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

Welcome, everybody. I'm Ron Josey, I cover the Internet sector here at Citi. Always happy to share the stage with Mark Zagorski and Nicola. Mark Zagorski, CEO of DoubleVerify; Nicola Allais, DoubleVerify CFO. For those that don't know, DoubleVerify verifies about 300 billion data transactions and 60 years of video content daily, I think, is the stat. So to say DV has scale, I think, is a little bit of an understatement. And exposed to multiple tailwinds, right? So social, short-form video, CTV, retail media, in addition to everything, online advertising. So we have a lot to talk about. So Mark, Nicola, welcome to the Citi's Global Tech Conference. Thank you for coming.

Excellent. Thanks for having us.

So let's kick this off here. And so I want to talk about macro, and then we'll get into DV. I mean, Mark, we've known each other for might be a decade-plus now. We've gone through several advertising cycles, ups, downs, sideways, et cetera. Talk to us how you view the current macro environment, and ideally compared to any product. Is it another cycle in the past? Or is it something new?

Yes, I mean, I think the year started off with a lot of consternation that we were going to have a pretty rough year, right? And I think, if you would ask me this question in January, I think some people did, we saw the macro is looking a little bit rough. The cloud is pretty much separated, I would say. By the end of Q1, we saw some stabilization. And now, I think, we're heading towards a new normal, which is a pretty balanced ad environment. I think, marketers are not as concerned about a massive recession impacting their budgets. And I think, one aspect of this new normal, though, which has occurred is there's definitely a shorter-term thinking than we've seen in the past. So a lot of advertisers are still making quarter-to-quarter decisions than we've seen in the past as opposed to kind of a longer-term commitments. And this was, I think, exemplified a bit if you start hearing rumblings after the upfronts and the new fronts weren't that great. Didn't mean ad spend was going to be down. I think, advertisers are really holding so much in their pockets. And there are so many venues like programmatic media, where they can just deploy as needed.

Quickly.

That's so much -- there's a lot less predictability, I would say. But net-net, macro, I think we've seen a stabilization in the advertising space. And I think, we're in a much better spot than we were at the beginning of the year.

Do you think the stabilization in online specifically, or call it, improvement, if you will, is coming from new formats like social, like attention, like CTV? Or is it just, you know what, we have -- these marketers have greater confidence that budgets aren't going to get -- macro may not fall off a cliff tomorrow?

I mean, there's definitely a combination of FOMO going on, which is I don't want to miss this new thing. I'd better get into short-form video. I better make sure that I have a social strategy in place because I'm going to miss out on this. And that -- and with also the fact that, okay, the sky is not falling. Things are going okay. We need to sell product. And that's where this all comes back to, is that for marketing -- the reason marketing exists is because, it's supposed to help marketers sell things, right? And the initial reaction to every downturn or recessions like, first budget gets cut, marketing -- let's cut marketing, right? And then every CEO realizes, wait a minute, how are we going to sell anything if we can't -- if people don't know about it. So I think, that realization in that cycle always occurs, and we are now in a much better spot than we were at the beginning of the year.

I think that's a great, great way to frame it. And I'd say the same thing to many people, like, no, no, no, you've got to cut marketing. Like, we'll see. You got to have sales. Let's take a step back and talk about DoubleVerify specifically. So tell us about the core model, the value prop. Obviously, it's about verification, and making sure. But tell us just about how you see the business today. How you see it evolving going forward. And just, yes...

