Taboola.com Ltd. (TBLA) Earnings Call Transcript & Summary

Perfect. So we'll get started here. Adam, welcome. Let's get started. My name is Ron Josey, I cover the Internet sector here at Citi. And I'm always excited to have my old friend, Adam Singolda here with us today, the CEO, Founder, Co-Founder of Taboola. We've known each other for a number of years now. I think I was thinking about like what the right icebreaker is, when we were talking about this just now as we catch up.

Talking about your latest LEGO set and then we get into like the real business.

I'm building the Formula One car. It's probably like 2,000 pieces and not that big. I was trying to count.

With the kids or on your own?

That's my notes we know each have their own. My son is doing this like art piece, my daughter is doing a unicorn and other ones doing, I think, a super Mario. So we have a -- I realize that I enjoy a matter 2. I can't just do things with them all the time. So that's where I'm right now. It's fun.

You matter, too. I think that's -- I don't disagree. So why don't we kick off here? And I would love to level set for everybody in the audience living on the webcast. Just give us an overview of Taboola just the evolution of performance advertising has been pretty extraordinary. It just continues to be a greater part of most companies' budgets. And so just talk to us about Taboola's approach and where things are going from a performance advertising perspective.

One, I think over the last few years, you've seen -- for those who may not know, Taboola is the largest kind of outcome or performance advertising platform outside of Google and Facebook. So we specialize in helping businesses big and small to reach a certain outcome that is their growth strategy. So they always go to Google first. They would go to Meta, then they get kept out and they're looking for other partners that can help them satisfy that. The definition of what's outside of batteries and social changed over the years. It used to be publishers at WW someone. And then over the last, really, 5, 10 years, it evolves into anyone that wants a piece of the advertising market. Uber is making money from ads. The #1 line of EBITDA for Amazon is ads and services. Netflix said they'll never do it, they love ads. Everyone is in the advertising business, and everyone needs a good friend that's not maybe Google or Facebook. And that's where we fit in. So I think the market has evolved to your question from just traditional Open Web websites to apps OEM manufacturers, utility e-mail providers, just anyone that wants to make money from ad. That's one thing to change. Two, the world is crazy. So the idea that you can spend money in not knowing what's going to happen with that money over time went down. You really expected to have an attribution model that allows you to spend and know that it is working. And so that's also changed. And the third minute changes AI, like that's reshuffled everything, conversational ads, the idea that consumers now talk to you, how is that going to be monetized, what type of data that creates. We just had on stage here before we do understand we had -- from us today, and that's obviously a big part of what they talked about.

So let's get into that in terms of how the online ad market is changing. Before I do, if you could just give us a level set of the online ad market in your view, healthy, building challenged? How would you look at the market overall in terms of demand?

I think it's, for the most part, stable. I think especially as there's more focus on performance. Performance advertising tends to be more stable is the last thing it stops, I would say, it's mostly stable -- of course, we're seeing pockets and geographies where things are a bit different. And then the biggest thing in the future, which I'm sure we'll get into is if consumers are going to be converting so much with LLM, how is that going to transpire into like what does that mean to the economy and advertising. But outside of that, I think today is pretty stable. .

Yes. Certain geos, we hear a lot about like decent in the U.S. from an advertising demand perspective. So yes. So with that, Adam, you actually -- you highlighted a topic that I think is coming up in every single conversation, which is how GenAI is changing how we all interact with the web and agents you can easily see how you can have that -- you're already having a conversation with agents. A lot of companies are launching these agents that are your best friend or can do everything for you on your behalf. Just talk to us about how the Internet can evolve or how it evolves with agents being your front door to a certain extent, like things are changing in terms of how we go online and what we do online.

I would divide it into 2 parts that I think are interesting to understand is investors as well that are looking to spend money in this ecosystem or not. So on one side, you have the demand side. How will AI affect billions of dollars kind of flowing into the Open Web and anywhere else. And I think the biggest change there is the evolution from programmatic, which is a protocol that has been used for the last 30 years. by agencies and advertisers to buy ad and you know this really well, you covered many companies in that space. .

I'm smiling because 30 years seems like a long time yet, we're still in the evolution programmatic.

Well, I think it's going to evolve into nothing. I think it's evolving into -- I think it may be replaced -- over time, we can discuss what's over time. But when you're an agency, and you're spending $60 billion a year on programmatic and ads, and you have an opportunity to build your own buying machine using Claude code or Codex or maybe SpaceX with -- and you can own the IP that buys those ads and you can bypass everything under way, stretch the supplier. You can go straight to the consumer and negotiate with a different agent. That is a massive revolution we have never seen in tech in a very long time. The idea that you don't need any more DSPs. You can just bypass everything goes straight to the consumer on the other side and negotiate with an agent to ask that agent questions to buy and place an ad. So I think that's one of the biggest changes that might happen to -- on the demand side, which affect agencies, DSPs, ethic in general, agents buying from agents versus dramatic protocol that is like an old let's such an old protocol. It's or inefficient. It can take advantage of all the first part of data that's available out there. I think it's going to be duplicated. How fast, I don't know, but I think faster than we can add in. So that's one change.

