DoubleVerify Holdings, Inc. (DV) Earnings Call Transcript & Summary

All right. Well, thank you for joining us. We're happy to have DoubleVerify. Mark and Nicola, thank you so much for stopping by.

Great being here.

Maybe we can kick it off. Mark, we reported 4Q last week. What were some of the highlights coming off of the call? And what are the biggest growth drivers we should think about as we head into '26?

Yes. So we ended the year last year growing around 14% year-over-year, delivered 33% EBITDA margins. So had a very solid year for DV coming off of an expectation. At the beginning of the year of about a 10% growth rate. So we are happy with the yearly drive. Coming out of the quarter, we had some really strong highlights in kind of 3 areas: Social, CTV and then AI solutions. On the social side, we saw social activation grow 60% year-over-year, which is a great growth rate for kind of our emerging solutions there on Meta pre-bid and Authentic AdVantage. We saw our CTV volume impressions grow at 22% for the quarter and ended up around 33% for the year. So we continue to expand into connected television. And then we had some really strong AI offerings launched into the marketplace. So our agent ID solution, some of the areas around combating AI swaps. So a strong quarter in product development, a strong quarter in delivering kind of solid growth across the 3 kind of emerging growth lines for us.

And we'll dig into each one of those as we go through here. But maybe taking a step back and looking at just like the broader advertising environment, and there's this narrative that's going on that even brand advertisers are shifting to increasingly performance advertisers. There's not as much brand advertising anymore. It's brand performance. Just how do you think about that impacting brand safety specifically? Are brands not as focused on brand safety anymore as they focused on pure performance?

Yes. I think it's an interesting narrative because there's this concept that brands never cared about performance. And all advertising is driven through the fact that you need to sell something, right? And building a brand is about selling more solutions -- or selling more product. So the idea of brand performance, I think, has always been out there. And it is an increasing focus just due to the fact that the tools to track that performance have gotten better and better overtime. With regard to kind of where does brand safety and suitability fit in, it's always been part of the equation. And just simply put, if you take garbage out of the system, if you take ad fraud, if you take impressions that are not relevant because they're geographically misaligned, if you take brand suitability or safety violations, that's garbage that you're pulling out of the system. What's left performs better. So we've always thought of ourselves as a performance player no matter what, because eliminating bad stuff helps the good stuff perform better. So brand safety has always had a role. Brand suitability has always had a role. And more interestingly, we've also leaned into solutions like Scibids, which takes that data and makes it even more powerful. So beyond just taking garbage out, what you have left, how do you make that perform better? That's where Scibids comes into play. And Scibids is an algorithmic bidding solution that we bought a couple of years ago that we're now integrating across our platform and has had significant uptake with our top 100 customers. I think over 50 of our top 100 customers now are using some level of Scibids in their engagements with us.

And then maybe another high-level question is it just I think with everything that's going on with AI, I think people are focusing on what are the structural assets that ad tech companies have that differentiate them from the competition. Can you maybe just take a step back and talk about DV and what are the assets that you have that really sets you up for structural differentiation?

Yes. So I think the core asset that we have that differentiates us is data, right? If you think about what drives an advertising decision, there are tools out there that make the decision, and there are people out there who inform the decision I think where AI disruption is going to be greatest is in the tools that make the decision, right, whether it's an agent -- it used to be a DSP in the future, it will be an agent, right? The data though is imperative to what that agent -- the decision that the agent makes or what that AI runs off of. That's where we have an incredible proprietary asset. And it's data based on trillions of proprietary transactions that we see every year from our customer base that no one else has access to. But more importantly, that's based on proprietary integrations into leading social, CTV and transactional platforms that are relatively limited. So if you look at someone like Meta, they work with 3 companies and brand suitability that have access to transaction data that you can make a decision on. If you look at Netflix, it's the same thing. It's 2 brand suitability or safety companies. Those integrations are contracted and they're limited. No matter how good your AI crawler is, it's not going to get into Meta. No matter how good it is, it's not going to get into Netflix. So the fact that we have these proprietary relationships that are based on our independence and unbiased and unconflicted perspective, give us a data pool, which acts as a really significant differentiator. And that data pool only grows over time and the number of transaction only grows over time. So it creates this really wonderful flywheel where we get proprietary important data that allows us to make better decisions. Those better decisions help us close new clients. Those new clients give us access to more data, and it spins over and over again. So we have a data engine that just gets smarter over time based on proprietary information.

That's helpful. And then maybe let's dig into some of these catalysts that are coming to 2026. And specifically, Meta activation is one of the big ones. Can you just talk about what you're seeing so far to date? And I think one of the big things you talked about before, Mark, was getting to the annual budgeting cycle and getting this into some of those contracts. Can you just talk about how we should see this progressing throughout 2026?

