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

[Audio Gap] Internet sector here at Citi. And this is one of my favorite conversations. We have DoubleVerify CEO, Mark Zagorski. We have DoubleVerify CFO, Nicola Allais. Is that correct? Is that the right way to say it? And there's just so much to talk about. We saw each other a few months ago, I guess, during the summertime. I think most people know by now what DoubleVerify does. We'll just say measurement, brand suitability, attention, there's all sorts of things. Suffice it to say, sort of required, if you're a platform, if you're a brand, to make sure your brand is safe and now it's across social and video and retail media and the list goes on and on. So with that, Mark, Nicola, welcome to Citi's tech conference. Let's kick this off with macro. And so we'll go back and forth here. And just this last quarter, we were pretty excited. There was a lot of good things going on from a macro perspective. But then there are a few cracks from a brand perspective as well. And so because DoubleVerify sits in a really interesting space that you see the OpenWeb, the closed web, across video, non-video. Just, how would you characterize the current mood within online advertising? And talk to us about that.

Yes. I mean, I think we're in a period now where from an advertiser's perspective, it's cautiously optimistic, right? So I wouldn't say that it's exuberance when it comes to ad spend, but advertisers feel okay. And what's happening when it comes to digital advertising is, we're definitely seeing the era of winners and losers starting to emerge, right? You're seeing dollars starting to move into social in a much more aggressive way. You're seeing connected television and retail media networks growing. And what they're doing is, in many cases, taking away from what people would consider traditional publishers, right? So news publishers, people like that. And I think that trend is continuing over time. So you see certain platforms doing incredibly well. You see other platforms struggling and traditional media platforms struggling on the receiving side of dollars. And then you see the brands that are spending, I think, like I noted are cautious, but they're also looking to drive performance, and that's becoming a bigger and bigger play here. Everything needs to have a return. Everything needs to have a return on investment. And digital media is media, -- like for the most part, everything is digital, and everything is trackable. So I think when we look at the mood right now, it's, "Yes, I'm going to spend. I feel good about spending. I'm a little worried about elections. I'm a little worried about what may happen with the economy. But ultimately, if I can drive our results, and I can track that result, then I'm okay spending."

And we got -- coming out of this past quarter, Meta had tremendous results. Google Search had very good results, but then YouTube not as much. Relative to expectations, they were fine, but there's some concern about brand. Use the term cautiously optimistic to that and then also the importance of performance. And anything to call out between performance or DR or brand or just cautiously optimistic across the space.

Well, I think, again, it goes back to the idea of there's some winners and then there's people who are struggling a bit. And the winners are the platforms that are able to actually close the loop -- that can show an advertiser that their spend is actually performing. And I think Meta has been really good at that for the last few years. You see the emergence of retail media networks, as I mentioned, because those are specifically built to drive results. I mean the entire universe of retail media networks. And if folks aren't familiar with that, that's where people like Kroger or Amazon or even Uber have built advertising businesses on top of their core business. But the beauty of those is, because they have so much data, they can show the efficacy of the advertising on their platforms and how it actually drives a transaction. Amazon has obviously been the big winner there. And they're also -- they're reaping the benefits of that success as well.

So let's transition then into DoubleVerify, because all these newer -- retail media call it new, newer format social, CTV, et cetera. DV's core model and value prop. I just want to get that level set with everybody. We know it's measurement, verification, brand suitability. Just talk to us how these newer formats that DV can now measure have expanded -- talk to us about the value prop overall and then how the business has expanded over the last, call it, 2 to 3 years, now that social is a bigger part, CTV, the bigger part, et cetera.

