All right. Thanks everyone for joining us today. My name is Cris Kennedy. I'm the Research Analyst here at William Blair, that covers the Fintech and payment space. For a list of research disclosures and/or potential conflicts of interest, please visit our website at williamblair.com. Next up is Riskified. This is the first time they attended our conference, so we're pleased to have them here. From the company, we have the Co-Founder and CEO, Eido Gal. Riskified was founded in 2013. They came public in 2021. At the core, they provide fraud management solutions that help their customers grow while managing risk. With that, let me pass it over to Eido.
Thanks. All right. Hey, everyone, Eido, Co-Founder and CEO of Riskified. Some kind of standard disclaimers would be up on our site later on. Just flip through these. All right, so Riskified, like we mentioned, right? We're really great at understanding online identities, who's trying to purchase something or, you know, return something, and using that data to understand who's a great customer, a customer with high lifetime value, versus who's maybe a bad customer, okay? A bad customer, a fraudster, someone using a stolen credit card, someone who's trying to abuse a merchant's policy, someone who wants to order something and then claim, "I never received it," even though he did. Someone who wants to leverage, you know, a 10% off first-time customer discount 100 times, okay?
So we're great at understanding the customer, and through leveraging that understanding, we allow the merchants to make smarter decisions, who to accept and who to turn away. Just some kind of general numbers and statistics to understand who we are. Started in 2013, went public in late 2021. Work with over 50 publicly traded companies, really a big focus on enterprise e-commerce clients. Last year, we did $120 billion in GMV or processed in the system, and we're really proud of the 98% retention rate that we have over the past few years. It's really a very sticky solution, a solution that generates a lot of ROI, and we have a lot of great long-term relationships. So what problems are merchants trying to solve for, right, when they think about risk and fraud?
Maybe some of you are familiar, but maybe not everyone. So just to recap, if I steal your credit card while we're speaking, and I go and order something online, let's say with a Best Buy, you would see it on your statement. You would ask around at home, say: "Hey, did anyone order an iPhone from Best Buy?" Everyone would say: "No, it wasn't, you know, not us." You can call your bank and say: "Hey, someone, you know, just stole my credit card or used it." You get refunded. You're protected as a consumer, but the merchant is liable, right? The money gets taken out of Best Buy's account or any enterprise e-commerce for card-not-present credit card transactions, okay? So merchants are incurring a lot of cost, a lot of loss because of online fraud.
Because of that reason, they're screening every single incoming transaction to understand if it's fraudulent or not, okay? And remember, that's what we can do better, understand who that is. Is that a legitimate customer, or is that really a fraudster stealing your credit card? Now, it's not just stolen financials, it's also around policy abuse, right? Fraudsters have understood that instead of stealing your credit card, they can just call the merchant and say, "Hey, I never received my package. Send me a new one." And it's really hard to understand. Are these like porch pirates, that someone stole their package? Did it really get lost in the mail, or is this an abusive refund and return request, okay? So there's a lot of other areas where merchants are experiencing loss, that we help with. So how have merchants been, you know, managing this process before Riskified? Right.
So they would have large internal teams. These internal teams would be using a plethora of different tools and systems. Some of these are data systems to augment their internal decisioning. Some of these are platforms that allow them to build the rules and logic, okay? But, but the basic premise was that merchants were building and managing this stack internally. When we started the company, when we started Riskified, we said: "You know, we don't think that's their core competency." This is something that's a combination of machine learning, big data, and network effects, a lot of, you know, kind of feature engineering, and scale, and we think we can do it better than any single internal merchant can, okay? And we think we can be better at identifying these customers, understanding if it's fraudulent or not, and making these decisions.
What we came up with when we started was a unique business model. We told our merchants: "Look, we're gonna take over the decisioning flow, and we're gonna decide who to approve and who to decline because we're better at it. We have a better network effect, a better engineering capabilities here. But we'll provide you, the merchant, with a guarantee if we make a mistake. That's our chargeback guarantee. That way, you can kind of let go, trust us to manage that part of your, your business, and we'll also provide you with an approval rate guarantee." All right? So basically, the value to the merchants is very simple. A merchant would say: "Look, my cost of managing fraud is 20 basis points, and for that 20 basis points, I'm approving 96% of transactions." We would come in, and we would say: "Hey, that's great.
