Over 800 global brands are now trusting Zeta with their data, and we merge it with our data, and that's where the models really learn.
So for the next three to five years, where do you really see Zeta going?
We're going to build one of the defining companies of the software.
All right, let's get to it. David, I have three pieces of paper here that I wanted to present to you. Do you mind telling me what these are?
Well, Mitch, they are 8 by 11 sheets of paper. No, I think this is our last three earnings reports, it looks like.
Yes, and that's exactly right. I have 17 more right here. 20 straight beat and raise quarters. Simply remarkable. And on behalf of not only the long investor community, but I'm sure for every investor out there, we just wanted to say thank you.
Well, I will say we leave it all on the field. We really, really work hard. As you can hear, I'm losing my voice. I showed up because I was so excited to see you. In the last, I think, seven days, I spoke at Goldman Sachs, met with 36 institutional investment groups, gave the keynote at the Palantir quarterly conference, have been on the West Coast, East Coast, West Coast, East Coast.
Just in the last seven or eight days. So you've been busy obviously, and we'll get into, you know, Goldman and the AIPCon, but I just wanted to ask from an operational standpoint, clearly that's not an easy thing to do, obviously 20 straight beat and raises. What does that look like, maybe from your vantage point, and then day-to-day in Zeta around the company?
Well, let me be very clear, this is just having an exceptional team. This is not me. I mean, if you look at what Steve Gerber and his operating group have built, they have built just a team that wants to win. And you know, it's funny because what you don't see in these reports is some quarters we're so far ahead that it's sort of you coast to the finish line, and some quarters you're playing from behind. And in all cases though, the team has come together and sort of won together.
Okay. And that takes... we have our senior leadership team, which is I think 64 people at this point. Everybody's really working incredibly hard. And what I always say, and this is what's really interesting, Mitch, is that there's a lot of great ideas out there. Executing is what separates the winners from the losers. And what I'm most proud of is the consistency of our execution. It's also easy to execute for a short period of time.
We've all seen that, right? Like we have friends who they would go to school and they'd be great for like some part of school, or then they'd be bad. We'd see companies that execute really, really well. It's hard to execute consistently. And that's one of the things I'm most proud of the Zeta team and our Zeta people for.
Yeah. I mean, again, for five straight years, 20 straight quarters, it's not an easy thing to do. And so I just wanted to pivot. You've come back from the back-to-back Goldman and AIPCon conference.
Say that three times fast.
I just say Palantir Conference. It's too hard for me to say.
And you mentioned that Zeta plans to make three of the largest company announcements on October 8th.
It actually might be four.
Okay. Well, here you go. New news. Obviously, we can't get into that, but part of which allows you to execute is having a plan and a roadmap. So for the next three to five years, where do you really see Zeta going and what this company can become?
I'll have to be careful here because a lot of what we're announcing is the five-year roadmap.
Okay. But what I would tell you is that if you look at our major pivot from being a marketing platform to an intelligent AI infrastructure company, and you look at the launch of our first use case outside of marketing, you know, dynamic business intelligence. The reason I say dynamic is because most business intelligence traditionally has been: we're going to take a lot of people, we're going to data enter a lot of stuff, we're going to give you a bunch of static charts and graphs.
You know, it's funny because when you think of one of the greatest quotes ever, I happen to love Ted Lasso, and he always says there's a reason the rearview mirror is much smaller than the front view.
Right. And so when you think about our dynamic business intelligence, we're using our data cloud and our inference and contextual-based AI models to help enterprises figure out what do they want to do next? Where do they need to pivot and move inventory regionally? What products should they be manufacturing or not manufacturing? Where should they put their next 500 hotels? These are all business use cases we're doing today.
But when you think about the five-year future, what can the data cloud and the inference and contextual models also help enterprises do better?
Definitely. And the funny thing, and a couple of people online have gotten this, the sort of flywheel of business intelligence back to marketing. So if you have to move skirts from the middle of the country to the northeast because you're a retailer, and we see that that's what women 18 to 35 in the northeast want, once you get all of that inventory in those stores, you need to let people know.
Absolutely. When you open that next 500 hotels, as you open them, you need to let people know they're open. Marketing really is integral to almost every component of how an enterprise runs, as is business intelligence. So we really see it starting a flywheel that'll keep our growth rate really at a very high rate for what we think is many, many years to come. And I'm really glad you brought up the inferencing point because you guys, and you've made it a point to keep your proprietary data within your own stack rather than outsourcing to third party.
It's hard to say in our walls when it's all cloud-based.