Yes. So we saw a really nice acceleration in the Meta pre-bid or activation part of our business running into Q4. We grew from 56 customers to 68 customers A large number of those, 28, I think, are top 100 customers of ours. So we're starting to scale with the kind of companies that matter with Colgate, with Unilever, with Fidelity, with Sony. These are big brands that spend a lot of money on Meta, and they're now starting to implement the tool. A catalyst for that was the fact that we launched in GA, but it was relatively alpha-based GA in early -- in Q1 of last year. That product has gotten better overtime. Data transfer from Meta to DV has gotten significantly faster. The number of categories that we're able to filter against has gotten significantly larger. And that's meant that the product efficacy has gotten better. That allows us to sell new customers because they run it -- they only have -- they not only have a budgeting cycle, but they have a testing cycle they run through. So they want to run the filter, run impressions across Meta with the filter and without the filter and see what their suitability rates are, for example. And we're seeing significant lifts in suitability after you run the filter. That continues to grow as the product gets better. So we've seen that scale over time. We mentioned we exited the year around an $8 million run rate on the solution. And we ended the year at an even stronger run rate in 2026. So we look at that solution if we start to scale it. We do around $40 million of measurement with Meta on just verification measurement. This is a solution that costs roughly twice as much. So even if we sell half of our customers through on Meta pre-bid at twice the price, you're looking at a $40 million business opportunity for us in the relative short term. So we're scaling there nicely. The uptake has really started to pick up. We've got some big brands against it, and we're very optimistic that that's going to be a real impact driver for '26 and beyond.

And then the other big one is DV Authentic AdVantage. Can you just talk about maybe step back and inform investors exactly what Authentic AdVantage is? And then two, what are you seeing as far as adoption and then also performance outcomes?

Yes. So I know for those of you who don't know a lot about DV, we talk about a lot of products, right? So we do have a lot of stuff we throw out there. But just to take a step back, the main goal for everything that we do is to ensure that ad impressions perform and that ad spend is protected from bad behavior, from fraud or brand safety or suitability violations and lack of geo alignment. We do that either before an impression is bought, which we call pre-bid, so filtering or we measure it or block it after it's purchased. There's one other aspect of this tool where we started this conversation, which is around performance. And performance is about optimizing what happens before you buy based on data that you have after you buy. That's where Scibids comes in. And Authentic AdVantage is a solution we launched last year, first on YouTube, which takes all 3 of those aspects. So pre-bid filtering, post-bid measurement and optimization and puts them together into one package. We launched that across YouTube in mid last year with some beta customers. We've now continued to scale that. We mentioned about $8 million in ACV have been closed running into the end of the year. And that business continues to grow really nicely. The value prop there is pretty straightforward. We are able to drive 20% to 30% higher reach, 20% to 30% lower cost per unit while increasing brand safety and suitability at the same time. So it's kind of like the Holy Trinity for an advertiser, give me more reach that's cheaper, but give me quality that's better. And that's what Authentic AdVantage does. And we've launched it across YouTube, and we have plans in the midterm to roll it out to some other platforms as well.

Got it. Okay. That was my next question. [ I wouldn't expect ] any additional platforms. But it sounds like they're coming, but it just the takes time.

Yes. So we're testing it on Meta now. Still very early days. We're testing on Meta. And we also -- now that we have prebid on TikTok as well, we think there's some opportunities to roll into there. But our main focus right now is expanding our Authentic AdVantage implementations on YouTube. We've got customers that spend $1 billion a year on YouTube, $1 billion. And if we can just scratch the surface of that $1 billion, and this is a percentage of media product. So when you buy Authentic AdVantage, you usually pay against the percentage of media. There's really significant upside for us. At our Investor Day that we had last year, we saw this as a $100 million opportunity down the road. And I think it's early days, but we're seeing really great uptake. We're seeing great results and some of our biggest customers are trialing it out. So more to come there for sure.

And then CTV, this is another key catalyst for you. I think some of the product launches or [ do-not air list ], the Authentic streaming solution that you launched. Just how should we think about those products scaling in 2026 and beyond? Where are we, I guess, in like an innings perspective on CTV?

