Senior Media Analyst and Internet Analyst at Needham & Company, and I'm here to welcome Stephen Walker. I think this is the first time. Isn't this the first time I've ever interviewed you alone on stage?
That is correct, yes.
Yes, the first time.
Thanks for having me.
Ever. Yeah, thank you.
Be gentle.
Okay, no, that's not a thing. That's not a request I'm willing to abide.
All right.
Nice of you to ask. How's your one-on-one so far?
Very good. Yeah, it's been a good conference.
Oh, fantastic. Great. Thank you for being here. I think maybe to level set between the people on the live stream and in the room, why don't we start with what Taboola does? You're the Chief Financial Officer, so why don't, at the end of that, you give us what your 2025 projections are, growth rates, margins, stuff like that.
Okay, sounds good. No problem. In terms of what Taboola does, we're a leading global performance advertising platform. What that means is we help businesses grow by basically helping them place their ads, performance ads, across the open internet. That includes thousands of websites that we work with, OEM, phone manufacturers, apps, so high-quality supply. We help basically the performance advertisers to achieve very specific goals: cost per acquisition, cost per lead. We do it for them. Our advantage is that we have very unique data. We see 600 million people a day, every day. From that, we learn what they're clicking on, what they're reading, what they're doing. We're able to use that data with our deep learning AI engine to basically target ads to them that will convert and help those advertisers achieve their goals.
For the publishers, we help them with revenue and growing their business as well. That is Taboola in a very simple nutshell. In terms of 2025, this year is going to be a strong year for us. We expect to achieve about $1.9 billion of revenue. Our kind of mantra is that we want to achieve 30% EBITDA margins, and that is our target this year, about $211 million to $212 million of adjusted EBITDA. From that, we want to convert 50% to 60% of it to free cash flow. We have said that we expect to be at the high end of that, so you can expect us to be closer to the 60% end of it. That is our goal this year.
One other big thing that I'll talk about, since I'm sure you're going to ask me about it next anyway, is we just announced a new platform called Realize. This is our new advertising platform, which basically takes us from being a native advertising business into being able to work with performance advertisers across all types of media, all types of formats, all placements on the page. They do not have to be native placements; they can be display placements. It does not have to be native ads; it can be display ads or vertical video ads. Whatever they want to work with, we'll work with them on it, applying that same kind of unique data and same unique deep learning engine to target the ads effectively and help them achieve their performance goals.
One of the things I run into with Taboola is that you guys are talking about this. You've sort of saturated the native market, so now the total addressable market is larger because you're going after display ads, which are like a three times larger ad spending part of the budget. Why don't you explain to the audience, what is a native ad? How is it different from a display ad?
Sure. You know the format is one thing, but it's also the objective of the advertiser. Native advertising has grown up to be performance-oriented. Again, performance ads means you've got a very specific cost per acquisition or cost per lead or some numeric goal in mind. You know what that goal is, and you know how to measure it. That's as opposed to, say, branding advertising, where you may have goals, but your basic goal is to get your brand known by people. You're not trying to get somebody to give you their email address or to sell a product right then. Performance is very specific to very specific goals. Basically, the difference, native has grown up as a performance advertising. Display has actually grown up to be more of a branding business historically.
A majority of display ads historically were to build a brand, to get awareness, to get your name known. Now it has moved over time to be more performance, but we see a big gap in the market. That gap is that there's not really a player out there who will help you to achieve whatever your goal is, CPAs, CPL, whatever your specific goal is, in a very automated way with unique data and unique assets to help you to do that. For instance, if you go to your standard DSP, for instance, and you say, "I'm going to try and buy ads on the open internet, I'm going to try and achieve a performance goal, a CPL or a CPAs," it's kind of up to you to do it. DSPs are great software platforms.
They let you bring in different data, connect it, it lets you target your ads in different areas. Really, it's up to you, the advertiser, to figure out what works and what doesn't. What's different about Taboola is you come to us, you give us those same creatives, the same formats, whether it's native or a display ad, which is more of a graphical ad, or a vertical video ad like you see on Facebook all the time. It's our job to figure out where to place those ads to achieve your goal. You don't have to be turning the knobs and flipping the switches to figure out what works and what doesn't. That's our job.