Yes. We started off -- the core premise of DV was to protect digital ad spend initially in the OpenWeb. And the OpenWeb is an area where there's lots and lots and lots of opportunities for impressions, right? So there was a data-driven impression universe in the OpenWeb, but there was a lot of concerns about lack of transparency. Where is my ad showing up? What kind of content does it showing up next to? That's where we started to ensure that the buyers and sellers of media in the OpenWeb -- that, that transaction was protected, that there was no fraud in engage that the ads were viewable and ended up in a brand safe or brand suitable environment. Fast forward to today -- that was 10 years ago -- over 10 years ago when we started, more and more of those dollars -- or actually a vast majority of those dollars are moving into platforms like social and like CTV, which have their own challenges. In some cases, it's not the lack of transparency. You know you're getting an add-on meta, right, but you're more concerned about what's the tone and tenor of the content in a social network. It's user generated content, right? So we ensure that ads are from a brand end up in a content that makes sense for that advertiser in social. So you've seen the emergence of social and our products have adapted to social environments. Platforms like TikTok, for example, which didn't exist at scale 4 or 5 years ago, is now one of our fastest-growing social segments because TikTok has incredibly high engagement with consumers, but has lots of challenging content, right? So brands need to be there, but they don't want to be around stuff that is not appropriate for them. We play a role in ensuring that their ads don't show up in those environments. So you've got that emerging. You've got the connected TV world emerging, which, again, several years ago, was not a big factor in advertising when there was only maybe 1 or 2 ad plays in CTV, right? There was YouTube, TV and then there was Hulu. Now you've got every major platform from Disney, to HBO, to Paramount, all being ad-supported, right? Brand safety not as much of an issue, but things like viewability is an issue. Now people wouldn't assume that viewability is a challenge in CTV, but we found that many ads were being served to CTV connected television devices when the actual set was off, right? It's a different world. A TV set today is basically just a big computer screen. When you turn that screen off, the computer can still run in the background, which means ads can still be served there, too. So is that ad viewable? It's delivered while the TV set is off? Obviously, no. So we measure things like that as well. So each new emerging media source comes with its own challenges for advertiser, whether it's viewability on CTV, brand suitability on social environments or even fraud transactions or attention engagement across retail media networks. There's always a challenge for advertisers to know that their spend is insured and verified.

So I was going to ask about win rates. We'll get to that in a second. I think we have all these different topics, and we'll open up for questions as well. But Mark, if I had to take a step back and you said 10 years ago, OpenWeb, now we're in social, CTV, retail media and of course, open and everything else that comes with it -- we would love to hear your thoughts maybe bigger picture, opportunity set. What gets Mark Zagorski super excited? We've been talking about CTV -- I mean, we've been -- brands know each other for [indiscernible] and maybe more than that. And so -- [indiscernible] how old I am. So just all those things, like help us understand the broader idea and opportunity around? Is it social? Is it video CTV, retail media sort of newer? How do you think about this?

Yes. So I think the biggest opportunity for TV as a company right now still remains in social. We've run some estimates. We're probably only measuring and verifying around 5 [Audio Gap] across social. We have an activation tool on YouTube right now, which is actually doing really well. But on the other platforms, TikTok, Facebook, Snap and others, we don't have solutions there yet. So to me, that's a huge opportunity considering how many ad dollars go into social. So that's a big opportunity for us. You mentioned retail media networks, I think we're doing...

So I'll come back to it. I think the prebid is a really key part and you talked about closing the loop and the importance of -- just the ability to do -- to improve performance and [indiscernible] and everything else. And so we have prebid and activation. I think DV has postpaid solutions as well, I believe.

We do -- across social.

So just talk to us more about the ability to close the loop and how that is differentiated. Because we all know social is gigantic. We know, to your point, 60% of nonsearch or some 60%, 70%. Talk about prebid, the importance there.

Yes. So when we look at our prebid solutions in the OpenWeb, we have a solution called ABS with authentic brand suitability. And it links -- the way it works with our measurement, as you think about we capture data in measurement that says, "Hey, here are violations that we've seen." We're able to take those violations in that data and pump it into our prebid solution. So the advertiser then can filter out those types of impressions in the next time that they bid. And this goes -- in the OpenWeb, that gets implemented through platforms like the Trade Desk, through DV360, through Microsoft Xander, through Amazon, so think of that as being the filter that we have is an activation of prebid in the OpenWeb and then we're able to measure and adapt. So you create this virtuous cycle of optimization, which takes garbage out of the system and allows the advertising to perform better. So we talk a lot about protection. But ultimately, when we take prebid activation and match it with postbid measurement, we're actually driving performance, right? And we started this conversation, what are the advertisers looking for right now. They're looking for performance and ROI, right? So it squarely puts us in the world of not just protecting ad spend, but helping it perform. Now let's fast forward to the social universe. Social to date, for the most part, we've really focused on measurement. Let's get the data in, let's find the violations. In some cases, we can block the ad before it gets delivered. But really, it's a verification measurement. We launched a solution in YouTube to allow us to do filtering, right? So it works with our measurement tool now. And the value of that is immense. And if you look at, for example, our YouTube measurement penetration in our top 100 customers, it's over 90%, right? So when advertisers can get measurement and put it together with activation, it's this amazing package of performance, right? And we know we can penetrate. If you look at, for example, our Meta penetration right now, where we don't have prebid, it's only around 50% of our top 100. So we know that one of the keys to unlocking value in measurement is having an activation solution there as well. That's why we look at it as a big opportunity. We're working on those channels right now to be able to provide those types of solutions, but we know prebid or activation solutions and post-bid measurement solutions work really well together because they drive performance.