Yes. So for those of you who aren't familiar with DV. What we are is a software platform that enables advertisers to protect their media spend and then ensure that it actually performs better over time. We do this by ensuring that their ads are delivered in an environment that is brand-safe or brand suitable that those ads are viewable, and that those ads are delivered in a fraud-free transaction. And we do that both in a pre-campaign basis, so the ability to filter out impressions that don't meet those standards. And post campaign, so we measure and verify afterwards. How our business has progressed over the last several years, is that verification is in our name, right? But just -- I joked around this in the office, like just like Dunkin' Donuts, right? Donuts was in their name, but they call themselves Dunkin' now, right? Verification is in our name because that's what they started doing. But we really -- we're DV. And we start calling ourselves DV because although our core is verification, what we're actually doing when we verify or ensure that a transaction is safe, when you take garbage out of the system, we're really helping advertisers perform better. So we've said we're evolving from this idea of protection to performance, right? And it's probably, more specifically, it's not an evolution from protection to performance, but what we do is protect and perform. And I think the second half of that equation is where we focused a lot lately, which is once we get garbage out, what do we do with what's left? And you mentioned solutions like attention measurement and like contextual targeting. These are things -- these are refinements of using -- starting with a scalpel to take the [indiscernible] take the garbage out, and then using a scalpel to find what's good, and what we need to take out of that, if that works. And that's really how the business has progressed, from one which advertisers has looked at us as insurance, like we're going to keep you safe, to one of which advertisers are looking as a utility partner that actually performed better and drive return on ad spend.

And that's a -- is that a different buyer when you move from insurance to performance? Or is it the same CMO organization?

It's the same CMO, right? We talked about marketers, and what they want to do is drive sales. Ads that show up in hate speech content don't drive sales, right? That hurts brands. Ads that aren't seen don't drive sales, right? So the same CMO who says, I want to protect my brand, is the same CMO who also is telling -- is being told by the CEO that you better sell more product, right? It's the same team, they've always looked at our role in that equation as being protection but protection with an eye on the fact that, I know if I don't protect this thing, I'm not going to be selling a product. So there's got to be performance there, too.

And as we move to -- define attention. Talk to us -- I think was it CES this year, we were talking attention. Maybe that was a big year on stage, and talking about that. Talk to us about what attention is, how you view it. And then, I still want to talk about like the mix shift to performance.

Yes.

I understand it's the same organization of buying. But maybe it's a different way of thinking about DV. Not DoubleVerify, DV in the organization.

Yes. So attention is something that in the media world, in particularly digital media world, is gaining a lot of traction as a proxy for performance. There's a few reasons. One of the biggest is individual identifiers. So cookies and IDs on users are going away, either through legislation or through corporate leverage on devices. The ability to track a user from place to place is going away. That means the ability to actually target performance is being challenged. Those things that advertisers have used for many, many years in the digital space. So the advertisers are looking for new proxies. Attention is a great proxy for driving an outcome. If someone doesn't pay attention to an ad, if they're not engaged with an ad, then that adds ability to drive an outcome, to drive a sale is severely challenged. So we've looked at all the data that we've been accumulating over the last decade. So we look at things like viewability, and we look at context on a page, what's going on, on the page, and what kind of information is there. And we put those together to build an attention metric. And what that metric looks at, it's a crossover of exposure. So how that ad was exposed on a page. And engagement, was their user engagement with that ad? When we look at those two things, we can draw a direct line between high attention, high attention ads, and actual brand lift, sales, and client acquisition. And we've shown this over multiple case studies with major brands. So attention for us is a further extension of driving greater media quality, but also getting us closer to that idea of driving performance as well. And for advertisers, it's starting to become super interesting because as cookies and IDs go away, they need something else to latch on to that can help them understand how they can drive performance, and what types of ad placements are driving performance.

That's something that DoubleVerify is very well positioned to do.

Sure. Absolutely.

Based on what you know. That's super helpful. One of the things -- we'll get into AI here shortly and other new ad formats but one of the things that -- Nicola, maybe you can help on this one. We often talk about just win rates at DoubleVerify, and a lot of them are greenfield opportunities. And the question always is, why? How are win rates getting to 80% -- remaining or at 80%? And these are newer clients. So where are these clients coming from?

Well, the first thing to clarify is, the market is still very underpenetrated. So there is a lot of activity out there that is not yet verified. And so the services are still something that a lot of advertisers are discovering. And we're always surprised by the brand names that come to us and become a new client where they did not have a verification services from a third party, an independent third party, right? They might have been using homegrown solutions, but not a third party like DV. So it's a large market. The way we are achieving those win rates is -- we are -- we're expanding internationally where we did not have a very strong presence for a long time. And that allows us to, basically, be there for the RFP and actually compete for the business. So that's opened up an entirely new category for us. And then the win rates that we quote is also for companies that were using homegrown systems, that all of a sudden decide to use a third party, or for new solutions. So in that number that we quote is also the win rate when we actually launch a new product, and a client that did not use that product is actually going to start using it. So it speaks to the fact that we grow because of new wins in completely greenfield opportunities, but also by upselling our new products to our existing base. And so one example that I'd give is, Authentic Brand Suitability, which is our premium price product. 80% of the growth in last quarter was from existing customers. So they may have been using it, but all of a sudden they're using it on more and more greenfield opportunities. There are more and more geographies. And that product grew over 50% last quarter. So we have many ways of growing, and that's reflected in that one, right?