Yes. I'd say we're solid second inning there. We're a little bit out of the first inning because we have a decent CTV business. As I mentioned earlier, it grew over 30% last year. In the last quarter, it grew 22%. But I think we're still underpenetrated on CTV to be direct. Part of it has to do with the fact that some -- how CTV is bought, but also it has to do with the kind of data that we've been able to get to make our decisioning based on what advertisers are really starting to get concerned about, which is greater transparency in CTV. Both of those things are changing. So through our relationship with IMDB, we're starting to now pull and create program level analysis on CTV impressions. And we've got significant scale across the CTV platforms as well. Both of those things together have helped us launch things like Authentic streaming TV, which allows an advertiser to basically have the confidence that they're buying real branded CTV impressions. I know it sounds nonsensical, but in many, many cases, advertisers think they're buying CTV and they're buying a gaming app someplace or something else. And we've seen significant amounts of CTV fraud out there. With Authentic streaming TV, both on a pre-bid focus and a post-bid, advertisers are able to measure and ensure that they've gotten an authentic streaming platform and filter out those impressions that don't that don't align with that. The ABS list right now, it's a relatively small program that we've rolled out only on Trade Desk. But the goal there is our attach rate for ABS on CTV, and ABS is our authentic brand suitability solution, which is hundreds of millions of dollars of revenue for us, only has an attach rate about 10% on CTV impressions, right? So we have a huge opportunity for ABS on CTV, but we've got to find the right fit. And I think with [ do not air list ] through ABS, we've got the right fit. What the tool does is it allows advertisers to literally exclude programs through an agentic interface, so they have -- we have an agentic-based interface, exclude programs that they do not want their programs that they don't want their ads to run on. This is something that's been done in linear TV since started. Don't run me on this kind of program and show. We've now moved this into digital in the agentic age by building this solution. The goal there is to obviously drive attach rates up and also drive the CPMs up as well. Because on CTV impressions, I don't think we're getting the value from our buyers of what that product is worth. We now have some new tools to start charging a higher rate for ABS, for CTV because I think now it has real value against CTV impressions.

That's great. And then maybe one that's more of in the distant future, not distant, but maybe near distant future, is just verification on AI platforms. Obviously, it's an early concept. It really hasn't -- no one's launched it yet, but it's coming. Just how does DV fit into that environment? And how do you expect it to evolve the company as a whole and opportunities that you guys could get into?

Yes. So Matt, I mean, to be direct, it's evolving way faster than we expected. With ChatGPT now launching an ad-based version much faster than I think we expected or anyone in the industry expected, this is something that's barreling forward. And the interesting thing is that we've seen this story before, and we've played in this story before. Rewind 2 years ago to Netflix, which was never ever, ever, ever going to run ads and then 1 day to decide to run ads. Within the first 60 days, they gave us a call and said, hey, we need verification. We're seeing a similar story play out with ChatGPT, where they were never going to run ads. Now they're running ads, they're launching tests. We have several dozen of our customers as part of that test. So we've been involved with our customers from day 1. And what our customers have told us is -- we expect the same level of verification that we have everywhere else across social, across CTV, across mobile web, across the Open Web on chat environments as well. So that to us is a good sign that we have our customers behind us. I'm sure you saw this week that Criteo announced that they're a third party now is going to allow us to be part of the trial test. So all those indications are pointing towards the fact that [ ChatGP ] is -- wants to play with the broader ad tech ecosystem. It wants to bring in partners that can make a difference for them and that there's tons of customers out there who are leaning in and they're customers of ours that are kind of saying, we want the consistency on ChatGPT that we're getting everywhere else. The other thing that's really interesting for the -- on this front is that from what we've read, most analysts are saying that dollars that will go into chat advertising are dollars that are coming out of search. And that's expected to be somewhere from $25 billion to $30 billion in the next 4 years. That's money that we don't go after right now. DV has 0 coming from search. So we look at this as a total incremental TAM of like $400 billion that we've never tapped into. And then if it rolls in there and it proceeds the way we expect it to, it will be a great new opportunity for us.

Great. And then Nicola, Scibids and Rockerbox were 2 key acquisitions you made. We actually kind of hit on Scibids and the adoption that you're seeing there. But can you just talk about what you're seeing as far as adoption and just how those products are going?

Yes. So maybe we start with why we acquired these companies and what the philosophy was there in terms of M&A. These were 2 companies that were in market with clients and already with integrations within the ecosystem when we acquired them. So we bought businesses that already existed with client relationships where they didn't have the distribution that we have with so much more -- so many more clients. And what it allowed us to do is expand from where we were a few years ago, it was just protection to then adding outcomes and now performance, right? And so you now have a suite of products that no one else really has out in the market, and that allows us to have conversations with clients that other companies cannot have. And so Authentic AdVantage -- the core assumption around Authentic AdVantage really is built on Scibids and Rockerbox will come into the fold as well as we develop products around outcomes and performance. So it's worked out very well in terms of our whole M&A strategy. it is really driving all these product evolution that we've seen in the past few years.