We think the gap in the market is that there's a need for somebody who is going to just help a performance advertiser to succeed in the open internet the same way they do when they go to Facebook or Google. If you go to Facebook, Facebook doesn't say, "Hey, which users do you want to target this ad? And what traffic sources are you willing to take?" and things like that. They just say, "Give us your creatives, and we'll help you achieve a goal." Same with Google. "Give us your creatives, we'll help you achieve a goal." We want to be that for the open internet.
One of the learnings that's been on the stage today is I would say three out of the, I think you're my eighth, you might be my seventh, are guys introducing products that cross from the open internet to walled gardens. Their big thesis is increasingly CMOs want to optimize their spend across all of digital advertising. Their complaint about the walled gardens, even though they have best-in-class feedback loops, is they, meaning the CMOs, Chief Marketing Officer, can't compare Google's to Amazon's to Meta's.
CEOs in the open internet that have products that cross into the walled gardens, they're finding it easier to bring in business because they can compare their open internet metrics against not only just against each of the walled gardens, but across the walled gardens themselves, which gives a common metric, which is super valuable to advertisers, even if they're not really using it for the open internet guy. They're using it to compare the effectiveness of their ad from Google to Meta to TikTok. Talk to me about that, because you basically just laid out a product vision that stays squarely in the open internet.
Yeah, and it's interesting because I think, by the way, that's a great, I think that's really smart of these companies who are doing that, because I think that is, if you're a performance advertiser, at the end of the day, what is it about? It's about achieving a very specific goal. Knowing how you're doing on that goal on Facebook versus Google versus the open internet is very important for sure. I think for us, and by the way, maybe we get there at some point, for us, I think our thesis is that the gap that we're trying to address is the fact that even just on the open internet, it's really hard to figure out how you're doing against that goal.
It's even harder to get to that goal because, like I said, standard DSPs provide tools, but they don't provide a platform that does it for you.
Okay. So I mean, one of the benefits you guys have, Steve, is you have the best of the open internet, sorry, you have the best of walled gardens because you are a DSP and an SSP. You sort of have that wonderful feedback loop that the walled gardens have, but you have it in the space called the open internet, which is incremental reach to any of the individual walled gardens. I guess my question is, when you think about this, there's something going on in the open internet for independence, where there's DSPs, is we're moving data to the SSP level, which feels like it's moving power away from DSPs towards SSPs.
My question is, does that, what's happening at The Trade Desk magnate level, which are independents, they only used to have one side, and they're now sort of entering the middle, does that affect you guys at all as an end-to-end open internet platform or not really?
At the end of the day, we're all competing for ad budgets. It affects us in that way because we have to show why our solution is superior to the solution that they can get from those platforms as well. For sure, it affects us in that way. I think, as you said, we have a couple of advantages. One is we're already end-to-end, so 90% of our revenue or more is with advertisers we have direct relationships with, and 100% of our supply, not 100%, 90% of our supply is also kind of locked in with publishers that we have long-term agreements with. Generally speaking, we're already end-to-end, so we're ahead of a lot of those players who are trying to get there.
The second thing is, I think we have a very unique set of data because we're hard-coded on the publisher pages, or in the case of working with an OEM, the OEM manufacturer gives us information, data that we can use to target users. We have a very unique data advantage that I think a lot of those players are still trying to build. I mean, again, it affects us because we need to keep an eye on it, make sure we keep our lead, but I think we're ahead of them in terms of both already being fully end-to-end and already having a data advantage that's going to be hard for them to replicate quickly.
Okay. Okay. So you have a unique data advantage. Do you have any ability to sell the data, the unique data advantage you have as a new revenue stream to LLMs on an aggregated basis where you do not break it out publisher by publisher?
Yeah, without getting into contract specifics, we have the right to use the data to provide our services. Could you interpret that as we sell it to an LLM? Maybe. Would we ever do it? No, because at the end of the day, that is our unique special sauce. If we give that data to somebody else, then the advertiser does not have to come to us to get it. I think as of now, we want that. Like you said, we are a little walled garden like that. We want our data to be our data. Let the advertisers come to us to get it. By the way, with Realize, one of the things about Realize is we are not just bottom of article native anymore. We are going to be in display spots. We are going to be bidding on spots in apps.
We're going to be bidding on spots across the open internet. We want that data to be proprietary to us across that whole spectrum. If we give it away, then they don't have to come to us for it.
You would charge for it.
Of course, yeah.
Okay. CTV. Let's talk about the future of CTV at Taboola.