Super helpful. And is it -- so understood. Competing versus homegrown, so better solution. Understood, markets underpenetrated. So we're still getting there. Anything -- Talk to us a little bit more about how ABS might have the technology behind it, the verification tools.

Yes, yes. So I mean, we talk a lot about ABS because it is the gift that keeps giving in our business. It's a 5-year-old product that grew 50% last quarter. It is a significant percentage of our overall revenue. I think it will do almost what, 1/3 of our revenue this year, almost $200 million. And it's a great product. The core behind that product, it's the evolution of what we call standard verification. So we started off in the programmatic world, basically doing brand safety, which is a standard segment that says, this is safe or not safe, right? Which was valuable because people at certain points want to have that black and white line. ABS evolved into Authentic Brand Suitability, which was every brand has different qualifications. What works for them. And they want to understand and change that across every brand, across every campaign, and then learn from after they filtered out things that weren't suitable, how they want to adjust that, those criteria every time they buy. So think of ABS being taking dynamic data, changing the criteria in each campaign, or even on -- on each campaign for each brand, and adopting that over time based on performance. And this was the first example we talked about how we went from -- going from protection to performance. As our tools started evolving from just saying, yes or no, protect me, to saying, let me pick out some things that work. Let's look at the cycle of pre-bid filtering, postbid measurement, and see what is working as far as filtering is concerned. But also what's working as far as our other KPIs on performance, and adjust those things over time. The tool has been incredibly successful because it's, again, moved us from being in the protection suite to the performance suite. And it shows that, when we have a product that can actually drive real ROI, as well as protect the brand, it will grow and it will grow significantly. It's a product that there is no [ comparative ] product in the market that we believe that differentiates ourselves from our competitors. So it not only drives revenue on its own, it helps us win clients. And that's what -- you asked Nicola before. Previously, our win rate is not based on the fact that we've got more people, or we've got more products. It's that the products that we have are different and better than our competitors.

That's enough said, right? That's -- so let's maybe switch topics a little bit more to a higher level, in terms of like the core drivers. And Mark, I'm just -- I just want the first word that comes to your mind here when I say short-form video, blank.

Money.

Money? I wouldn't have thought money.

It's money for everybody. Look, the social platforms are making money. We are generating money from protecting it. Now if I'd say money, and then in (say caution) right? Because short-form video is the ultimate UGC environment where -- I just read today, something like 1 -- I'll probably get this wrong, but something like 1.8 million hours of video are watched on TikTok in every minute. Every minute of the day, there's 1.8 million hours of video being watched. It's insane. The scale is unbelievable. The scale of new videos that are being uploaded. And if you went to the Meta upfronts this year, you would think that Meta doesn't have any other product except for reels. And I'm not exaggerating, 90% of the upfront, they talked about reels. The reason why the platforms love this is because; a, is incredibly high engagement, right? So consumers who are on this spend tons of time on it; b, it's a great marketing environment, which you can implant a ton of ads into. Now if you think about it, there's so much buzz around CTV. And CTV is the transformation of television, right? But Netflix, when they announced their ad-supported tier, they said we're going to do 4 breaks an hour, with 4 spots an hour. That's 16 impressions, 16 ad impressions in an hour of viewing. That's how much they can monetize, in an hour of a short-form social video engagement. So if you're on TikTok for an hour, there's hundreds, if not thousands, of ad impressions against that. And that's why I talked about money. The idea of the revenue generation this can create is massive. And since we're knee-deep in this space, for us, it's probably one of the fastest-growing areas of coverage. Our TikTok business, we mentioned last quarter, we did more revenue with TikTok last quarter for verification, social verification than we did all of last year, right? So it's a great opportunity. Now that comes with the (of caution and concern) because the content there is UGC, things like brand safety and suitability become issues. And the complexity around that can't be understated when it comes to trying to verify whether something is safe or suitable because you're dealing with video. In one 20-second clip, you can have video, audio, images and text. And each one of those can be totally different, right? You could have a brand-safe image with totally nonbrand safe music or audio on top of it. You can have brand safe, you can have an unbrand safe image with really nice safe image or text or metadata around it. So the ability to analyze that and do so at scale is incredibly challenging. But that's why we get paid, right? And that's the importance, because advertisers want to be there. And we started off talking about FOMO, right? Advertisers know they need to be in short-form video, but there's definite concern of like, I want to make sure I'm safe when I'm there. And I think, that's the role we play. And it's an increasing driver of growth for us and increasing opportunity for advertisers.

