just the buybacks and given the stock price and everything. So the answer we've shared is that while we are discussing it internally, we've always felt that it's something that's maybe better suited for profitable companies. Now that we're somewhat -- have at least a clear line of visibility into profitability and how long it's going to take and how much cash we need there, we're just trying to understand the dynamics around a more limited float and knowing that some investors are looking to open larger positions, how that would impact. So those are some of the things that we discuss internally around how do you best use the cash.

You touched on another important point. I guess, the path to profitability and, I guess, how that can scale beyond this year. And so maybe update us on the time line for profitability, and then how quickly you can scale, and how are you managing expenses going forward.

So for us, I think it was about 18 months ago, maybe even more, when we were speaking probably during the IPO process, we said, hey, we're going to invest into expanding our global go-to-market teams because we feel that there is great demand in international regions, and we want to have field sales in Latin America, in APAC, and set it up in Japan, and we went ahead and did that. And we also said, we need to invest in some of these product platform capabilities and increase the teams to invest in policy and Dispute Resolve, so we went ahead and did that. . Now once we finish those strategic investments and have them in place, we said, okay, now we're going to pause and we're going to kind of accelerate the march towards profitability and beyond, and really happy with the progress we've made there. So you can see that OpEx has remained relatively flat over the past 2 to 3 quarters. When you think about it year-over-year basis, that's probably kind of gone down by a few million. And we think that there's a lot of leverage in the business model so we can continue to scale revenues definitely this year, probably also beyond. We haven't made final decisions on '24 yet, so keeping OpEx essentially flat or relatively flat while scaling revenues.

Yes. That's good. And speaking of scaling revenues, we kind of talked about it earlier, but there's -- I guess it's like latent potential within your existing customer base. I know in the past, you've kind of framed out what that is. Could you provide us some details on that?

I mean we've historically framed that what we've seen is that e-commerce was growing in, call it, low to mid-double-digit growth, right? And because we're a percentage of a merchant's GMV, we enjoy that growth as well and that our new and up-sell business was growing at a 10% to 15% growth, right, let's say, call it historically. . Now ever since we've shared that historical projection, the e-commerce landscape has not grown at that level. It's grown -- sometimes, tickets and travel has grown a bit less, a bit more, then COVID beneficiaries have had a more challenging time. But what we are seeing in today's environment is that our new, up-sell, cross-sell business is actually performing better than historically. It has. And that's probably a combination of more products, more conversations, more global go-to-market, the current environment where guaranteed cost savings is resonating more. On the other hand, the macro environment or the same-store sales for some of our merchants or the merchants that we might have some larger exposure to are still slightly down on a year-over-year basis, right? So we have some merchants in the home, home improvement, home furnishing category. We don't think there's anything structurally wrong with those clients, but they still have some -- digesting some of the COVID growth. And once it reverts to the normalization, we definitely anticipate that would be a tailwind to the business.

Is there like a rule of thumb in terms of how much volume you think you should get from a merchant at -- or maybe an optimal amount for the merchant that they get the best price and the best performance and stuff like that? Maybe, it varies by vertical. Is there a number or a way to think about that?