Okay. This is where Adam needs to be here, right? CTV is interesting. CTV today is primarily branding. Now, I know there are a lot of companies out there, including one who I think is trying to get public, who are pushing performance through CTV. Generally speaking, my understanding is the way they are doing that is they are connecting CTV ads to something else, either a user that they can then say, "Hey, that user bought this product later on," and try and connect it in a performance way. Very hard to do. We do think there is an opportunity for us at some point to do that connection.
If a company is getting video ads and they want to then drive down funnel and get a lead or get a conversion from it, we think there's ways that we can connect our data with that company and help them go down funnel. Use the CTV for what it's good for to get the awareness, and then we'll help you go down funnel from there and get a lead or get a conversion or get a purchase.
Are you happy selling your data off your platform then? I just asked about selling to LLMs. You said, "No, we want to keep it proprietary." Wouldn't this answer you just giving me be using data to monetize the data in a different way off your platform?
We would still want them to come through our platform to do it. We would not be giving the data away. We would want them to spend the advertising dollars with us to help go down funnel with our supply and our demand. Actually, it would be our demand and/or the, so if we are working with a broadcast company, for instance, the broadcast company could say, "We will get our advertisers to work with you too, and then you help them get down funnel to leads and conversions.
Okay. Okay. Okay. I mean, I guess I'm just trying to think of new revenue streams that don't involve your DSP and SSP that opens you up to revenue from off platform, which is a lot bigger TAM than your platform. Yeah, that's all. Actually, somebody was on my stage, and I'm forgetting who, but was saying that for Outbrain or something, they have unique supply like you do, but they don't have enough demand, which a lot of these guys say, and you guys are saying this too, you want more demand. In order to generate more demand, they've gone to, maybe it was The Trade Desk that was staying there. I'm trying to think if there was another DSP. They've said, "The Trade Desk, why don't you bring an incremental demand to our supply?" Have you guys done that yet?
Yeah, so we.
I think we've done this too.
Yeah, we allow programmatic bidding into our platform. In fact, programmatic bidders win about 10% of the time on our platform.
Over your own bids.
Over our own bids. I mean, it kind of shows the value of our data, though, because these are some very sophisticated advertisers. They're winning about 10% of the time, and the rest of the time you need our data to win on our platform. I think in terms of getting incremental demand, our focus, we talked about this at our investor day. We think there's three ways we grow the business from here. One is we will bring in other forms of demand. For instance, we have an import your social campaigns button on our platform now. We think that we can show advertisers that the diminishing returns they're getting from Facebook, we can help them achieve better performance, at least for the tail end of their budgets. That's one thing, new formats, new types of demand that we'll bring on.
Second thing is we also are, we're big enough now that we've been able to verticalize our sales team and really focus on kind of an optimized go-to-market. We think that also will help us differentiate from smaller players who may not be able to be kind of experts in the health industry or experts in certain verticals. Third, I think we continue, you saw us after we signed Yahoo last two years ago and then launched it last year. We signed Apple after that. We just announced a partnership with a company called Line, which is basically WhatsApp for Asia-Pacific. I think our other kind of focus now is let's go get forms of supply that is, again, unique to us, that comes with unique sets of data that will help attract that advertising demand.
Because at the end of the day, the ad dollars follow the supply, and they want to be on high-quality supply. We're the only company who can say you can be native on the Apple iPhone other than going to Apple directly. It is things like that that I think will help us also get that next level of growth and bring in more demand.
Question about all of your supply is unique to you because your code on page. So what happened? We had all this unique supply, but it turns out it's not as valuable as we thought, so we're going out and getting different forms of supply?
No, it's very valuable, but I think what we found is that the bottom-up article native is fairly saturated. It's hard to find a publisher website who doesn't have some form of native advertising at the bottom of their page. Therefore, there wasn't supply expansion. We are continuing to grow kind of our revenue per ad spot, but that's dependent upon the ad budgets. What we found is, and this actually was a big learning that came from Yahoo, we found two things. One is there are a certain set of advertisers, like the ones we brought over from Yahoo, like Intuit and Hulu and others, who we thought, "Okay, we're showing great performance on Yahoo supply.
You should advertise on the rest of our bottom-of-article native." What they said is, "No, thank you, because we like, even though we're a performance advertiser, we like 100% share of voice. We don't like to be in that stream of other stuff around us and ads and things like that." We said, "Okay." The second thing that we learned is that there's another set of advertisers. If we have Hulu, we went and talked to Netflix. We said, "Look at the great performance that Hulu is getting." I'm making up Netflix, but it's an example. Netflix, they said, "Yeah, but I just don't want to learn how to do native. That's not something I do. It's not part of my skill set." It was interesting to us.