You only need to pay us 15 basis points, and we would approve 97%." All right? So you get pretty... You get great ROI in a guaranteed way. It's a very simple ROI and sale in that sense, okay? So that's the process of Riskified, and that's the guarantee that we provide our merchants. How are we able to do that? What's unique about our technology that allows us to provide that guarantee and have that, you know, kind of higher performance? So just reading kind of from left to right here, we start with the labeling. The labeling is something internal that Riskified owns, that is proprietary to us. What we run is called supervised machine learning models, and the way we actually started the company, is that we would get transactions that other systems flagged as high risk. Merchants would send us their declined transactions.
That was our first pitch, the first two years. We would come to merchants and tell them: "Look, you already have an existing risk system. That risk system is turning away, you know, 3%, 4%, 5% of your transactions because it says they might be fraudulent, and you don't want to get a chargeback. Give those transactions to us, okay? And we'll only charge you if we approve and guarantee that transaction. There's no risk to you because we, you know, we'd pay you back if we make a mistake." So we would get these very, very dangerous transactions that everyone else thought was bad, and we would need to find a way to make sense of them, and to understand why they're good transactions and how to approve them. There's a lot of mismatches, some risky indicators, but find a way to alleviate and remove those mismatches.
And we had to be confident enough in our decisions because we had the guarantee, right? So that's how we started the company, and that's how we built the automation. So we have the most precise labeling and tagged dataset on which to train our models, and that's a big part of why they're the most accurate in our opinion. Today, the network that we have, those 50 publicly traded companies that we work with, they each enjoy the network that we have, right? If we see an abuser in one store, that propagates into our data and features that are used by another store, so that network effect is very, very powerful. The integration that we have brings us into more parts of the business than any other competing payment solution or payment gateway or payment network, right?
We also have, you know, kind of SKU-level data from the e-commerce merchant, together with shipping and reshipping and customer support data, and obviously, anything around the transaction. So it's a much deeper set of data per transaction than other players in this ecosystem, which is important. And things more around performance management is because we have the guarantee, it's not enough for us to just say: "Hey, you know, there was a fraud ring, you know, past two weeks, and now you're exposed to so and so losses." We need to continuously stop fraud the second it happens, right? And that's the way fraud behaves. They look for a loophole. Once they find a loophole, they go at it again and again and again until someone finds that loophole and plugs it in the system.
And anything within our network, we're continuously monitoring in a way that's very hard for individual merchants to do. And once we find an issue, any place in our network, any type of fraud vector and attack, we can close it for everyone else within our network. So we do believe that our network is probably the most--the safest and the most protected, right? And again, continuously innovating, most big portion of the company is R&D, is AI and machine learning, data scientists, has been and continues to be the biggest focus for us.
Some of the flywheel or the impact that you can see is as we add more merchants, right, we have kind of more GMV, more data, more consumers in the system, allows us to become more accurate, which in turn allows us to create, you know, more compelling ROI for our clients. And we think that's only increasing. Yeah, so some of these numbers, just, you know, 3 billion life cycle e-commerce transactions with full tagging and order-level data. We think that's very unique. So we're proud of kind of the 120 billion in GMV that we've done in 2023. That's obviously still a tiny portion of e-commerce today, or, you know, where we see e-commerce heading.
On our recent earnings, we shared that, you know, about a third of the business is in kind of the different industries of fashion, a third of the business in ticketing and live events, and another third around, you know, kind of electronics, general retail, delivery services, payments and remittance. But there's still a lot of opportunity, both in the core markets and in newer markets that we don't have a significant presence in, to increase. And we believe that as we add more functionality to our platform, and, you know, we started as just solving the problem of fraudulent chargebacks, we expanded into solving the problem of different forms of policy abuse, to helping merchants manage their representment cycle.
We believe we'll continue to layer in additional products that are focused on helping enterprise e-commerce merchants manage their business better, become more profitable, understand their customers better. We think we'll be able to expand the GMV that's really available to us, and also some of the take rate dynamics. I think this is really great to show just the wide range, and obviously, we're not allowed to name everyone we work with, but these are just some of the names that we have approval to use. You can see that there's a wide range of categories, industries, geographies, low-risk merchants, high-risk merchants. Really, the common theme here is that we can always outperform an internal team, okay? It doesn't matter if you're, let's just say, a Wayfair, right? Which could be considered lower risk, mostly domestic U.S. transactions, okay?