One of the things we hear often is the Lo-Fi environment of creating ads for social video, for short-form video, allows brands to just get into the mix right away. But I do hear the caution comment because you need to make sure you are advertising next to brand appropriate content.

Absolutely.

So I was going to ask about advertiser demand. We just talked about that with FOMO. Maybe social overall. So obviously, we just talked about social video, we talked about TikTok and reels and et cetera, et cetera. But I think social measurement like the feed. Talk to us about the feed. I think that's now a newer format for DV. And we would love to hear your thoughts on just social overall, demand for social, and specifically what you're seeing because you're talking to Fortune 1000 companies.

Yes, yes, yes. Look, social growth in our business, last quarter was over 30% growth across the social networks that we work with. And we work with every major platform out there. So Meta, YouTube, Snap, Pinterest, X, TikTok. I mean, you name it social platform, we're engaged with...

You rank order of those by ROI? No, I'm just joking. If you're comfortable, we'd love to hear it.

I can rank them by size. Meta is by far the largest. YouTube, neck-and-neck, close there. And then we've mentioned, for TikTok -- has rallied to third place, really quickly from 0 to third in a very short period of time. The overall social environment continues to be in high demand. And I think that's what finally spurred, you mentioned, News Feed, Meta to open up their News Feed to third-party brand safety and suitability vendors. We're knee deep in the development process there. We're hoping to get our initial products out the door early next year so that advertisers can feel comfortable that there's third-party independent verification of what's going on in the News Feed. It is a big event that was many years in the making. When we IPO-ed over 2 years ago, the question was, well, why aren't you guys in the News Feed. Why aren't you in Meta? When is that going to happen? When is that going to happen? And we always said, "Well, really, it's not up to us at this point." We have the tools to do it. We have the scale to be able to handle News Feed. It's really up to Meta to open it up. And they did. And I think a lot of it had to do with the fact that advertisers were increasingly looking at alternatives like short-form video, like TikTok, who had been incredibly open to third parties to come in. And I feel like the pressure just was too much. And they said, okay, we want to be good players in the space. And I think there is -- the reality of it is, we are partners to the platform. And our entire goal is to make advertisers spend safe and secure and perform. But in conjunction with that, make sure that the platforms there can help attract those advertisers. I think, that's an interesting dynamic in our business that's changed over the last few years is, we were always kind of seen as the interlopers coming in and calling someone's baby ugly. Like, we don't want you in here to tell us our content is bad. And that's changed to platforms, new and emerging platforms coming to us, like TikTok and others who have said, "Hey, we want to make sure that advertisers know that our stuff is great." And putting your [ imprimatur ] on that totally helps us. So we used to do all the door knocking. Now they come and knock on our doors.

What do you think made that change? Like was it the brands -- the larger brands saying, we're not coming unless we have DoubleVerify -- DV. Or is it Meta saying, we need your solution because we're maybe losing dollars because advertisers are not...

I mean, it all starts with the advertisers, right? When they say they're not going to do something at some point, that speaks louder than anything that we can say, or the industry can say. And it starts there. And once one domino falls, then usually all start falling.

Let's stick with new formats, CTV. We talked briefly about that. Just accelerating core driver going forward. I hear you loud and clear on the Netflix comment. Just talk to us about your view on CTV and DV's overall approach and strategy.