We learned that we actually had to be a bit different in order to get those big brand advertisers to work with us in some cases. That is part of the core strategy of Realize. Realize is going to open us up to different forms of supply. We will bid on display spots, which are 100% share of voice. We brought in Apple, which is 100% share of voice. Actually, Yahoo is good because it is 100% share of voice. We are bringing in different types of supply that now appeals to those advertisers, and we are not requiring them to learn native. We will tell them, "Okay, you have got social campaigns that work? Great. Give us those. You have got display ad campaigns that work? Great.
Give us those." What we realized is we needed to kind of break out of this bottom-of-article native in order to really bring in those incremental budgets that we need to continue to grow at the pace we want to.
Okay. All right. Okay, that makes sense to me. Okay, so then when we think about the new Realize, one of the things I just need to tell the audience is the best thing about Realize, it's a new platform. It's a brand new platform, and you guys changed it over a weekend. Over at The Trade Desk, they're trying to get people, the people who use their platform off of Solimar onto their new platform that's already been out for maybe 18 months called Kokai. It's really given them a lot of trouble. They won't be able to get that done until the end of the year this year. It's going to have been 22 months or 24 months. You guys did it on a weekend, and you shut down the old platform.
So the cost savings of that are meaningful if you can just transfer everybody over on a weekend. Does that mean the changes you're making are less dramatic?
First of all, I do not know exactly what the Trade Desk is. I will not compare to them. I think the changes were fairly substantial. We basically took a platform which was all geared towards native ads. Give us a thumbnail, give us a title, give us attribution text, and was geared towards you are going to show up in this feed at the bottom of the article. We turned it into a, "Give us whatever format you want. Oh, by the way, we are going to put it in a bunch of different spots." I think it was a pretty dramatic change. It was risky. I mean, it was a risk we took saying we thought it was better to do a full conversion flipover rather than try and do two platforms and migrate.
It was risky, but at the end of the day, we think it worked out well. We do have our native advertisers have stayed with us. We now have the ability to upsell them on other formats and other placements. We are now open for business for those people who said, "I don't want to do native." It was risky, but I think it worked out well.
Did you just create tabs? When they go into the native tab, it looks like what they did on Friday, but they have to go, there is just a homepage that now gives them other offerings. You can watch sports or you can watch horror. You can watch like Roku does where they have these tabs.
Kind of. I mean, it's not really a tab format, but yes, we tried for the native advertiser who were used to the platform, we tried to make it look and feel as much like they were used to as possible while still having this extra functionality built on it. We did completely rebuild the platform. The interface is different. It's much more modular. It's something that we now feel like we can really build upon as opposed to this kind of legacy native advertising platform, which would have been harder to kind of keep building features onto.
One of the things that where Meta tells us they're going, and I'm curious as to whether since you guys want to be sort of the open internet walled garden equivalent, is this where you're going? Meta's saying, "Look, we want a performance advertiser," and they are really 2 million advertisers. They are small and medium business, apparently. They want to put in a budget, and then Meta, the Gen AI, will go find all their assets that are IP protected. They will make the content. They will then, let me call it, customize the content. They'll go out, they'll put the ad anywhere that Meta goes, and then they'll get that return. Let's say you say, "And then all they want is budget.
You give us the budget," and I don't even think they really want budget, but you give us a minimum return, so $4 for every one you spend on Meta. Then let's say you do a maximum $1 million budget. You hit go, they do everything else. They make the ad, they deliver the ad, and all they do is then give you the feedback loops with attribution that's only creating their own homework, but only on them. Is that where you guys are going to end up going as the walled garden of the open internet?
You know, I think, well, we've certainly followed Meta and Google in terms of how they do things. A lot of this is, "Give us a budget and we'll figure it out for you." I think we'll end up somewhere close to that probably over time. I mean, you've seen some of the advances that we've made. We have a tool called Abby, which, yeah.
Self-service.
Self-service. Basically, it helps advertisers come in and just get set up. It's a generative AI assistant who talks to you and says, "What's your goal? What do you want to do?" We do have an ad maker, which will actually make the ad for you for native today, not for display yet, but I'm sure that'll be coming. Yeah, I mean, we're headed that way. I think we'll end up in a similar spot over time, probably. Certainly, we want to make it as frictionless and easy as possible, just like Meta does.
