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

Okay. Good afternoon, everyone. I am Mark Murphy, Head of Software Research with JPMorgan. Great pleasure to be here with Mark Zagorski, who is CEO of DoubleVerify; and Nicola Allais, who is CFO of the company. So first off, welcome to the conference, and thank you so much. It's a great privilege, and thank you for making the trip.

Of course. Of course. Thanks.

Do you want to give a quick one-liner, Mark, perhaps just to -- for the benefit of anyone out there. And I think people know what DoubleVerify is, but just in case we have anyone that isn't fully familiar.

Sure. DoubleVerify is an ad platform that enables advertisers to determine whether or not their ad spend in the digital media universe is viewable, is brand safe, is aligned to the right geography and is -- the transaction is fraud-free. We basically look at ourselves as an independent arbiter of digital media quality.

Okay. So let me begin with the dynamics behind your Q1. I think it was one of the best positive surprises that we had this entire earnings season. And when I kind of -- when I zoom out a little bit and I think back, revenue had been growing in the 20s, right? During 2023, it dropped to kind of a 10% to 15% range. Last year, you had had this cohort -- you had a cohort of a handful of customers that they themselves were -- they're struggling with their own business. And then now we have this improvement, and again, pretty big positive surprise where you showed 17% growth in Q1. Can I ask you just at a high level, what do you think improved for the business during Q1? And I think we're all sort of wondering why would that have happened. The headlines have obviously been very volatile around the trade and tariffs, but business confidence was presumably in a kind of a worse place, right, in February and March, we would have thought, than where it was in November and December.

As you noted, we had a good quarter in Q1. Some of our premium priced solutions like ABS grew at the best rate that they had since 2023. And across the board, we saw a strong customer adoption of new solutions as well as good scaling of current customers. And I think that those growth levers, so, a, core customers scaling on new products and new customers scaling up faster, is a testament of things that we can control. We can't control the macro. We can't control discussions around tariffs, but we can control adding new customers and adding new products and upselling those products to our customers. We leaned very heavily into both of those things in Q1. We delivered better-than-expected results and the fact that we were able to launch new products with new customers. So we had folks like Kenvue and Microsoft take on some of our biggest and most important solutions last quarter, and scale up current customers across new solutions, so people like Lilly and Novartis who extended their spend. So again, we saw good core growth, good new product growth and good new customer growth, and that helped us deliver a better-than-expected quarter.

You -- in social, you -- that's something to us that it has had fits and starts. It certainly holds a lot of promise. How would you sum up the Q1 performance if we said customers adopting social activation on Meta? What was -- was something -- was there any difference in kind of the lagging indicators, the leading indicators, the vibes you were seeing from customers?

Yes. So think of our business on social having 2 aspects in the same way it does on open web. We have a post-bid or measurement business on social and then a newly emerging activation business on social. We just launched activation solutions on Meta in February, which have gotten really good traction since that launch. We've closed 20 new customers, 8 of them are top 100 customer of ours. So big brand names have started to scale. So we see social as a big opportunity for us. Right now, on the social side, Meta does around $40 million of measurement revenue with us. The Meta activation product that we launched is priced anywhere from 2 to 3x what we charge for measurement. So if you just would say, hey, let's take half of our current customers using Meta and upsell them this current solution, you're looking at a $40 million to $60 million opportunity just for activation and -- which makes Meta a $100 million opportunity for us over the next year or 2. So we look at social as a growth vector for us. One of the reasons why it's been a smaller part of our business than the open web part of our business is that we haven't had the same balance of solutions there. In the open web, we have what we call measurement solutions where we determine the quality of the media after it's been purchased. But we have activation solutions, which filter out impressions that don't align to those quality standards. We just started launching those activation solutions on social now, so we have a balanced portfolio of tools, which I think gives us a lot of runway in social to continue to grow and drive that business.

I missed one comment you made there, Mark. The 20 new accounts you closed on -- that was on Meta social activation...

On Meta activation.

And that was since when?

Since February.

That's since February. Okay.

They still have to scale, and they're still relatively new, but folks like Nike, AARP, others, relatively big brands and big spenders.

And then when you mentioned -- you were talking about, I think, Kenvue and Microsoft, but also more broadly, you were talking about rapid scaling by new customers. What -- to see that in Q1, can you -- it's hard for me, I think, externally to understand what are they scaling more rapidly and why. Can you give any insight into that?