Maybe before with that change, agents buying from agents, we're hearing that more and more. That's on the, call it, B2B programmatic side, I guess, to the question. Is that happening today? And give us an example of what that could be.

So for us, we announced a quarter ago that we've launched a Claude skill and an a bridge from anyone who wants to buy Taboola and place ads on Apple News on ESPN on Yahoo Mail, like amazing parts of the Internet through Claude code, never talking to Taboola no need to know our dashboard, no need to talk to an account manager. You do need to rely on our reporting everything through your Mac, like open your $200 subscription, open an account in Taboola and tell the Claude, place my ads on 1 sport as we know, is a big category. I want to be on ESPN. Can I do it? And you buy from Taboola. We just made it available and it's already making news of dollars for us. Some people we don't even know. And I think so we're seeing signs of that. So I think it's happening in a big scale, No. But I can tell you all the holding companies, the agencies who spend tens of billions of dollars. That's all they do now. They're all building their agents buying arms that can interact with different suppliers that have unique data and unique distribution. I just don't know how fast it will happen. But look, we're seeing mergers between different agencies we're seeing acquisitions for data purposes. We're seeing actually things are happening. So I do think it will happen, and we're seeing small signs of that as of now, would I mentioned.

And these are -- I mean -- so with the skill that you built, so I've got 2 follow-up questions here. How long it take you to build it, 1 and then to performance advertising, performance marketing. So this expands your funnel, your number of advertisers that could use Taboola do to reach everybody within Yahoo Mail that meets a certain parameter. So talk to us about how you're advertising -- what type of advertisers are coming on with this new skill. So how long does it take using AI in the development cycle to give us a sense? And then who's...

The time to build it was not that long because it still eventually uses Taboola's engine in the back end. So Claude is just talking. Claude became the best account manager, right? It's like if the best accountant we've ever had, and they're just using our system. So it's -- but it's still our engine called Realize, so Realize is our PMA or our kind of ad engine. So Claude is just the best account manager using that. So that's obviously still most of the work is on our side. But Claude just studied all of our documentation. -- everything that was ever reason and then just built all the directories that need to come an expert. So that was a net long. The type of advertisers is still roughly the same. So we're still very good with like mid-funnel kind of mid-sized advertisers. We're making a push towards bigger advertisers towards agencies because we think that's the right time to do that. And I do think still that most of the money that will come through MCP will still come from bigger advertisers and agencies, I think, for a while. For a time, I do want a lot of advertisers -- but I don't suspect that will be most of our growth at least for the foreseen future, okay? We'll have more advertisers but sizable ones. We call them scaled advertisers that essentially means over $100,000 a year, you become someone that has found a product market fit with Taboola.

That's helpful. That was one aspect of AI in the front door. What was the other.

You have very good memory, Ron. The second thing is on the consumer side and the supply side. And I think over there, we're going to see A big part of the Internet as we know it transforming into conversational. So I mentioned use today, we've launched with them in September of last year, deeper dive, which essentially allows anyone to go to use today and ask any question they want about something that is covered by trusted humans that have covered the space. So the Nicks obviously won the championship. I hope you all have the Nicks and otherwise, they should leave the room immediately. But for the second.

We have several Villanova graduates here. We thought.

So it's my son's dream. So when that happened, and I went to it today, and I talked about the Nicks all the time and on of speaking 8, 9, 10 minutes about what the chances we're going to win compared this to a few years before. And I'm doing this no hallucination, just using user today content, I mean -- and then it says, by the way, if you're on that topic, read that article, you should read that and watch that. And it's a way better experience for me because when you love sports, if you love sport. You're never going to ChatGPT, hey, tell me yesterday's score. That's because it's not what you want in motion. You want to watch the highlights. You want to get a point of view of the post maybe New York post, you're part of the community. So to have that experience in the community and talk and watch and interact is such a better experience. So I think on the size side, we'll see a huge surge in different types of companies, making conversation available for the consumers. Look, the Lakers were bought for $12 billion. People believe that communities matter in the world of machines. So I do think that we'll see deeper that, which is like services. I'm not saying we'll be the only one, but kind of like AI layers that offer consumers to do what they are now being trained to do, which is talk to me.

How do you think we access the talk to me? So who are -- is it going to be from one of ChatGPT's latest model? Is it going to be Meta's hatch that we're talking about? Or is it going to be these newer bots that are coming on.