I mean, historically, your captive advertisers, which you say is 90% of demand, have been very large companies, not Meta-type advertisers, much larger companies, which is why we thought Abby, which is your self-serve Gen AI product, which in three steps launches an ad campaign, which is very close to what Meta does for a living. We really thought it would open TAM, but it's just been hard to get awareness and drive attention.
We're working on it. We're seeing some progress. We actually have a team now who is, historically, we just said self-serve, not for us, because before generative AI, the idea of the complexity of getting an advertiser up and running was just too much. With generative AI, we now see a path. With Abby, I think we can see ourselves getting there. We put a team behind it. They're starting to work. We haven't disclosed any numbers yet, but we're starting to see some traction there. We're starting to see some growth. We're even putting some of the advertisers who are at the smallest side for what we call our growth team, which is our smaller advertiser team, into self-serve because we think Abby's good enough that Abby can help. I think we're getting there.
Abby needs to get better at not just launching, but also helping to optimize and upsell and things like that before we'll be ready to kind of really turn it up. We are getting there.
Okay. Okay. That's interesting. So basically, I can follow Meta. I can listen to Meta, and that's where you're going two or three years before.
We do not just copy, but yes.
Yeah. Okay, fair enough. Fair enough. Let's talk about costs. You guys have guided to long-term adjusted EBITDA margin expansion. What are your biggest levers to improve profitability? Is it Gen AI, tech efficiency, OpEx rationalization, publisher terms, revenue mix shift? Is it a cost advantage over your ad tech competitors that so many of your engineers are sitting in?
Tel Aviv.
In Tel Aviv, yeah, in Israel.
Yeah. That's a lot of questions all in one.
Yeah, it is all about costs. All about costs.
Okay. First of all, the biggest thing we can do for margin expansion, frankly, is more advertising budgets because it is really about if we get more advertising budgets, our XTAC margins, so kind of our top line margins, will expand. That is like free money for us because you do not have to do anything to support additional XTAC necessarily. There are small things, but that is the biggest thing. More advertising budgets will expand margins at the top, and that will expand margins at the bottom. Beyond that, we are focused on our President and COO, Eldad, who you have met at some of our investor days and the like. He is laser-focused on using generative AI throughout the organization and really trying to get our teams to embrace it, use it. When somebody asks him for incremental headcount, he says, "But are you using generative AI?
What could generative AI do to do that for you?" It's funny, he pushed me on it for our FP&A team. As I started thinking about it, I'm like, "Wow, you know, you're right because Google Notebook LN, it's like it's amazing. You put a bunch of documents in it, including spreadsheets and stuff, and then I can just go to it and say, 'Hey, how did Taboola News grow over the past five years?'" It will give it to me.
Does it chart or on a table?
It'll do it however you can tell, put it in a table. It's amazing.
Do you have to talk or do you have to type?
Yeah, you type. But yeah, and then you can even say, "Oh, but actually I wanted that quarterly. Oh, but give me the year-over-year quarterly growth rate.
That's using Gemini? It's using Google Store.
It's using Gemini, yeah. So it's a Google product. But it's like it just shows like, and I told our FP&A team, I'm like, "Look, I know you say you're overwhelmed, you've got a lot going on, but you could stop having questions for me if you just get all the data into the right place, and I can do it myself." It's eye-opening how you can do that. That's like an admin function. Now, if you talk about the R&D function, if they're not having an AI code for them, they're doing it wrong, right? There's just amazing levels of productivity that's available. We will, we think, be able to kind of at least hold costs and kind of keep costs lower over time thanks to AI.
Obviously, we've always talked about the fact that we'll continue to get more leverage out of GNA. I think even our sales teams, we think, can become more productive with generative AI and also with self-serve and other things can have some effect there. I think we'll see productivity gains across the board. I think actually it's kind of an exciting time because I think the economy in general should see productivity gains if people really embrace generative AI. We won't have enough power to run it all, but it's going to be amazing.
Down the U.S. Mark Zuckerberg is saying that by the middle of next year, he expects the coding excellence of Llama 4, which is his open AI model, to replace mid-level software engineers. They're on the cutting edge of innovation. If they can replace mid-level coding engineers, like everybody else can replace their frontline most innovative, I mean, that's amazing replace. That's in a software-first business. I don't know if you guys think of yourself as software-first probably.