Yes. So some of those customers like Microsoft, Kellogg, we closed mid- to late last year and started them off with a core set of solutions, but we were able to upsell them to premium solutions, and in this case, our Authentic Brand Suitability or ABS solution. So they implemented those solutions and started driving volume through them pretty quickly. Kenvue was a late 2024 close, but we did a ton of legwork to get them up and ready and launched on January 1. So they also adopted the premium solutions and started scaling them faster. So what we were able to do is upsell quicker than we expected, implement quicker, and it was based on a lot of legwork and sales work that allowed us to convert those folks better than we expected.

So it was kind of internal execution more than the external customer [ pacing ]? Or is it kind of dependent on that?

A little bit of both.

Yes. Okay. And then, Nicola, there was also this comment that the business momentum remained steady through April. You had said that on the earnings call. But then you did not elect to pass through the Q1 upside to the full year, but that's been very common across our coverage. I think it's a wise move. Do you think that that is buying you -- does that buy you a little cushion to maybe kind of observe any macroeconomic softness that you haven't seen yet in the back half of the year? I know it's hard to know, but...

We chose not to change guidance, which, as you said, we're not the only company who did that. But we did see consistent outperformance in January, February, March, April and then even now into May. So the choice of not changing guidance is offset by the fact that the trends are actually quite positive in the first quarter going into the second quarter. If you look at what that implies for the second half is growth rates that are pretty low compared to what we've seen in past experience. But as you said, it just buys us time to see how the macro is going to impact the business.

Yes. Okay. Great to hear that in your words. So coming back again and I think trying to zoom out a little on the -- as you know, I'm not really an advertising industry expert, but we -- just when you go through a period like this, you -- we have just so much stock market volatility, the headline -- a new tweet can kind of change the narrative over the weekend or between the morning and the afternoon. And then we did see a bunch of sensitive industries that had preannouncements or clearly had issues, the airlines, retailers, energy, the -- and like I think it's a pretty good bet that you're doing business with all of these industries. What do you think is going to determine whether they pull back on advertising spend this year or kind of stay the course?

Yes, it's a great question. And let me slice into a couple of pieces. The first is the nice part about our business is that we're pretty well diversified across different verticals, okay? So the ones that are expected or we're really worried about getting hit, like auto, for example, is only around 6% of our total revenue. So vertical diversification is a positive, and we lean into that. When it comes into advertising, though, I think we're dealing with advertisers and marketers who've gone through multiple shocks over the last 5 years. So this is not -- nothing new to them. They had COVID. They had supply chain issues. And I think what that's done, it's made them less trigger happy, right? They've become incredibly agile, right, and the fact that they know they can pull spend pretty easily in and out when they need to. But they -- also, they become very performance-oriented. And what they've learned over the last couple of shocks has been, if you don't spend, you don't sell. And every advertiser has become a performance advertiser. And I think the platforms to their credit, the folks like Meta and Google and others have really trained them to feel like if you're not putting a dollar into our platform, you're not going to sell product. And we've shown you the relationship between point A and point B better than I think we've ever seen in the ad industry. So there's not this immediate knee-jerk reaction to say, okay, times are going to be bad, I'm going to pull because I still have to sell product. And if I do need to pull, like there are no longer upfront commitments that are either binary and either spend or not spend. The programmatic universe in which we exist in -- on the activation side of our business is very fluid. You can turn pies on and off on an hourly basis, a daily basis. So interestingly enough, we did not see much fluctuation in that at all. And you would expect like, oh, announcement comes out, programmatic goes down. It's not -- and I think because of the tie between the ability to spend and the measurement against performance. It doesn't mean that ultimately, at some point, if there's a drastic market change and we're into a massive recession that the overall pie may not shrink. But ultimately, advertisers have become, a, a bit more resilient, and two, a bit more focused on performance and spend in connection to performance.

Okay. So what you see in terms of the activities in your dashboard, it's not as volatile as what we see in the stock market for the reasons you described?

No.

The -- do you think then -- I mean, is it fair to say they're going to look -- as they digest all the headlines and whatever each of your brands is expecting for the year, are they going to look at the verification piece as a must-have type of insurance for them this year?

So I mean, one of the things that we've been doing as a company is evolving from this idea of not just protecting spend, but helping spend perform. And a big part of our narrative is when you take garbage out of the system, right, what's left performs better, right? Ads that are seen by bots don't sell products. Ads that are fraudulent don't sell products. Ads that are sent to content that's not suitable or aligned with my brand don't sell products. So, a, we look at the essential nature of our core solution as being focused on pulling garbage out, which makes what's left better. On top of that, we've invested in direct performance solutions like our Scibids AI tool, which is a bidding algorithm that allows advertisers to compress the cost of their media by finding better impressions that meet their quality criteria at a lower price. And then most recently, we bought a company called Rockerbox, which closes the loop, shows what's work. So if you think about how our business has changed in the last 2 years, we went from core media quality, which is getting garbage out of the system, to measuring media quality, finding that quality cheaper through Scibids AI and now determining whether or not that impression actually sold something. So that entire closed loop means that we're approaching customers differently and selling them a value prop that goes beyond just being essential, being desirable because I can actually show what's moving the needle. And to me, as a CMO, that's critically important, what is working, and I need to know what that is.