No, that's great. Maybe I can sneak another one in here because I think I think it's important for everybody is just as we think about guidance for 2026 and just -- I know for 1Q, I believe it's a slight acceleration that people are seeing. What can you give investors as far as just comfortability around like hitting the guidance targets and then moving forward, what's embedded in those? And just -- yes, what can you give as far as like is it some of these new product initiatives? What are we seeing in there?

Sure. Yes. So the anchor for our guidance is the core products into the core base of our clients. The average tenure for our top 75, 50 and even 20 clients is almost 9 years. And our NRR last year was 109%. So off that base, that's the baseline for what we expect for 2026, 109% NRR. And then, of course, we've guided to on the higher end, 10%. So you have an extra point there that comes from everything we just is cut. It's going to be a product-led growth from all the adoptions of these new products. The speed at which the products are adopted. It depends a little bit on us, but it also depends on the clients and opening up budgets and adoption of the new solutions. But that's how we're thinking about the guidance. And the only other thing I would say is it is going to be more towards the second half of the year that you'll see the acceleration just because last year in the first half, we had a strong first half. But that's how we think about it. And the upside to the numbers will be the speed at which we can have an adoption of these new products.

No, that's great. Yes, it definitely seems like there's not too much that's embedded into these new products. And if things go well, there is potential. You've operated this business typically at 33% EBITDA margins, at least for the past couple of years. You just guided to 34%. What's changed and what's enabled you to maybe bump up those margin targets?

Sure. So what's not changed is the need for us to continue to invest in the business. So we are a company that invests in the business and this is how we were able to launch all these products that others cannot have in the market. What has changed, though, is that the advent of the AI tools really allows us to create efficiencies within the business. The investments we need to make around AI for us is not the infrastructure of AI, right? It's the usage of their tools to continue to develop our own proprietary databases and our own proprietary classification of content. But that is now at such a speed that we're able to actually extract efficiencies that naturally will fall to the bottom line. So we're able to continue to invest yet show efficiency growth. We expect to be able -- simply said, we expect to be able to grow with fewer people because the tools are going to allow us to do what we were in the past doing with a lot more people.

Yes. No, that's great. And then a $300 million share buyback you announced last quarter, I think the biggest in the company's history, if that's...

Correct.

Can you maybe just take a step back and just talk about capital allocation priorities? Should we expect any more M&A from Scibids and Rockerbox that you bought previously? Like just what are the priorities as you sit here in '26?

Yes. So the 3 pillars of the capital allocation, again, haven't really changed, which is we invest in the business. We look for M&A opportunities like Scibids and Rockerbox, which, by the way, we're able to fund with the cash that we generate and then buybacks, right? So '26 will be the third year that we're doing buybacks. $300 million is the largest available amount for buybacks. What we said is that we will do buybacks that is at a clip that's faster than we did in 2025. For '24 and '25, we were anchoring buybacks to the stock-based comp expense. In '26, we said we're going to do it at a clip that's going to be faster than that. Last year, buybacks allowed us to reduce outstanding shares by 3%. $300 million would create a much bigger reduction in shares if we use it all. So that is the beauty of the business, right? Like the free cash flow conversion, it was 70%. Last year, it was 60% the year before. And so we have a lot of options to use cash, and we're always focused on return to shareholders and buybacks as part of that.

And look, we firmly believe that we've got a great plan in place. We've got huge opportunities ahead and who better to bet on than ourselves. And I think that's a great use of our capital.

Yes. No, definitely. And then Mark, maybe a last one here just because last year, IAS got acquired. And like just can you maybe take a step back and talk about the competitive landscape? Has anything changed since then? Just like as we sit here today, how do you think about the competitive intensity in the...

Yes. I mean it's a great question. And if anything, we feel like the competitive landscape has tilted more in our favor over the last 12 months due to the fact that we just spent 30 minutes talking about all these amazing products that we've invested in. And out of all of those, our competitors have launched none, right? They have no optimization solution. They've got no solutions in social to address kind of AI slop. They've got no solutions like Authentic AdVantage out there that bring together pre-bid, post-bid and kind of performance-based in 1 package. They've got no CTV solutions. And why that's happened is because when you're distracted by acquisitions and not up deploying capital, you can't deploy products into the market. And you could listen to me rant or you can look at our 90% of our wins in Q4 were greenfield wins. And those were wins in which there was no competitive product on the other side of the table that we were displacing. That means we're going into totally new areas with a totally new value prop that our competitors can't meet. So I think in addition to the great uses of capital that we do in investing ourselves, we're also investing in product and tech that our competitors can't do because they're distracted by $1 billion of debt sitting on their balance sheet when we've got 0.

No, definitely. Well, it's setting up for a great year for 2026. Lot of product catalysts. Mark and Nicola, thank you so much for that.

Thanks for having us.

Appreciate it.

Thank you.