Yeah, I mean, that's basically the way we think about ourselves, yeah.
Really interesting. Yeah, Gen AI changes everything, I think, but okay. Free cash flow. How should investors track the bridge from EBITDA improvement to free cash flow growth? Are there timing or structural differences investors should be aware of? Working capital needs, CapEx, earnouts, etc.
Yeah. So we've said fairly consistently that we think that our adjusted EBITDA should convert into free cash flow at about a 50-60% rate. We continue to say that. In fact, actually this year we're saying we should be at the high end of that. In fact, over the last eight quarters, we've actually converted at about a 74% rate. So we've been doing a fairly good job of that. This year we have a little bit more CapEx because we're pushing to be back in office more. So we're doing some office upgrades. It'll probably be $40-45 million instead of $35 million last year. We actually think we can offset that with some other working capital gains. Net-net, I think we're at 50-60% plus, and hopefully over time we can kind of continue to push that higher.
Okay. Okay. Cool. Yahoo. What has surprised you most about Yahoo? It feels to me like you ended up being a big supply integration, but now demand sources that we hope would come from Yahoo haven't materialized, which has lowered your overall auction prices, and you gave up 25% of Taboola's shares to Apollo to get the deal done. What's gone right and wrong over the past two years related to the Yahoo deal?
Yeah. So first of all, we're very happy with the deal. I mean, at the end of the day, we went from, we more than doubled our adjusted EBITDA, we more than doubled our free cash flow last year as we launched Yahoo. Yahoo had a very profound effect on the financials of our business, and we're very happy with the deal. Having said that, nothing ever goes exactly as planned. As Mike Tyson likes to say, everyone has a plan until they get punched in the face. What went well with Yahoo was we transitioned all the supply over, and as we expected, our advertisers love the supply. It's amazing, unique supply. The way they've laid out their websites and everything is very attractive to advertisers. It comes with very unique data. Their users are heavy users.
They're back all the time, which gives really unique data insights into their users. The supply is unique. It performs really well on our platform. The supply has gone really well. We have transitioned a lot of very big brand name advertisers that, frankly, are of the type that historically we didn't have. Like the Hulu, Samsung, Intuit of the world, those are advertisers that just hadn't worked with us historically.
I thought you said they didn't come because they wanted different things. They wanted to swap.
No, they did, but they're only spending on Yahoo. They continued, and they were previous customers of Yahoo for Yahoo supply. What surprised us, and that's actually what hasn't gone as well as planned, was that, one, some of those that we thought would, why wouldn't Intuit then advertise on the rest of our network? They didn't want to be in the feed. The second thing that didn't go as well as we thought is we thought once we had the Yahoo supply and we had the proof that it worked for these other big brand name advertisers that we could get people like Netflix and others to come along. That hasn't happened as quickly because they don't like native. That's why we talked about earlier, that's what Realize is about, us solving those things.
The one other thing I would say that's gone well with Yahoo is it's really been kind of a flagstone deal for us or a.
Flagship deal, cornerstone deal.
Cornerstone, flagship, yeah. I'm going to mix those two. I'm going to call it a flagstone. Yeah, it's been a cornerstone deal. We probably wouldn't have gotten Apple if it weren't for the fact that they saw us doing Yahoo. It's leading to Line also took notice. That's the Asia-Pacific version of WhatsApp. Companies took notice, and I think it's really opening up doors for us that probably wouldn't have been open without that deal.
Like the brand had glitter value or something.
Yeah. Basically, I think a lot of companies kind of went, "Oh, wow, if Yahoo is going to outsource that whole native business to you, maybe Taboola kind of knows what they're doing, maybe we should talk to them too." Remember, Adam would like to call it advertising in a box. We now have, for any company who wants to get into advertising, we'll help you on the performance side in particular.
Okay. Cool. Even your display is going to be tied to performance.
Yep. Our display. Now, with Apple, for instance, they sell their own display when it's branding, and if it's performance, they send them to us.
Gotcha. Okay. Let's talk about capital allocation. How do you think about capital allocation, like debt repayment versus M&A versus share repurchases for investment and growth? Has this been changing over time?