Yes. I think CTV is an incredibly important part of our mix. And a big part of our business thesis, and the way we approach our customers is that we want to be able to verify their spend everywhere they go, right? Wherever they spend on social or mobile or display or short-form video or CTV, we need to be there. So I think it's probably more of an important part of the basket of goods we sell than an actual scale driver. And the reason I say that is because, it goes back to not only the number of impressions that you delivered against CTV, but the cost of those impressions. So it's a big market. And CTV is a big market, and important to be part of that. But our business model is a volume-based model, right? So we charge a fixed cost for every impression or every transaction we manage. And there are significantly fewer transactions on CTV than there are in other forms, like short-form video, et cetera. So for us, it's an important part of the mix, and it's growing. I mean, growing double-digit, high double-digit growth rates each year as far as volume. But I think it's still a very small part, [ start over revenue, ] maybe 3% or 4% of our revenue, impression count. So it's still small, growing, important. I think there's some value unlock there, too, that will start coming when we start doing brand safety and suitability on a show level or a program level [ on cases. ]

Super helpful. That makes a lot of sense. I think the basket of goods we sell versus -- and maybe the last one of these new or the big 3, the retail media side. So you have to be where your clients or where the consumers are. I think, DV is working with 12 of the top retail media platforms, 39 retailers. Just talk to us how you see retail media.

Yes. Retail media is a really cool one because the one thing I didn't mention about CTV that is nice and is beautiful, as CTV is bringing money from linear, of which we have no linear business, right? And linear is not a place for digital verification. So that's all new dollars. Retail Media is a lot of new dollars, too, because if you think about who participates in retail media, they're the platforms. So the Best Buys and the Targets and the Walmarts. But then there's all those advertisers. And a lot of those advertisers are taking co-op budgets, merchandising budgets that they used to do on the retail side, and moving it to digital media. So like that's new dollars. We love new dollars coming into digital. The role that we play -- and our retail media business grew over 80% last quarter. So it's actually becoming a real contributor to not only growth, but overall size of the company. The role we play is both as -- on the platform side, so we provide, for example, people like Amazon and others, a base level of brand safety and viewability and invalid traffic protection for their platforms, [ so the media they're sending out. ] But we also provide filtering and pre-bid solutions to people buying into those platforms, so they can ensure that their impressions are viewable and fraud free. That's interesting for us, too, because it attracts an entire different group of customers that we wouldn't go after, which is, I'd say, midsized retailers, small and midsized retailers, who are buying into these retail media networks that we would never go out to directly, right? So these are small OEMs, small retailers who are looking for broader distribution. But they wouldn't be big enough for us to go. So really, retail media networks act as a channel partner for us. So again, new dollars from new type of customer, new dollars from new budgets. And retail media continues to grow for us, it's a great click. and as you noted, we work with Macy's, Best Buy, Target, Walmart, Kroger -- some of the biggest -- Amazon, I forgot about those guys. Amazon, a little retail guy, some of the biggest guys out there, and it's great. There's one other benefit too, which is really cool, which is not only do we work with them on the platform side and with the guys who are buying in, but in many cases, these relationships have come out of the fact that we work with them as advertiser clients. Now totally different groups. The retail media group, for example, within Macy's is different than the advertising group within Macy's. But it doesn't hurt to have that relationship. So we just announced a deal with Uber. Now we have Uber now as both an advertiser. And they're a huge advertiser, digital advertiser, whether they're recruiting for Uber Eats or for new drivers, huge advertisers, so they're advertiser partner of ours. But we're also a platform partner on their new retail media network, which is their journey, right?

Platform partner, meaning verifying.

Yes.

Got it.

Yes. So it's a great place for us to be, as more and more of these advertisers become advertising networks, right?

And advertising is becoming a bigger part of all these marketplaces that are out there. Uber has been very clear about that, and we cover others. Let's -- I want to get to Scibids, and the acquisition in a second. But we're 10, 11 minutes or 10 minutes left, 11 minutes left. We haven't even talked about GenAI. So one of the things that we just hear often, of course, is just -- well, we all hear everything about GenAI. But I'd love your thoughts on the broader just -- where do we go from here? A lot of buzz. We've heard in this conference, just how GenAI has improved, creative has improved, just the ability to really target folks. So we just heard -- just now the ability to go from 0 to 90 languages overnight. Talk to us how you see GenAI impacting the advertising environment today and then going forward.