I can tell you how we do it. So we had a bunch of -- when we started that, we've had a bunch of open source models, and then we used our unique data about the consumer leadership on the Internet to train those models and find them and refine them. And then we are lucky because we have a huge division of servers around the world. So from a CapEx perspective, where I think a lot of companies struggle because it's very expensive, and the APIs are very expensive. We were very fortunate we've had the CapEx to support hosting those models on our own servers and training the models based on our own unique data, which made it very affordable for us to make a lot of for free. And the second thing is that how do you make money from that? Because otherwise, AI can just be a cost center. So we show -- again, we're lucky because most of our $2 billion in revenue is direct demand. So we could have sourced the right ad to monetize a conversation. And I can tell you and I spoke about this publicly, when you show an ad that's relevant for the consumer inside a conversation, the CPM of that ad and the value to the advertiser of that ad is 5 to 10x higher -- then an ad on the Open Web on a display in a home page. It is ridiculously valuable for the advertiser. The ROAS is so much better.

Because it's super targeted. And it's exactly what humans are having this conversation .

And I think you -- I think the consumers are having conversations to become super humans. They're like so intentional about what they want to do next. So it's like you prequalify those consumers. You are having 10 signals for the first time. We know Google is a good business. But for the first time, now use it today to get a taste of Google Search business through deeper dive because those signals become available to them. And they use today, I mean it's more than 10 million questions a month. Can you imagine it took to work and you get access to tens of millions of questions of your consumers on your own site. Now you can decide what to cover from a news perspective, you can show as they are relevant. So I think that if you can imagine, 20%, 30% of the Open Web evolved maybe 50% evolving to conversational type of experience, the revenue upside is huge. So much more money can be made and advertisers will find so much value in the Open Web. So I think that's exciting. And we're -- we've always been very close with publishers. So we're lucky that we get to have these conversations, and it's now live on many great partners, Advanced Media -- many great partners globally.

They are leveraging deeper dives on their site. So talk to us about deeper dive specifically because I think there's a lot of discussion out there in terms of all of my conversations might go through let's call it Gemini. One might go through ChatGPT or might go through whatever Meta is doing with Hatch. But with deeper dive, if I understand it's correctable, help me understand a deeper dive and b, then everyone become -- then all these publishers can be that engine conversation. But tell me more about deeper dive.

So Deeper Dive is the first in the world free LLM answer engine supported by ad for the open web. You can be anyone if you work with Taboola, you access Deeper Dive to make an answer engine, Gemini like on your homepage to -- on the ops as today anything. We're using our first-party data to also generate questions for you. So a lot of consumers even know what that enter engine means -- so when you go to use today, if you're a nice because you've been to ESPN, we may say a question we'll say, the Nicks season is about to start in October. Can we win again. And if I click on and first alike, that's so nice, you know me. And if I click on it, I enter Deeper Dive on USA today. It looks like a Gemini entered AI mode. You get the short answer. You can say I like it or I didn't like it, you can follow up and then Deeper Dive says, while I have you, check out this video and news for today about the mix. Checkout is article just look at this point of view about why we won and why we may win again. And then I read it -- and then the Deeper Dive, again, this is no lag for this article, but this question, maybe you want proximately more. And that is a completely new experience. It's free, and it's generating revenue every single time someone is using that. The adoption rates and you know the web really well on. I mean, if you look at companies over the last 20 years, click-through rates on things online, like what is the likelihood I will click on an ad, what's the likely that will click on an article tends to be 0.5% to maybe 1%, 2%. Deeper Dive is being used by more than 10% constantly. So it's a double-digit adoption, which is -- I've never seen anything like that. And the CPM, like I mentioned, could be 5 to 10x. A single time ad is making such a great CPM, which is also why Google is not only they're not concerned about BlueLinx search. I'm convinced they just want search as we know it now to die as fast as possible and all move to Gemini because they probably know what I know now, which is ads on in conversations are worth a lot more than anything else. So that's really is now. And then our -- obviously, advertisers are excited to be part of LLM. And about a month ago or 2, we got a lot of companies coming to us and say, "Can you help us monetize our AI because we don't have your ads, and it's very expensive. -- and we kind of launched deeper dive network -- so if you have your own utility up and you're making it, you have a commerce AI app that you raise money for from investors. -- and you want to make money, you can come to us and we'll show ads on your own LLM. So that's new.

And talk about the implementation. So how long does it take? Did it take or as an example, USA today, to install it because essentially, what we're talking about is natural language search and conversational search across every single site that most every single company is trying to do. So talk to us about the implementation process maybe I'll back up. What gives table the right to win here? Then talk about the implementation process and then where are we on adoption rate. of your customers.