Fortunately, we're in a position where we generate enough cash flow that we can fund kind of the investments we want to make in our business from our operating cash flow and still generate last year almost $150 million of free cash flow. We don't need money to do what we want to do in terms of investing in our business. What do we do with that excess cash flow? Right now, it's pretty clear our number one priority is to buy back shares, and we've been doing that very aggressively. We bought back over $90 million of shares in the first quarter and a half of the year. We're going to continue to do that. That's our number one priority. We will look at M&A, but probably more tuck-in type of M&A. We're not in the market for anything big.
Of course, never say never. If the right thing came along at the right value, you got to look at it. I think for now, we're thinking it's probably more tuck-in. Historically, I would have said, and we'll look to pay down debt over time, but we actually just got done with a debt refinancing where we took on a $270 million revolver, paid off our fixed-term debt, and this revolver is at much better terms from an interest rate perspective. It allows us to take advantage of our working capital swings so that our average debt balance will be lower. I'm not at this point in any hurry to take our debt balance down. We think buying back shares should be the priority.
Okay. Do you feel that after stock option exercise, you are actually buying more shares than account for the stock option exercise? You are actually shrinking capital base?
Yeah. So we're actually down from something like 340-something million shares to 336.
This morning I had breakfast with Roku, and they asked for the meeting, and their entire thing was, "Laura, you're measuring productivity wrong. Here's how you need to measure it. Next quarter, I will not be using revenue per FTE, which is what I've used for two years now. I'm going to try theirs," which they believe will capture this issue of FTE productivity and also the share dilution. Because so many companies like Google and Meta are constantly diluting their share bases, and revenue per FTE does not really capture the share dilution from stock-based comp.
That's interesting, yeah.
We'll see how it turns out. I have no idea. You'll see it in print when it shows up whenever I get to it.
Great.
Anyway, let's talk about Gen AI because I think everybody in this space being ad tech. Let me go to questions. Any questions from the audience? Okay. Let's go to, I mean, these are ad tech. Tech is in the name. How are you guys using, I know you're using a lot of Gen AI and a lot of it, if you're really moving towards this meta model where more and more is going to be automated, that's all Gen AI. I got to tell you, the most interesting meeting I had with Taboola in Israel was the Gen AI guy that came in because there's a specialist you have at Gen AI. He's talking about a worldview that I never even visit. You and I think of, you think, you and I think of quarters, years, maybe multiple years, quarter.
He's thinking, anyway, so tell me what you're doing in Gen AI and what the path is and how to integrate it into driving revenue or cost savings.
Got it. Actually, I'm going to speak more broadly about AI holistically because I think it's also the biggest impact we can have on our business is continuing to develop our deep learning AI, which is how we take the unique data we have and target ads to consumers. I think that is actually something that we continue to invest a lot in. Max Conversions, the bidding algorithm that we released last year, was an advancement on that. We have another.
Which is automated.
Which is automated, exactly. We have another iteration called Max Revenue coming out, and we'll continue to iterate on those because, frankly, that is the number one leverage point for our business. If we can get yield, basically, if we can get more revenue per ad spot and have the advertiser happy, that's money in the bank for us.
What do you call Gen AI?
That is, well, Gen AI, it's AI. It's deep learning AI. It's a distinction. Adam will talk to you about it. But so it is AI. And then on the Gen AI front, we've got two opportunities. One is make the world dramatically easier for our customers. That's where Abby comes in. That's where the ad generator comes in, or ad maker. That's where it's Meta's vision, right? It's like, just do it for them. Don't make people think. Do it for them. We are working on a bunch of initiatives in that area to make it dramatically easy to use our platform. The second one we talked a little bit about earlier is internal productivity. We do think that this is something that you can use to really dramatically change the cost structure of your business over time or increase the productivity, which ultimately means margin.
Ultimately, we think there's a lot of opportunities in that front that we talked about earlier.
Okay. Great. So we're using AI, these deep learning AI, which sounds like it's a midpoint between old AI and generative AI because generative AI has to sort of make new things. Old AI just took a bunch of data, but didn't, I don't know what it did with it. And deep learning AI makes better algorithms, which makes better recommendations or better targeting.
Right.
Okay. Got it. Okay. You're probably using all. One of the things Wall Street believes is that everything's going to be tied to purchase, not just outcomes. Is that too simplistic? Is that not a trend you're seeing? You think performance advertisers are staying at other metrics other than purchases?