Right. And that's key and sticking on the social theme, you talked to think about TikTok, you talked obviously about Meta. Would love to hear about, where are brands focusing their time on these platforms. Obviously, TikTok is all video. Meta, we think, reels and we all know reels are driving a lot of this. I'd love to hear your thoughts on what you're seeing between the feed, maybe stories, if you're Meta or reels or TikTok or [indiscernible]...

I mean, that is the future. That is where advertisers want to spend. That has incredibly high engagement and attention. It drives tons of volume. And it's -- you get a little bit of the sound in motion from television, right, along with the engagements and the customization that you'll get in a social media environment. So you bring those 2 things together. It is a powerhouse. That's where advertisers want to be.

And do they care of a platform? Is one platform better than -- I'll ask from a DoubleVerify perspective.

I mean, we see impressions across all the platforms. And really, I think advertisers don't think they have a preference really when it comes to short-form video. It's wherever they feel that the creators are most active, but every creator now is pretty much on every platform, right? They'll build something and distribute across TikTok [indiscernible]. So creators, they follow creators, technology of the platforms does matter a bit. So the ability for the platforms to target. We think that our ability to be engaged with those platforms and be able to verify is important as well. But considering that we're across TikTok, we're across reals, we're across shorts. There's really no place that we aren't. That's not a real differentiator for the platforms yet. But it's great for us, right, because they feel like if they don't want to have verification, then that puts them in a disadvantage. So I don't think there's a real -- I've not seen a real preference for platform right now.

And I don't know if you -- so we talked about the 5% of social impressions, the DVCs, maybe that you mentioned that earlier, that was a '23 metric. Does that get to 100% over time? Do the networks allow you to get there -- platforms?

I mean, we see every impression that our advertisers are engaged with, so it's not the platforms. It's how much advertising engagement that we have. I don't think we get to 100%. And the reason why is DV is an enterprise solution, and we work with the largest brands in the world. So think Unilever, Mondelez, Pepsi. These are huge brands, Colgate. Meta has 10 million advertisers, right? A big portion of their business, and a lot of the platforms' business is self-serve, SMB, kind of engagements. That's probably not a place that we're going to be at any time soon. It doesn't mean we can't be there. But I think initially, right now, we've got enough enterprise clients to go after that we're really focusing on the big brands.

I cut you off on the CTV comment. We talked about all these different platforms and...

[indiscernible] you keep talking -- you better cut me off.

Maybe the bridge to CTV is YouTube. And then we talked about YouTube before with prebid or what have you. But I thought the comment around, I think, Performance Max and the Demand Gen are now live. And getting back to the importance of performance, the importance of closing the loop, just talk to us about how brands are looking at YouTube and relying on DV to get to that performance side.

Yes. So look, I think all the platforms, whether it's PMax or what Meta is doing with their automated performance solutions. All those platforms are focused on driving results. The idea of branding without an ROI on the other side, it doesn't exist anymore. What we do plays a role in that because there's one aspect that's usually nonnegotiable. And that's the brand safety and suitability, right? That is an inherent part of what advertisers look at as performance. Yes, I want to perform. But I don't want to perform if I'm next to content that I'm not comfortable with, right? And for the most part, content that's not brand safe or brand suitable doesn't perform anyways, right? Videos that are incredibly incendiary don't sell soap, right? And Unilever doesn't want to be next to those...