I struggle with that too. Like I haven't come up with a good way of putting it.
Yeah. Exactly. Other than we've never exposed our data cloud to any foundational or large language model ever, which includes even our partnership with OpenAI, and it includes our partnership with Palantir.
Can you go into that a little bit and talk about how that benefits obviously Zeta's moat versus maybe a legacy company that can't do that?
Yeah. Well, most of the legacy companies own no data at all, right? But when you look at our data cloud, we believe, and listen, we love OpenAI, you know, we're close to the Anthropic guys, we're close to the Google guys. Palantir is different because they're running a different type of business from what I'm going to talk about. And I think the way Palantir is running their business is the way we're running our business, where you really think about data sovereignty and maintaining both your and your clients' alpha.
Right. So when you think about that, you've got to think about the fact that first and foremost, 51% of the Fortune 100, 24% of the Fortune 500, over 800 global brands are now trusting Zeta with their data. We import their data to a consumer data platform. We merge it with our data, and that's where the models really learn. And that's where sort of I joke the magic happens, right? And that data has zero exhaust. Zero. It's interesting. It doesn't even go back to our data cloud. It's their data plus our data. And that's why the models get so smart. And then as you're creating, maintaining, and monetizing customers, the CRM platform feeds back more information.
Understood. So the longer a client's with us, the smarter the algorithms get, the better the return on investment. If we were to expose our data to a large language model, we'd be giving away our alpha. Now, it's interesting because what a lot of people don't know is that in 2017, we repivoted the whole company. People don't... the famous people know that, right? But what they don't know is the data cloud was fully operational well before the marketing cloud.
So I would say a year, year and a half, because we'd already been a data company. Like so converting from a data lake, antiquated, shouldn't say antiquated, an older version of a data cloud, and a lot of people still use data lakes, so try not to be offensive, but they are antiquated. So I'll say it. The bottom line is converting from our original data lake into a data cloud, we were able to do that in a year, year and a half, whereas the marketing cloud we had to start with an entirely new architecture. So now I'm sitting there, I'm a year and a half in. We've got this unbelievable data cloud, and at that time Oracle Data Cloud was gobbling up everybody. Bought BlueKai, bought DataLogic, bought all these companies, Moat, and they were telling everybody that the way to manage data was to synthesize it into actionable data and then sell it.
Okay. So now I'm sitting there, I'm sucking a little wind here, right? Because I've gone from making a lot of money to not a lot of money because we've got four or 500 full-time engineers working on this marketing cloud, and we've got this data cloud asset that's just sort of sitting there now. We were using it in social, and we were doing certain things with it to target, but there was a massive push on me to start selling the data, like massive, because why don't you load your data into The Trade Desk, or why don't you load your data into third-party DSPs, sell the audience data, make a lot of money. And I remember, you know, Steve Gerber and I, and Steve Vine as well, who's our general counsel, we were sitting in a room. Chris Grinder had not joined us yet. He would have been in the room for this.
But we were sitting in a room, and there was a lot of pressure, and I sort of looked at Gerber and I said, 'Steve, I just have a gut feeling that one day never having exposed our data to anybody at any time is going to be valuable.' And he looked at me. We'd been together at that time, I don't know, 10 years. We're now together 17 years. I don't know how he stands me, quite frankly. But he just looked at me and said, 'Your gut hasn't been wrong yet.'
Well, that's some pretty crazy foresight to have.
Well, a lot of it was luck too, because now you also have to understand, and this is going to sound really crazy, but my stepfather, who was one of my mentors, had been in the direct mail business back when that was a really big thing. And the big differentiator for his direct mail business, which he ended up selling for a bunch of money at the time, and it's not what we'd consider a lot. I mean, it's maybe some people would...
Yeah, it was real money. And he was, I mean, my stepfather was an amazing guy. You know, born the ninth of nine children into abject poverty. He would say, 'Where'd you go to school?' The corner of Clancy Delaney, you know, Clancy Delaney in Essex, which is I guess a corner in Brooklyn. And I remember growing up that his biggest asset was his proprietary database.
Wow. And he used to talk about that like, yeah, yeah, we don't share our data, we don't sell it, you know. And maybe that had some influence. I don't even know. But in that case, we decided to go buy a DSP because I was like, well, if we're not going to load the data into one, let's buy one and let's build a proprietary data set inside of our own DSP so that we could begin to run targeted internet marketing to our data without exposing it to a third party.
Gotcha. And that was the big sort of move there which led us to doing 100% of our own activation.