I'll put my dark brand in hat on now and be the evil AI guy. So let's talk about how AI is creating chaos in the advertising space because it is, right?

[indiscernible] over a year, but...

GenAI is creating more nefarious content. There's a categorization of content that's called MFA, right? Made for advertising content. This is stuff that you've seen out there. And I'm sure you've all clicked on it, which is kind of like the, hey, look at 12 facts about the Brady Bunch that you didn't know, right? I mean there's like 17 million ads on that page. MFA content has been increasing exponentially. There's -- we've seen some studies that say, sites that we're creating maybe 10 or 15 new articles a day, we're now creating 1,000 articles a day on a single website. Now those websites now are being -- and those websites now are proliferating by the tens of thousands. So think of an environment out there that advertisers are trying to manage, and they're buying against their programmatic platforms that's filled with what's called MFA content, that's almost entirely created by CGI -- by GenAI. And the second aspect of that is, and there is another study that just came out that said that by 2030, 99% of Internet content will be created through AI, 99%. So almost all news content and almost all other content. So AI creates...

What does it say about equity research?

Well, look at the financial articles. Half the financial articles now are AI generated.

And how detailed are they, really?

They're actually stealing your data half the time. So I think it's creating a very challenging environment for advertisers. There's tons of MFA content. We're seeing increases in fraud. We've seen twice as much mobile fraud this year than we've ever seen in the past. So it's creating more fraud, it's creating a more challenging environment. What that means is, for us, it means greater demand for our solutions. It's creating a more challenging environment for advertisers to participate in, to go out and buy advertising in. They've got to have more tools to do so. So tools that we have that help them avoid hate speech that's being amplified at a greater level, that avoids MFA content, that helps them to avoid fraud that's being generated. Those tools become more important for us. And we use AI to fight AI. As you noted, our classification engine is being driven by AI. Our Fraud Lab uses AI tools to start detecting patterns. So it is -- the arms race has just escalated from slingshots 10 years ago to rifles to now bazookas, right? And that's where we are. And we just have to have the arms on this side to combat that, too.

I mean, very few people have the amount of data that DV has, it goes through. And so I'm curious, how is AI or maybe -- you talked about AI and ML. But I'm curious how -- if GenAI is helping to safen through that, so you know all of this, short-form video content, you talked about music and audio and video altogether.

Yes. I mean, look, your models are only as good, and your tools are only as good as the data that goes into them. And we see over 5 trillion transactions a year, right, that we get paid for. We see more that we even don't get paid for because as we do analysis, we look at more and more data. . So we see trillions of transactions. That gives us tons of data to build our models around, which allows those models to actually be incredibly efficient and effective. Now it doesn't mean, again, there isn't some type of arm race between bad guys do bad things, and good guys trying to protect advertisers. But we think, we're pretty well armed based on the fact that we have so much data. We have a legacy of data that we have to pull from. And the advantage we have, too, is that we're working with 1,000 of the biggest brands on the planet. And that data that we work with them to generate is information that we can use to then create smarter solutions over time. And that's access that we have exclusively, right? The guys on the other side don't have access to that data. The guys trying to do bad stuff have to come up with their own modeling to spin things out. We can learn from things that we've seen in the past.

That's great. Creating chaos, I think, is the right way to say. And of course, we're just looking for, what it can be and where it goes. Maybe I'll switch topics. We have about 5, 6 minutes left. So if there's any questions in the audience, start teeing them up. But talk to us about the acquisition about Scibids.

Yes. And maybe I'll let Nicola talk here for a moment...

He had to pay for it.

That's right.

He pay for it.