And that still exists today?

Yes.

I got served a few things about that recently. That's helpful with the brand safety, suitability and you need to have brand. Let's talk about CTV. I think bigger picture, there's just a lot of debate going into this year, we heard and we saw Amazon Prime now open up. Every single platform has an ad-supported tier for the most part. And so there's always challenge around. Okay, we have more inventory, pricing might come down. And I thought it was fascinating, this past quarter, I think you -- DoubleVerify reported 55% growth in CTV measurement volumes, which talks about the volume and the demand here. Where are you seeing this increase in volume? And I'll stop there and not lead the witness here, so to speak. So another question is bigger picture.

Yes. No. So we consider -- it's interesting. Because YouTube is both social and CTV, right? Kind of fits into both universes. YouTube is still the largest portion of our CTV measurement business. But we've seen really nice growth with Netflix since it has grown from a very small base. And we're across every -- we measure every major CTV platforms from Warner and HBO to Disney. So we've seen growth across all of them really. And the interesting thing about it, which you brought up, is the flood of inventory into the market is not a nonfactor. I mean it has absolutely impacted CPMs. I mean Amazon deciding one day just to say, "Hey, we're going to flip on literally billions of impressions," changed the market overnight.

Is Amazon fully live? I know they're fully live, but meaning that they flipped it on or they announced it in February, but I feel as if a lot of their new programs are coming on...

They're holding back.

They're holding back?

Yes. They haven't even started to actually put real volumes of impressions...

So the pressure on CPMs is sort of freak.

Yes. I mean, I think I would say 2 or 3 years ago, CTV was a seller's market. It was absolutely a sellers market. They could demand pricing that they wanted. They didn't need to provide transparency in many ways. But I think that's changing, which ultimately from a selfish perspective, benefits us, because we're in the business of transparency and impressions because we charge a fixed rate per impression. So CPMs go down, impressions go up. That's a good thing for us.

I agree. We've talked one-on-one about show-level transparency. I think just talk to us how that maybe is changing the purchasing path of CTV. And for those in the audience, making sure that right now you buy an audience or just help us understand show level from where we are today.

So from -- if you take a look at how CTV is bought today, it's very different than how linear TV was purchased in the past. Linear TV, you bought American Idle on Thursday nights, you know you've got a guaranteed spot there. You bought the show. CTV, you don't buy the show. You might get the show that you want, right, but you're really buying the audience and most CTV platforms are very concerned about selling the show. And the reason why is they have huge catalogs, right? And a majority of viewing goes to their long tail. And they can't sell thousands of programs individually, right? So they sell on a genre basis. They sell on an audience basis. Good for the platforms, challenging for advertisers. Because advertisers want to know, "Hey, I want Stranger Things. Hey, I want Bosch, actually -- [indiscernible] Bosch of old guys -- [indiscernible] looks like a 60-year old white guy who liked Bosch. So it's -- when advertisers want specific shows, they don't get that. And why I mentioned before the idea that is becoming a buyer's market versus a seller's market is, in the past, the platforms didn't need to provide that transparency or that accuracy. You're going to get an audience, but you're not going to get the specific show. We're not going to really tell you where you ran. That's changing because of the amount of inventory that's coming to the market. What that means for DV is currently, we don't get what's called show level data. We don't get granular data on where impressions ran on a show. We get very broad data on the application itself. Earlier this year, we launched a partnership with NBCUniversal to start providing that data on a granular basis. And we think the platforms are starting to realize that, that data is going to need to be put out into the measurement world -- into the verification world for them to compete. Because every other platform, digital platform, shows you where you run, right? Even short-form video shows you these are the types of videos you ran against. Here's the video you ran against, did I run against this creator, et cetera. CTV doesn't show you that, right? So I think that's changing. That show level data, of which we're getting a very limited about today. I think, there's an opportunity for us to get basically the same amount of data we get across all the platforms on a broad basis to get it on a granular basis in the future, which will change the dynamics of how CTV is bought, but also changes the dynamics of how we sell verification across connected television.