So there's two really interesting things you said there. First, the return on investment, which we'll get into. But you had the foresight back in 2017 to obviously kind of pivot the company, see where it was going. Now with an ever-changing landscape and where agentic AI, the AI agent economy is going, how do you see that shaping up moving forward? What does the agentic AI economy look like in marketing, and maybe where does Zeta fit into that?
Well, we're going to lead it. I mean, if you think about it, it's not a coincidence that we've started and are now one of the fastest, if not the fastest growing GEO platforms in the world, right? Because as the HTTP layer gets less and less relevant, and let me separate this out because I think you're going to have e-commerce and publishing transactions that are on more commoditized products, okay, that are going to become more agentic based, which like airplane tickets. I'm not sure you're going to have to go to a website to navigate to figure out which flights you want to take. Your agent's going to talk to their agent. It's going to come back, here are the three available flights, and you'll pick one. But an agent's not going to help my wife pick out her next little black dress.
Right. She's going to want to see that. And that's going to happen. But what a lot of people don't understand yet, Mitch, is our GEO platform sees the information in and out from the chats. So we're able to garner large amounts of information as we help enterprises better target and place their brand and their products into search results.
Of large language models, not traditional search engines. So that's actually much richer data than the vast majority of the web page we ingest today. So I think that was a really good decision. I also think that you're going to have two or three big winners in the foundational ecosystem, right? And I think we all know they're going to be the four big guys, right? You're going to have Claude, you're going to have GPT, you're going to have Gemini, you're going to have Grok.
Yeah. I mean, and you know, there is an argument that you could sit around over a bottle of wine or two and really start to think when you get very philosophical. And oddly enough, I get very philosophical often. If they truly accomplish everything they're trying to accomplish, won't I one day be able to go into Claude and say, 'Build me my own foundational platform.'
Correct. Now, they might stop that, but the open-source models won't. Nvidia certainly is making a big bet on open source, and we saw them acquire Hugging Face.
Well, but that's my point. They're the largest open-source platform for foundational models, and you know, their biggest clients are now going to become in some ways competitors. Now, Jensen's very clear. I've actually spoken to him about this personally. They are not going to operate it. They see it like Git with Microsoft, where it's an open platform even though it's owned by another company, and they see that as very open, which is great. And we're obviously partnered with all the large language models, and we use open-source models internally as well.
I believe data is going to be the differentiator because even if you have only four companies that do this and no other platforms ever work, which is highly improbable, but let's go with that for a minute. They can only create so much or ingest so much data. And when you get into synthetic data, you can't build deterministic profiles on individuals with synthetic data. Hey, it's, you know, probably illegal. And don't want to take that chance.
I wouldn't want to take a penalty.
Yeah, exactly.
And if it's not, it will be.
And I think that it's going to be something that's just nearly impossible to do. So, enter, we're going to bifurcate into two different types of enterprises as it relates to software. You're going to have software companies that are sort of workflow management tools, and then you're going to have software companies that create meaningful intelligence for enterprises.
And so we saw obviously with the SaaS apocalypse. I would love to get your opinion on it. Now that we're kind of getting out of it,
My opinion, it's ridiculous.
Okay. I'm glad we're on the same page because it just seemed like every day, you know, whether it was Anthropic or another model would release something, and it was just that sector of the software sector would get destroyed.
I saw a great video this morning on the Wall Street Journal online. Obviously, they can't put a video on the paper just yet. A great video where it was like a SaaS conversation, and the moderator started with a bunch of enterprises and said, 'Okay, okay, 500 people in the room, put your hand up if you vibe coded your CRM platform.'
No hand goes up, right? And it's just, listen, when the internet was created, everybody said it was going to destroy every company that had preceded it. And yes, it created amazing companies like Amazon and Meta and others. But don't forget Apple, Walmart, and JP Morgan Chase, they all predated the creation of the internet. I mean, JP Morgan Chase probably goes back, you know, a hundred plus years at this point. Companies that embraced the internet accelerated and changed the world. Companies that didn't embrace the internet are gone.
And so with that, they not only dominated, but they were probably awarded a premium in the market. And when I look at Zeta in particular, there's some software companies that trade at 60 times sales. There's others at 63. There's others at say around 40 and growing slower. And then I look at Zeta and I'm dumbfounded because it's 3.6 to 4 given on the day. Now you're starting to garner more respect, and I wanted to ask, you know, do you have a thought on maybe why that is? Is it because of the sector, like marketing?