But then we could also work with SSENSE, which is very international, high risk, very high luxury. Our pricing is risk-based, okay? So the way that we price is, for example, in the way, for example, they would say: "Hey, we're a relatively safe merchant. Our chargeback rate is 15 basis points, and our approval rate is 97%-98%." And we'd say: "That's great. You know, we can still provide you savings to that cost and incremental uplift to that approval rate." But whereas the SSENSE example, they would say: "Hey, we're, you know, a higher-risk global fashion brand with luxury. Our chargeback rate is..." And again, made-up number, "60 basis points, and our approval rate is only 90%," right? So the pricing would be based on that.
So it's risk-based pricing, which means it's relevant for lower risk and higher risk merchants across categories and geographies, with the common theme being we can outperform an internal team. When I say an internal team, okay, that internal team is using third-party tools and services and data. It's not Riskified or competitor or internal team. It's Riskified or internal team using these competitors. I think that's just an important distinction. So just a bit about kinda summary financials. $329 million, midpoint of guide for kinda 2024. 50+ accounts generating over $1 million in revenue. That's a 2x increase in this metric since 2020. Over 53% gross margin. OpEx has been declining for the past, I wanna say, yeah, two, three years.
So we've continuously shown a lot of leverage in the business model, and have marched, kinda shown really great improvements towards profitability. And we think together with the high retention rates, this is a very attractive long-term opportunity. Just to highlight some of the different industries and geographies that we work with. So I think I mentioned that fashion and kinda tickets and travel is about a third, respectively, each of the business, and you can see the different breakdowns across electronics, home, general, payments, and money transfer. The U.S. continues to be the largest geography for us, but we're seeing the fastest growth in APAC and other Americas, predominantly Latin America. There might be some... No, it's just the color here. And chargebacks. The biggest impact on our gross margin, why they're kinda lower than industry standard SaaS margins, is the chargeback component, right?
The amount that we've made a mistake that we need to pay out to merchants. That's something that we model a lot internally, have a lot of control over. We're talking about we're not doing a single transaction that's, you know, $50-$100 million and guaranteeing that. We're doing millions, if not billions, of smaller transactions, and within a few days to weeks, we start seeing the chargebacks come in, and we can always see if they're coming in based on the graph that we expected them to return, or slightly lower, slightly above, and then make adjustments across our portfolio of merchants. That's how we manage it as a business.
And being a machine learning company and having both, you know, better engineering within our machine learning platform, more data as we onboard more clients and more experience over time, what we see is that our CTB, chargeback to billings, or the amount of chargebacks to our revenue, reduces over time. And I love this chart because you can see it, that it pretty consistently happens with every single cohort, right? The great reduction. The offsets to this are the addition of new merchants and going into newer categories, which tend to start at a higher chargeback rate just from day one. Yeah, so really pleased with the improvements. You can see here just the quarterly improvements since IPO we've made towards adjusted EBITDA.
I think part of the storyline we've been sharing, around the IPO, right, around the IPO, we said, "Hey, there's a lot of investments we wanna make to build our global go-to-market teams to go after the opportunity in APAC and Latin America. There's a lot of investments we wanna make in order to go after the platform opportunity to build out our policy product, our dispute product, and we're probably going into an investment cycle." We've also communicated that we feel that this investment cycle, as of last year, has kinda finished, and we're seeing the positive results of that. We're seeing the traction in the newer product, we're seeing the growth in the newer geographies, and we're seeing kinda the leverage that comes with that. And we're really pleased with the results, especially over the past two quarters.
When we think about the free cash flow, you know, kind of even more significant. So again, this is a company that has over $450 million on our balance sheet. A significant part of our current valuation is our cash. Free cash flow, you know, we've already shared that it should be above $30 million this year. So we feel it's great and significant and only increasing. And how do we think about the growth opportunity longer term, right? So number one, you know, we're tied to e-commerce. That, that's usually a positive. There have been, you know, some periods post-COVID where it was slightly a negative, but longer term, obviously, we're big believers that more and more volume is moving online, and we'll continue to benefit from that, that same-store sales growth.