So we announced the acquisition of Scibids on our second quarter call. Scibids fits on the activation side of the business, right? So we have the measurement side of the business, where we evaluate how well an ad performed after it was actually run. And then we have the activation side of the business, where we help advertisers evaluate where they're about to place their ad dollars. And historically, the way our data was used in that part of the business was, our data was ingested by programmatic partners, such as The Trade Desk and DV 360 as two of the large ones. And the data was used to allow an advertiser to bid one way or the other based on the data that we provided. So it's a very binary type of decision. They would use our data, and we'll take the example of brand safety, and say, is this brand safe, is this not brand safe. It's brand safe, go ahead and bid. It's not brand safe, we don't bid on it. The evolution of that product was Authentic Brand Safety. Authentic Brand Safety allowed an advertiser to not only do it at the brand level, but go down to sort of the campaign level and be able to create more sophisticated filters before they were actually going to say, yes, bid on this, don't bid on this. It's still sort of a binary decision. What Scibids allows us to do now is use our data, and algorithmically predict with our data, whether they should bid or not bid on an ad impression. So down to the impression level. So from just the brand level when it was brand safety, to the campaign level when it was a Authentic Brand Safety to really with any of the campaign, can you, rather than just say, yes, no, use the Scibids technology to pull the lever one way or the other towards yes or no, depending on what you're looking to do with your campaign. So a lot of advertisers were coming to us and say, look, I understand what you're telling me about brand safety. I understand what you're telling me about viewability, whether it's viewable or not. But within that, I would like to have the ability to say, I'm okay to sort of throttle that lever a little bit, if the CPM is going to be low enough that I'm actually going to be able to have a return on my campaign. Scibids is that tool. It allows the advertiser to really throttle that. It allows the advertiser to bring in additional data sets into their decision. Data sets that are specific to the advertiser, so data that they're collecting elsewhere, such as pricing or intent of buying something. And so it creates a much more sophisticated way of using our data. It moves from just yes, no decision to a far more sophisticated way of using. So the impact it's going to have for us is a lot more usage of our data rather than just having a yes, no decision. Our data is used for yes, no decision, but we're only getting paid when it's a yes, right? This allows us for data to be monetized in a far greater amount of transactions. That's intuitively what that tool is going to do for us. So basically, it allows the advertisers to make an even more informed decision using our data, on the activation side of the business. So we're very excited. We acquired the company. We acquired talent, which is not easy to come by a lot of data scientists. This is not going to be a cost synergy exercise. This is going to be the stepping stone for how we're going to develop all of our data science work around AI. It is a product that's based on AI, on algorithmic predictive tools, right? There's a lot of conversations about AI, creating efficiency in your model, creating ways of cutting costs, which, of course, we will do as well. But this is a product that's based on that technology, which is pretty unique.

And the integration with your data to have that -- I was going to say nuance, but to be able to go up and down whether to say, yes or no, how difficult is that integration?

Well, Scibids is an existing product. It's an existing company that already has...

And you've been using it?

Yes. Yes. So one thing we did do is, we actually -- we tried it before we bought it, right? So we had a partnership with them to try out the product. They have an existing client base. It's a small client base. They have, importantly, they have the integrations for the DSPs. And so for us, the decision of buying versus [ build ] is pretty obvious, because they kind of have all the proof points in place already.

That's great. Super helpful. We have about a minute left or two. Any questions in the audience that we should ask the team?

Could you talk a little bit more about the international growth opportunity? Maybe how the land and expand approach differs from your domestic markets? And then maybe also the health of the international ad macro.

Do you want to talk about it?

Sure. So international, we've invested in the international -- the infrastructure of international growth about 24 months ago -- 24, 12 months ago is really when we established our presence in markets where we weren't. Just for the interest of time, the simplest way to think about international for us, it's not even 1/3 of our measurement business. It's well over 50% of spend for large enterprise clients. That's the opportunity, right? So the investments are made. Last year was pretty tough. In EMEA, in particular, and APAC, Q2, you saw huge growth for us based on the infrastructure investment that we already made. So this is a matter of being on the ground, being able to compete on RFPs. And our growth rate is really tremendous on international.

Yes, our EMEA business grew over 30% last quarter. Our APAC business grew over 50% last quarter. And we just -- we announced and we have -- we've opened 3 new offices in APAC and 2 additional offices in EMEA in the last 2 months.

Super helpful. Well, with that, Mark, Nicola, we're over time. Thank you very much for coming today to our Global Tech Conference. And really learned a lot. So I appreciate it.

Thanks for having us.

Thanks for having us.

Awesome stuff. Thank you.