Yeah, I think so. I've given this a lot of thought. I mean, why do we continue to trade at a 70 to 80% multiple discount to the comparables in the intelligent AI infrastructure space? And by the way, just to point out even further, Mitch, if you put all seven of those companies in a row, we would be first, second, or third in every KPI. From net retention rate of 120% last year, we're number one in free cash flow yield to market cap. We're number two in net retention rate. As I just said, we're number three in operating margin. We're number two in the rule of 40, which we were at the rule of I think 67 or 68 last quarter. It's gone back and forth with that.
And listen, I think A, we were very misunderstood as a company. B, I do think there's a negative connotation to the marketing ecosystem. And I'll tell you why I think that is. Up until Zeta, every marketing company that's ever really become a public company has been a point solution.
Okay. So it's either a CDP, ESP, a DSP, all the acronyms. Point solutions ultimately blow up. Why? If we had stayed with just being a DSP, we would have failed. If we had stayed just an ESP, we would have been a mediocre company. If we would have never added connected television, which is now one of our largest and still one of our fastest growing businesses, we would have never grown at the rates we're growing. If I hadn't been sitting in a meeting with somebody two or three years ago who said, 'Yeah, yeah, you got to look at TikTok,' we wouldn't have integrated into TikTok.
The key to the Zeta model is the data plus the AI equals the intent. The activation is totally agnostic, and that's something nobody's really understood. We don't care. Yes, we have slightly higher margin on some things or other things, but at the end of the day, we want to get the highest return on investment for our clients. But back to the ecosystem. So because all of these marketing companies, high-flyers blow up, high-flyers blow up. Then you have guys like Salesforce and Adobe and SAP who've come into the marketing space.
but they've come into it. They do well, and that's, you know, sort of like what we're looking at. Obviously we want to be more Palantir, Snowflake, CrowdStrike, so on and so forth.
But here's the thing. Here's the rub. Many, many, many shareholders, institutional and retail, have bought those point solution stocks and lost money. They've blown up.
So there's a bit of anxiety there.
Correct. So they're always like, 'Oh yeah, yeah, marketing's the first thing killed in a downturn,' or 'marketing's going to be disintermediated,' or 'you're going to do this.' So that sort of negative connotation has hung around our neck. But what none of those other companies ever did, excuse me, is they never adapted. They said, 'No, no, no. We're a DSP. We don't want to own data. We don't want to go out there and create intelligence. We're creating the world's best workflow management tool.'
And now that intelligence is the moat,
but here's the thing. It always was. They just didn't understand it. You know, to me the great companies out there, guys like Palantir, guys like Snowflake, guys like CrowdStrike, Datadog, MongoDB, you could go through these companies. These are great companies. They either create intelligence or help enterprises better create their own intelligence while protecting their alpha. And that is the literal definition of Zeta. But the truth is when you look at our pivot from being a pure marketing platform to an intelligent AI infrastructure company, it's not just a narrative. It's a pivot.
And what we're going to announce at Zelive is going to take that pivot 10 or 20 levels up.
Okay. I'm excited to see that. I just wanted to pivot back into, you know, the return on investment. You stated that it's been 600 to 700%.
Not just I've said it, Forrester has on an independent study certified that we are at the 6 to 700% return on spend.
Which is truly an anomaly nowadays.
And so a scaled customer, if I'm not mistaken, in their first year spends around 700,000.
We don't break out scaled anymore. We just break out super-scaled now.
So super-scaled is over a million.
Okay. Gotcha. And so the reason I referenced that is because it was at scaled was 3.9 million and then 5 million for how, or sorry, 3.9 million.
And I was just going to say, you know, where did you see that limit, I guess, with super-scaled going, and how high could that eventually get to?
I mean, I think, I mean, I remember sitting in a board meeting in 2021 telling my board that one day we would have customers that spend $100 million a year, and they thought I was nuts. Now we have many that spend $100 million a year. Wow. I see no long-term barrier to getting customers to a billion a year in revenue.
And I'm not talking next week, and I'm not even talking in 5 years. But, you know, listen, we're dealing with companies now that are spending, some of them, 10, 12 billion a year in marketing. There's no reason we can't get 10% of that. Now, once again, we're trying to build a $10 billion company. We're probably not going to have a 10% client at that level, but we have vision into it. And I would tell you today, our largest customers are our fastest growing.
Well, that's all you really need to know then.
No, but it's crazy. Like we look at the math, and some of our super, I call them super-duper scaled internally, like because we're still not even 7% of their wallet.
It's wild to me. And you would assume at 600, 700% they would want to pour as much money as they could.
Well, some of it is, you know, these are very large institutions, by the way. Now sometimes I have to deal with the question, well, why aren't you growing faster?