Okay. And I want to come back to the performance vector, including Scibids and Rockerbox. I want to do that in maybe about 5 minutes. Can I ask you first about -- Nicola, about some of the price and volume trends? When we look at the media transactions measured or which we think of as the volume part of the equation, it grew 19% last year. The measured transaction fee or the price, right, was -- it fell about 4%. And so I think our simplistic view of it is then, well, what happened is the -- you had advertisers that are -- they're scanning a lot more of the impressions. They're doing that through DoubleVerify, and then the unit cost kind of came down a little. Can you unpack that unit pricing trend? Because I think the -- there's a difference in whether there's a scenario where it's mix shift and there's a scenario where it is kind of like like-for-like. And I think we're operating under the notion, it's really the first thing, the former is happening, not the latter. Is that correct? Is that correct, and how do we know?

Yes, it is absolutely that, which is the -- it is -- the change in price is a function of changing product mix and changing geographies, most and foremost. As more volume comes from, say, outside of the United States where the CPMs are generally lower than in the U.S., that commands a lower price for the product. The price per product is remaining stable. It's just that the mix is shifting towards areas that are growing faster for us. Specifically, about 2024, we also had the Moat wins and [indiscernible] I mean, it was a very aggressive, very compressed time line to win those deals as moats shut down in the third quarter. And so those deals were acquired at introductory fees for introductory price -- for introductory product. So it was a basic product that command at a basic price. We're actually already in the motion of upselling those clients to our higher paid products, which will impact the overall fee over time. But the answer to your question, it is a mix and geography product.

And are you able to speak -- should we assume that that general kind of change in direction continues? I know you've got -- it sounds like Moat might be at a downward drag and then somewhere [indiscernible] but the geo mix and the other product mix, should we kind of pencil in the same thing?

Well, if you look back 2 years in '23, fees went up 3%. In '24, they went down 4%. The reason why you could expect rates to go up is because our premium priced products are being upsold into the base, right? So the social activation product, for example, is a premium priced product, 2 to 3x more expensive than the measurement product. So as that gets deeper into our base, you will see a positive impact on price. But again, it's -- the volume that we measured in Q1 went up 22% versus 14% in Q4 '24. There is more volume that we're seeing, and we're able to upsell more products on to the same impression, which is our motion, right? We're growing with volume, and the price -- the effective price is essentially an output of where the growth is coming from.

Yes. It's -- a good thing to think about is [ think of it ] if we bring a customer in at an introductory product, it's layer one. And what we want to do is start stacking products on that same impression, right? So if we bring a whole bunch of new customers in, you're going to see volume go up. And price may actually become slightly depressed because they're coming out at a base rate. But then over time, we're going to sell them ABS, and we're going to sell them activation for social. And then you're going to see that dollar per impression number actually increase as we stack these solutions on top of each other for the same impression.

Okay. Understood. Well said. I want to spend a moment as well on -- even if it's a bit in the rearview mirror, but the idiosyncratic behavior that you saw in this group of 6 customers was happening last year. And then I think more recently, you had -- you have the one that is reducing, I think, its social spend, right, because of their -- they had an increase in the commodity costs.

Spend across the board.

Across-the-board spend. Okay. The -- so there was this isolated behavior. It didn't seem like it was kind of contaminating into other customers. It seemed like it was kind of ring-fenced off. Has that settled down? Or do you think -- and kind of going along with that, is there any risk of it -- in this kind of backdrop where you have the threat of a trade war at some level, we don't know at what level, is there a risk of running into other companies that are seeing something similar?

Yes. I mean, look, we -- those challenges were specific to those customers. It did not spread, which we did not expect them to. And what it really forced us to look at last year was ensuring that we have a broadly diversified top 20 to top 30 group of customers. And we did lean on bringing new customers in to fill that base out, so folks like Microsoft and Kenvue and others, to make sure that we're not overly concentrated. So we respond to that well. But then also to look at ensuring that we're very close with those customers and that we understand where their spend is and that we make ourselves totally essential to them. So we can't control their ad spend, right? If they have challenging times and they need to spend less, we can't control that. What we can control is what we sell, right, and who we sell, so new products, scaling new products with the current customers and bringing new customers in. And we did a really bang-up job of that at the end of last year. So we closed folks like Google and Microsoft and Kenvue and BlackRock and Charter, all new customers that have joined us. And then they're starting to fill in that top 20 so that we're diversified across that group of customers moving forward.