Super helpful. And we'll talk about show levels for years to come, probably, early days on this. Maybe last one of the big, major verticals that are driving a lot of the growth, retail media. I think DoubleVerify's measurement tags are accepted by 100-plus net retail media networks, supply-side revs are growing 50-plus percent. Would love to hear just the demand from your clients as they expand beyond maybe the socials of the world, CTV, the OpenWeb and how retail media sort of expanding that out?

Yes, retail media is a really interesting space, and you mentioned 100 retail media networks. It's crazy. If you think, like, Dollar General has a retail media network, right? You've now put retail -- everyone is selling advertising. If you're in the retail space and even people who aren't in the retail space -- so Uber has an ad network now, right? We were with people like Uber, with Kroger, with Amazon, with Macy's, with Target. These are all pretty large retailers. And they're using the data and their engagement with consumers to actually build an advertising business on top of it. If you think about it, there's no better way. If you're a CPG company and you're selling -- we work with Kroger. You're a CPG company who has the best data on what people are buying in Kroger, right? So to be able to target a message against those users is incredible. Same thing with someone like a Macy's, if you're Ralph Lauren or a clothing manufacturer. You know exactly what someone has been looking at on Macy's. You know what they've been shopping for. You know their purchase history. Using that data to actually target those users is incredibly valuable. So our business in the retail media networks, as you noted, grew 50% last quarter. It helped drive, what we call, our supply side growth to over -- to almost 29% growth. And it's an entirely new segment for us. Now some of those dollars were dollars that used to go to the general OpenWeb, but some of those dollars are totally new. Those are co-op dollars coming from certain brands coming out of other budgets. So it's pretty interesting for us. Again, we talked about all dollars are digital now. Think about it like, everything is going into digital media. Retail media networks are taking like dollars out of co-op, dollars out of in-store advertising that now are going into digital. And it's a continually growing area. The neat thing for us as DoubleVerify is, we work with a lot of these companies as both retail media networks. So Uber is a retail media network customer, Macy's is, Target is and as brands themselves. So we had this great opportunity with someone like an Uber who uses us for their retail media network, but they're also an advertiser that uses us to verify their ad spend, too. So we're in a really unique position to take advantage both of the retail media network business, but of those retailers themselves and their ad spend, too.

The core and then supply, like both supply and demand. So let's talk DV specific, and we'll open up to questions in a little bit here. I think we have maybe 13 minutes left to be exactly.

You got with this [indiscernible].

Let's talk about it. So I was going to ask about win rates here, and I'll let both of you sort of talk -- I think you were at 80% win rates in 2Q, 70% of those are greenfield, 50% penetrated the top 1,000. These are stellar numbers. And I think they've been 80% for a while, maybe since going public. Would love to hear your thoughts on how do you expand that penetration of the top 1,000. And then the cohort level data as you get into year 2, year 3, because I think you've talked about some accelerating growth there.

Yes. So I think -- I mean, the win rate is really what's driving our continued growth, right? So if you think about greenfield that's really more outside of the U.S. than in the U.S., we are not as penetrated outside the U.S. as advertising spend, right? So it's only about 30% of our revenue for the measurement side of the business, but really sort of 50% of all global ad spend is outside of the U.S. And that expansion is greenfield. It's clients that are coming to brand safety and brand suitability and verification in markets. We invested quite a bit to get ourselves situated in those markets so we can grow not only with U.S. brands that want to expand outside U.S., but actually local brands, and we're seeing very strong growth outside of the U.S. That's the greenfield part of the business. In the U.S., we work with 50% of the top 1,000 advertisers. That means there's the other 50%, and we can go after with products that we feel is superior to what's in the market. So that's kind of the growth pattern that we see in the market. It is a growing market, and we're kind of going where the opportunities are. The 80% win rate has been a pretty consistent number. And that's on wins for any product solution that wasn't available or wasn't being used by a client, right? So it's also just sort of the upsell of our new product solutions, which is really where we're gaining a lot of traction, right? So for example, we bought a company last year called Scibids, that's a product that's unique in the market. We're able to go after our existing client base and upsell them to a solution that's not available anywhere else. So it is the driver of the growth. We have net revenue retention that's about 120%. And the way we get to that is, we upsell new solutions to our existing base, and then we go after greenfield opportunity.

And that's year 2, 3.