I think they're on to something in that I think everything is moving down funnel, and I think everything needs to be, that's why people are talking about CTV for performance, right? Because everything needs to get tied to that end result, which is somebody buys a product or a service or does something that makes a business money. You can't skip directly to it. That's the challenge. A big part of our business is cost per lead. It's Wayfair trying to build up their email list for their home furnishings newsletter. They pay us to help them get email leads. That is their first step towards then getting somebody to buy. It takes consumers a while to get educated and to learn what they want.
If you get them on a newsletter and you start telling them about why the outdoor fire pit thing is the best thing since sliced bread, then you can get them to buy it. If you just drop an ad on them, say, "Buy this," it is kind of like, "I do not know why I need that." I think there is a journey for most consumers. I do not think you can skip to the bottom, but I do think the end result needs to be the bottom at some point.
Okay. Okay.
By the way, I think that's where we're somewhat uniquely positioned because we can go from that mid-funnel helping to get the leads, helping to get awareness. Historically, we were called content discovery because what we did is we helped consumers read content about products and services, which helped educate them, which helped them on that journey to the bottom of the funnel. Now we also have, we bought Connexity, which allows us to now get all the way to the bottom of the funnel and get actual sales for the retailers who want to sell the products and services.
I would say that's been my biggest surprise. Any questions from the audience? Okay. How is Connexity? So much silence around Connexity on the last couple of calls. What the heck is going on with Connexity?
You know, it's funny because the reason for that is because we're now, we verticalized our sales force. We now have a finance vertical. We have a health vertical. We have an auto vertical. We have a travel vertical. Frankly, what we've decided is that really e-commerce is just a vertical. It's a unique set of advertisers, which happens to be retailers, who need a specific type of targeting to achieve their goals. The goal in this case is a CPA. It's a cost per action or, yeah, exactly, or acquisition, exactly. Ultimately, we've started thinking about it that way internally. I think that's why you're not hearing us talk about it as much. We will talk about it as one of our verticals and how it's doing, but not as a standalone business per se.
I would have guessed it actually had applicability to finance, health, travel, and autos, but you're saying it doesn't. It really only focuses on retail.
No, because the whole Connexity business was retailers: Walmart, Target, Wayfair, and others, but it really was helping them sell product at the end of the day. It was not helping Citibank sell credit cards, for instance.
That's what I'm saying. It seems like it would have had applicability to these other sites.
We've certainly learned from how they did business to be better at it ourselves. I think as we verticalize, they are a very mature and well-established vertical that the other verticals can learn from.
Okay. When you look at that e-commerce, let me just call it Connexity, even though it's like a vertical, what percent of your total ad budgets are that? You think of finance, health, travel, autos. If you think about all the money that's spent, how big is e-commerce as a % of that on your platform? Is it 10% or 2%?
No, it's roughly 20-ish. Yeah. I mean, it's a big vertical.
If it's 20 each, then something has to be smaller. If it's 20, what's?
Not all of those are 20. I mean, and I don't have the numbers in front of me, but I believe that finance is probably close to that. The others are probably a bit smaller, but again, I don't have the numbers in front of me.
That's because finance and health, I totally understand finance and health because they have so much regulation that they're really high-value-added silos. I don't understand travel and autos. Is autos just because it's so big?
Yeah. I mean, the way basically, if you were to go back to the early Google days and look at the search and where did they get the highest revenue per search, it's those same verticals. The reason for it is because if you get a new customer for a credit product, it's high value. If you get a new customer for a travel plan, it's high value. If you get a new customer to come buy a bag of potato chips, it's not as high value. It's really about the.
A lot of them, that's why they're 20%.
Exactly. Yeah.
Whereas the other ones might be five times the CPA or CPO because the leads are so elevated.
Exactly, yeah.
Okay. Let me see how we did on time. Oh yeah, we're right on time. Okay. The only one we didn't do is what's Wall Street missing? Three key takeaways you would like Wall Street to take away from this fireside chat about why the stock's undervalued.
Yeah. So hard for me to say what Wall Street broadly is missing or not missing. I think our sense, though, from talking to investors and kind of listening to their questions closely and trying to get a sense of what they're thinking is we're in a transition right now. We've admitted that the native business was growing too slowly and we can't just be a native business. We launched Realize. We're expanding our TAM. We're going into new markets. I think we're now a prove-it story to a lot of investors where they want to see the traction in that before they're going to jump in. I think if we can show them that and we can show them that the growth is there, then I think we'll get the valuation that we deserve. We got to prove it, I think, at this point.
Okay. Great. I will call it there. Thank you so much, Steve.
Thank you.
For your time.