Okay. Understood. Those are some big logos as always. I want to go back to this whole thought around performance. We think back on DoubleVerify as being primarily a verification specialist. So the companies that would think of DoubleVerify and they'd say, okay, brand safety and fraud prevention, keeping it very simplistically. And then you're kind of pushing beyond verification, you're going into performance and analytics. And I think the -- so rather than just kind of measuring what happened, like how do you affect -- and we had talked about this before, it's the seat belt, and then you want to become like the navigation system and that whole analogy. What's the investment significance of it? And how do you operationalize it? Because I don't know to what extent do we want to boil it down to Scibids and Rockerbox versus something that is pretty big and overarching.

Yes. I mean, the thesis behind it is quite large, which is, again, not only protecting spend, but helping be the independent objective arbiter of helping advertisers understand if that spend works and doing the most -- do it in the most efficient way possible. We've done that through a combination of acquisitions. So you mentioned Scibids, which we did a little bit over 1.5 years ago, and Rockerbox, which we just closed recently. And the idea behind that is, again, core -- it all starts with core media quality, getting garbage out of the system. But if we can work with advertisers to make sure what they're buying is -- and the quality they're buying is actually cheaper and most efficient and then ultimately tell them what worked and didn't work, it's a very different value proposition than anybody else in our space because we stay outside of the media transaction. So we still don't compete with the DSPs of the world or the media platforms of the world. But we help all of those solutions perform better. And we answer the 2 questions that CMOs have all the time and that keep them up at night and that risk their job, which is, I want to ensure that my brand is protected because we've invested in it over time. And I want to make sure my CFO does not fire me for wasting ad dollars on things that don't sell products. We can now close the loop on that. So we've done that again inorganically through acquisitions, through 2 acquisitions that, in total, cost us around $200 million and through organic products that we've built and advanced, so things like attention metrics, which show engagement, and through connecting our Scibids tools into our quality tools. So the idea here is it's a cohesive solution set that advertisers look at from end to end and saying, I can protect brand, I can understand performance. And the interesting thing about that is we are now selling them together. So for example, we're able to show a brand the amount of dollars that you push through Scibids are optimization solution. For every dollar you push through, we're going to save you $4 in media cost. That media savings can pay for our quality measurement. It is a win-win situation. We're compressing the cost of media while actually using -- they can use those savings to actually pay us for the measurement. And no one else in our space has that, and it's been a value prop that's been incredibly valuable for us in pitching customers.

So part of what's happening in the background that has always been there is that you're measuring such a volume of, you call it, billable media transactions. That number was 8.3 trillion last year. By the way, we -- you don't usually see numbers in the trillions. And I actually stopped myself when I said I better go make sure that's not millions or billions. So trillions. Can you make that tangible for the audience? Because the -- what is it -- what are the elements that you're measuring? And then can you take that -- can you repurpose the verification data and use that? Can you tie this in with measuring the performance?

Yes. I mean, if you think about the digital ad space itself is -- there's been numbers that -- it's around $1 trillion a year is what people think of digital ad spend, right? We measure across the open web. We measure across social, so people like Meta, TikTok, Pinterest, Snap. We measure across connected television. We measure mobile app and mobile web. So basically, we cover just about all of digital ad spend. And we do that for about half of the top 1,000 brands on the planet, so think Unilever, Microsoft, Colgate, Kenvue. These are huge customers, P&G, right? P&G alone spends $1 billion on YouTube and -- right? So we see tons and tons and tons of transactions. That data that we capture from that transaction, we use to measure quality, right? But we also get signals like viewability. We get signals like device type. All of those different signals can be used to build different products, but also help us determine performance. So those signals can be fed into Scibids if you want to optimize against viewability, you want to optimize against quality. That's one take, and we see those transactions. With Rockerbox, they actually get to see the actual -- not just the ad transaction, but the commercial transaction as well. So for our customer, they will get their sales data that comes in from their website. They will get also the linear TV spend. They will get the transactions that happen off their website. They get everything in that box because they're there to prove what's called view -- what's called multi-touch attribution, which is try to attribute ad spend to actually a sale. So we're going to see even more data over time, which gives us a better view to actually determine what's working and what's not working. So all of this data comes together into not only a contextual view of the ad environment and whether it's safe or not safe, but a performance view as well.

And if you are -- it's nice to see that how you can close the loop and how all that data is potentially synergistic. If you are successful in the long run in kind of growing this performance vision and closing the whole loop like that, what mix of business do you think it can become?