We add, continuously add more merchants, right, in our existing categories into newer categories, into newer geographies. I think it makes sense for people to understand why we're more accurate than internal solutions at these e-commerce companies and why the ROI is positive, and over time, that results in more client wins. Because we start to work with large enterprises, we don't always work, or even traditionally, we don't work with all of their volume on day one. They would usually test us on a smaller segment. They're international, they're high risk, and over time, we expand and capture more of their volume for a lower risk-adjusted fee. Okay, so that's also another growth driver for us. We're continuing to see, you know, a lot of value in the geographic expansion and going into newer categories.
You know, I think a year or two ago, we would not be talking about payments and remittance and delivery services, and now I'm happy to share that we can. And just the platform sales approach, going from just the chargeback guarantee product to the policy product and dispute product, is both helping the chargeback guarantee sale process, but it's also great cross-sells, and we've actually started to see a few new lands just on these newer products. So in Q1, we already shared that we had our first standalone sale for our dispute and policy product, which was very exciting for us. And the way that happened is we, we approached a client with a platform sales approach, and they said, "Hey, that's great.
You know, I actually- I'm locked in for the next, you know, nine months on, on the risk management side, the chargeback side, but I can really use help on the policy side. That's my biggest issue. Let's integrate, you know, the solution there." And we definitely think that's, that's great, full stop, but it's also definitely a great cross-sell opportunity over the next few quarters for the core chargeback product. So really exciting things happening there. And that's it. I ran through it a bit quicker than I should have, so sorry. But I guess we have time for Q&A if-
Yeah. Please don't be shy. Feel free to ask a question. I'll just start, I guess. Chargeback Guarantee is your core product, and you have these newer-
Yeah
products. Can you just talk about the economics of these newer products relative to?
Sure
- Chargeback guarantee?
No, no, that's a great... So let me, let me give a wider answer on the newer products. In Q1, we shared that they contributed about 0.5% ±, to our gross margin. They have traditional SaaS like, you know, kind of gross margins, 80%-90%. So I think you can kind of reverse engineer the revenue rate they're at, and that grew about three times year-over-year. So again, still small numbers, but very new products seeing good demand. And we think that could be meaningful if they continue this type of growth. So that's a bit about their economics, their growth rates. And I would say that Policy Protect has been in the range of 10%-20% of from a size perspective.
It's a chargeback deal with better margin unit economics, and Dispute Resolve is smaller, probably in the 5%-7% range.
Just talk about the risk related to new verticals as you expand into new categories. How do you manage that, and kind of how the economics are?
Yeah. I think we have so much confidence, especially just looking at this chart, that we continuously see that whether it's a new client, a new geography, a new vertical, we're able to improve significantly over time, and I think that we've been able to shorten the time frame to make some of these improvements, right? Stuff that used to take us, you know, a multitude of years, with a lot of the advancements we've made to the ML platform, they actually take even less time. Right? And when you become more accurate, you can do two main things with that. Number one is you can increase the acceptance rate for our merchants, right? Because you're now more accurate, so for the same chargeback rate, they get more approvals.
Or you can, you know, kind of decrease your cost ratio, the amount of chargebacks that we incur. And we try to, to balance the, both of those from a business perspective, to, to maintain both, you know, kind of the client satisfaction long term, but also kind of those improving and healthier margins for us so that we can reinvest and build more services for them. And as we think about, you know, going after some of these newer regions, we find that it's really the similar, if not the same, models and data, more with, you know, some slight adjustments than anything brand new. So we feel good about the ability to service it.
Yes.
So a few things inherent in that question that I wanna kinda respectfully push back on. This isn't insurance—balance sheet insurance for a large merchant that's afraid of having, like, fraud liability and wants to move away from this. This is the ability to look at a transaction, okay, and detect if it's fraudulent or not, and based on detection, you know, and make that decision. Right now, in order for a merchant to trust us and to say, "You know what? I don't wanna have a team of 10 data scientists, 15 engineers, 30 managers, continuously manage this process." Fraud never goes away. It only shifts and continues to increase.
And I believe," this is the merchant speaking, "if I'm gonna continue to invest in this, like, $10 million a year, I'm gonna get a 95% approval rate, and my cost is 20 basis points." Okay, at what point do you say, "You know what? This isn't mission critical." Okay? Is that, like, the core competency of whatever enterprise e-commerce, name your thing, is that what they should be focusing on? Or if you can guarantee them results at a lower cost, at a higher acceptance rate, would they say, "I don't wanna manage this part of my business?" Right? I think that inherently, the network effect, the ML platform that Riskified has, right, enables them to do a better job. The fact that they're guaranteeing the results means that I can actually trust them. That's all it is. It's about the trust, right?