Yes, exactly. So on average, our top 25 clients are with us for over 7 years, and that's our ability to not only expand in new geographies with them, upsell them new products and continue to kind of go deeper with the client wherever the growth is.

I just want to hammer one thing here. Nicola just gave a number out there. Our top 25 clients have been with us -- actually, it's our top 50 clients have been with us for over 7 years. I mean, I would challenge you to find a software company out there who has data strong as that. I think that's an amazing testament to not only the quality of our solution, but the stickiness of our relationship with our customers as well.

Super helpful. Let's stick with the finance part, and I'll come back in a second on the business part. Let's see. So we report revenue in 3 different lines, I think, activation, measurement, supply side. So activation, DV is the largest line. How do we -- we just talked about social and CTV and retail media. Talk to us how we might reaccelerate the growth there?

Yes. So just to level set, all 3 lines are growing double digits, right? So we're in a position where we're able to benefit from wherever the advertisers spending their dollars. You're correct that in the past few quarters, years, activation has been the main driver of the growth, because the dollars are going there, right? And so we kind of follow wherever we need to verify for the advertisers, but also because we had a product called ABS that really grew very strongly. Now activation still grew double digits even though ABS didn't grow double digits in the last few quarters, mainly based on some advertisers reducing their spend levels and they were using ABS very strongly. But within that line, we still grew double digits. And what drove that growth is everything that Mark just spoke about, which is the Scibids product and the social activation were double-digit growth along with the core product also growing double digits. So we are entering a phase where more lines of our revenue are growing as opposed to having to really focus just on ABS. It is a bit of a transition, right? And as Mark said, we do need the social activation products to kind of close the loop with the measurement side of social, which is where the main growth is. But if you think about the profile where we're still growing double digit in a time of transition towards social and closing that loop, that's a very strong proposition, right? Supply side, I just want to mention the supply side. You mentioned 50% growth on the retail side. That is what drove the 26% that we saw in the overall supply side line. It is a smaller revenue driver for us, but the fact that it's growing so fast around retail media, which is one of the opportunities we have to continue to grow. It's also a sign of where we're going to see growth. Overall, I would say -- I will give 2 stats, which is social is not even 20% of our total revenue. So if it's 60%, 70% of all spend, it's only 20%. You do see the opportunity there to continue to grow. International already mentioned, it's only 30% of our measurement, it's over 50% of all revenue, digital advertising spend. Those are the drivers that we know at some point, our revenue mix will look like our advertiser spend, and we have a lot of opportunity to continue to grow at double digits on all of our lines.

And Nicola, just to hammer home on those few brands that might have pulled back. People talk about the 6 cohorts. Talk to us about where we are with those. Any update on...

Yes. So it's an uneven pattern of spend. That's what we've said all along for the whole year. It is a large gap towards our expectations that this group would continue to grow. So the unevenness remains, right? We said it didn't help or hurt. It didn't get worse or better through the second quarter, but it is an uneven pattern for large advertisers, 3 of the 6 are in our top 10. So obviously, it has a material impact on our overall trajectory.

And why did they pull back?

These were specific issues related to the advertisers. So what's important to us is, they didn't turn off services, right? But they were heavy users of ABS. So as they lower their ad spend levels, we saw an outsized impact on ABS in particular. What's important for us is, they didn't turn off the service, they didn't go to another provider. This is just intrinsic issues within the advertisers that has created a cohort that spending less. We think, overall, this will be a cohort that will have a moderate decline in the year, right, in their ad spend, which is reflected in our revenue. So it's obviously a drag to our numbers. But even then, we still did grow double digits on all of our lines.

Right. And then hard to -- I got to get you. So let's open up for questions. I got several others, but there are questions in the audience. Yes, wait for a mic, yes.

Yes. Thank you for sharing so much information. So you just mentioned that advertisers are increasingly focused on the short-form video than the creator actually because [indiscernible] doesn't make a difference, but the technology. So because you are doing the measurement and the verification for these clients on these 3 platforms, TikTok reel and the YouTube shorts. So how would you rank their accuracy, you say TikTok is 100, like where are reels and YouTube shorts in terms of their targeting accuracy.