I mean, we've got a couple of layers to get to before we go. Right now, our social business, as we talked about, even on our core verification, is relatively small compared to how big social is in the universe, right? So a, number one, I think our social business can be as big as our open web business, which means multiples the size that it gets. And then when you look at things like performance, we've said Scibids -- we're on target to hit $100 million just with our Scibids business by 2028. Rockerbox is small in itself, but I think looking at overall, its impact on our platform, I think, is going to be sizable. So we've grown the TAM for our business by growing performance because of the fact that we're not just top of the funnel with brand advertising and brand quality, but we're now moving bottom of the funnel. So you're looking at expansion in the billions of potential TAM for us. And I think the short-term growth opportunities, again, are on social, but as well as performance metrics like Scibids and performance tools like Rockerbox.

Okay. So if we step back -- great to have a TAM of an additional $1 billion. We step back and kind of think about the -- I'm sorry, you said billions. When we step back and think about the broader kind of value prop and the differentiation that you've got, and you're constantly mentioning Procter & Gamble and Microsoft and Google and Kellogg's, and on the earnings call, you were talking about Nike, it's kind of a who's who. You are still outgrowing your competitors, right? And your -- you have had a pretty -- you won a big chunk of the Moat business. You're getting competitive displacements at a pretty regular cadence. There's a lot to like there. What do you think is the kind of bigger picture differentiation that is attracting them? Like it is -- I think back to the time of the IPO, we would hear a breadth of offering, we would hear global footprint, we would hear scale. What do you think is getting these deals over the finish line?

I mean, all those things are still really important. So scale begets scale. The more data transactions, so you talked about the trillions that we see, the smarter our engine becomes, right? The more of those big brands we work with, the more we learn because we're just one big learning engine, right? It's the contextual engine to look at transactions and learn more. So that becomes an advantage for us. The more platforms that we work with becomes an advantage because some of these platforms only work with 2 or 3 companies and will never work with more. So they don't want to let lots of start-ups or small companies look at their data and analyze what's going on. So that becomes an advantage to us. The breadth of our offering is now unmatched with bringing in these performance solutions. No one else has an embedded optimization tool like Scibids AI. No other company in our space has a closed-loop attribution solution. So we've got a differentiated platform. We've got more scale. And we're -- and now the solutions that we're providing are providing better results because of the fact that we can tie all of them together. So I really do believe we're in a -- we're in the leadership position because of breadth, scale and overall quality of the results we're able to deliver.

Can you give us a moment worth of discussion on the impact that AI is going to have on the landscape? Because I want to also ask you about the Innovation Day coming up, and the 2 things are probably going to interrelate. But I think similar to what we see in security software, it's kind of creating challenges on both sides of the ledger. You've got AI can create this huge volume. It can auto-generate web content. You've got deep fake -- you've got fake videos out there. You've got AI bots that can mimic human behavior and how are you telling now what is human behavior and what is not. Can you speak to the [indiscernible] not a lot of time, but can you give us a moment on the challenge and the opportunity?

Let me talk about AI in 1 minute and 2 seconds. So it's created a ton of challenges, as you've noted, but we're using AI to address those. So it is an arms race, right? So if AI creates what we call AI slop out there, which is tons of AI-generated content or deep fakes, we're using AI to detect those. If AI is creating fraud, we're using AI tools to actually look for those abnormal patterns that were hard to find before. So we use AI to fight AI. We're also leaning into it for operational efficiencies. We're somewhere up to upwards of 20% to 30% of our code is now AI generated internally, which means less engineer time spending -- doing -- tapping and more thinking. We're using it in our unified content intelligence solution, which allows us to analyze massive amounts of short-form video on TikTok and Reels in a very efficient way. So we're leveraging it for efficiency. It's creating more noise outside, but that just creates a greater demand for our product, too.

Yes. Okay. And then finally, maybe 30 seconds on this. Tejal does a great job with all respects in all of our interactions, she says -- it includes the Innovation Day and highlighting it, making sure we're aware and making sure we're all going to be there. What do you think we should be looking for out of Innovation Day? Are there 1 or 2 topics or themes that you think are going to stand out or that you're most excited about?

Yes. I mean, look, the broader long-term vision of becoming an essential performance platform for advertisers is going to be the key theme. Plus, Nicola here on my left will be talking about interesting financial future for the business as well, right, Nicola?

That's right.

Yes.

Interesting financial future for the [indiscernible]. That's a bit of a...

I'm not going to give you the specific [indiscernible] financial stuff [indiscernible].

We've got something to look forward to. Thank you so much for taking the time to be here. We really appreciate it.

[indiscernible].