'Cause if I was not guaranteeing the results, and I would come back to them and say, "Hey, there was, you know, an issue over the weekend. You just lost $100 million because of a fraud ring I didn't catch," they'd say, "Okay, thank you. Goodbye." But they would also need to keep that team of 10 data scientists and 20 engineers, and they couldn't really let go of that process. So we don't believe that there's a clear cap over the size of merchants. And you can see that that's the focus, right? Like, those 50 publicly traded companies, what they've said is: "I don't want to be in this business. Okay, I think that I can get better economics, and also focus on the areas that are more important to me by leveraging Riskified.
Yeah, two questions. I guess, competitive threats, and Visa, Mastercard are still alive in this space, and I know there's a lot of fraud, and the fraudsters are really good, but is there? You know, what would you? Would you view Visa, Mastercard as competitive threats? If not, where, where are the competitive threats? And then maybe the product roadmap, where do you see opportunities separately?
I don't view them as competitive. I mean, it's always confusing because people are like: "Wait, Visa, Mastercard, aren't they into this space?" They're into so many different things, like the issuer authorization and a lot of the different risk systems on their end. But when you really think about the merchant risk systems that the e-commerce merchant is using in order to make a decision to accept this transaction or not, they have very limited capabilities. They have a very limited data set relative to what we use to make the decision. And I think, like, the best proof is that the actual competition, when we come up against within merchants, is not that, right?
It's some of the newer generation players, whether that's like, Sift or Signifyd or some of the others, but they're some of the systems that internal teams use, and that's really what we're up against a lot of the time, right? Saying, "Hey, you can either continue this kind of the way you've always been doing things with different tools or this different way of approaching it, leveraging Riskified." And that's kind of the biggest single thing that we need to convince in the sales cycle. And then on the product roadmap, look, we think we have a great ML platform. We think we have a great network and deep relationship with really, like, amazing blue-chip names in e-commerce and amazing retention rates. To us, the question is: what other services can we build and offer them in addition?
You know, we started just chargeback, and now it's policy, and then it's dispute. To help them optimize their relationship with their clients, understand who the-- give a better experience to their best clients, block the bad ones out, and I think, you know, we'll continue to release in that vein. Sure.
Yeah. Are there opportunities that we provide with, Shopify or Etsy or other large platforms?
I think that's a great point 'cause we talk a lot about, like, the direct relationship with the end enterprise clients, and definitely, as we think about servicing the SMB, that's not something we can do, and you would want to work through the intermediaries, whether it's like a Shopify, an Etsy, to be an infrastructure risk. That's definitely the approach we employ when we think about the SMB side of things, right? How can we be the infrastructure risk for some of these other... whether it's e-commerce platforms, whether it's like PayFac or payment gateways. We don't do anything significant with them today.
The guarantee part treated as a contingency for cash or accounting.
Contingency for cash on the accounting? Not... No.
No, it's not. It's recognized right at the point of our sale, and the chargeback is actually what gets kind of what fall over to you.
Well, I understand. You don't guarantee the results for your company, the merchants. You don't say, "Okay, there's 3-4% connections, our system is so good. We take the risk, not take the risk to say, 'We help you, but we don't take the risk if it's if you are wrong on your systems.'
No, we do. If we make a mistake. So the way to think about this is that, let's say, we would bill a client based on the amount we approved and the take rate we're supposed to take, let's say $100, but for that given period of time, $30 came back as chargebacks, mistakes that, transactions that we mistakenly approved. So we would actually invoice the client $70, right? Net of those chargebacks. And that, that's how predominantly it works. The invoice becomes smaller by the amount of the chargebacks.
Like a 2018 number. What is happening in your PNL as that gets better or better? Is that, like, those markets are ticking up? Is that how that works out?
Correct.
As you roll over, I don't know, do you have, like, multi-year contracts? Do you guys have...
We have anywhere from 1 to 3-year contracts.
What, what's happening to that basis point of the take rate?
It can be some combination, right? It can be, "Hey, look, you're actually at 93. We think you can be at 95 for a slightly higher fee." Or, you know, "Let's package some of these more platform e-sales around policy and dispute, you know, kind of together with incremental discounts for long-term contracts." But that's part of the discussion.