With their internal tools and their ability to -- I'll just be direct. I'm not an advertiser, so I couldn't tell you how accurate their advertising tools are. I know that our engagements with all 3 platforms are similar. So we provide the same level of verification, the same level of suitability measurement across all 3 platforms. So I couldn't tell you how strong their other tools are across the different platforms.

Is there a volume way to think about it? TikTok volume here, shorts -- are you seeing more volume on one platform in particular?

Yes. In TikTok, I think, from a short-form video perspective, we're seeing the most, I think, just because again, we're going to be biased based on where our advertisers spend. So I think TikTok is still probably the largest -- the short-form video engagements that we've seen. But I mean, look, shorts and reels are big. And YouTube remains our largest social platform that we engage with.

Any questions? Yes, we have a mic coming.

Just a follow-up on the large operators [indiscernible]. Have you seen any increased ability since Q2. And like what gives you confidence that [indiscernible] ABS solutions is not reasonable [indiscernible].

Yes, I'll take that. So as I've said, they didn't turn off the service. So really, the question is about ad spend patterns within the 6. And we've spoken in the past about some of the 6 advertisers that have intrinsic issues where they're not likely to come back to the same levels of spend as they were in the past because they have internal considerations that have reduced their ad spend. So the visibility hasn't improved. It's still an uneven pattern and that is not a common that's changed. It's still uneven. And coming out of the year, it's pretty clear that some advertisers won't come back to the same levels as they were before.

You have 3 lines of revenues. And could you comment on the revenue model? Are they pretty much all based on volume? And then maybe you have different tiers and the more you spend, you probably have to cost less per impression.

Sure. So we have 3 lines of revenue. We'll report on 3 lines. 90% is for measurement and activation, and that is the advertisers paying us and the way they are -- that we're getting paid is a fee per impression that we measure. So the driver there is obviously volume, right? The more volume we see, that is what's driving our revenue growth. Within that, if you think about the model, we're not really tied to the CPM with a media buy, right? So we don't take rate on the ad spend. So we're less susceptible to deviations on the CPM side. For us, it's all about impressions and just being able to verify more and more impressions. So as the market grows and more dollars go to digital, we benefit from that directly because we're just paid for impression. The third part of the business, which is growing very fast, 26%, that's got some 10% of our revenue, and that's more of a SaaS model where we sell our data to platforms for them to do basic level brand safety and fraud before that inventory is sold to the advertisers. So ultimately, we really are tied to an advertiser using our services. Supply side, that's correct.

Yes. So you might have time for one more or we're about to go into over time? Okay. We have one more and then we got to wrap.

You talk a little bit about the Scibids go-to-market strategy, just maybe how that's evolving and any early learnings that would be great.

So GenAI did not come up until now.

So very quickly, Scibids is our algo-based optimization tool that we insert into our activation business. So think of -- our current activation business is very much a binary decision. It allows an advertiser to say yes or no to bidding on an impression based on data we provide. What Scibids allows us to do, it allows us to say, maybe. So based on a KPI on the other end, based on the price of the impression, maybe this is not considered viewable by industry standards. But if I get it cheap enough and if it drives a result, I'll bid on. So basically, what Scibids does is it allows us to look at a spectrum of decisioning as opposed to a binary decision. [indiscernible], so Scibids initially -- we brought it in, and we've now integrated the entire solution into our solution. We're putting, for example, our activation data into Scibids now, which will allow advertisers to make decisions based on attention scores. So think of that as a version of viewability, how engaged the user is. The go-to-market right now, we have got 8 of our top 100 customers that we've sold so far this year using Scibids. We continue to scale that business, we've said, in the next 4 years. So we've announced last year that it would be a $100 million business. I think we're well on our way to that trajectory. And the nice thing about it is, as Nicola noted, it's a real differentiator. None of our competitors have a solution like that to the point where we now have advertisers who use our competitors' solutions for measurement, using Scibids for optimization because they don't have a tool to do so. So the go-to-market now is, we're bundling it with a lot of our solutions, our activation solutions. We're reaching out to advertisers who don't even use us for measurement, and we've been able to penetrate those as well.

With that, we didn't get a chance to talk about Oracle and we did not get a chance to talk about gross margins that are expanding, and a bunch of us that [indiscernible] -- so anyway Mark, Nicola, thank you very much for the time, very much enjoyed.

Thank you.