00:01:17 · David Steinberg
[00:01:17] S0: That's a very polite way of calling me old. So I appreciate that. You know, listen, I I started really my career in the wireless space and then moved into the internet space. Now of course we're focused on AI and cloud. What I would tell you is in every case, everybody believes that this technology is going to be completely different than every technology that's ever followed. Quite frankly, in every case, everybody has felt that every company that preceded these new technologies is going to go away. I remember sort of the beginning of the internet. Yes, I'm old. My kids laughed. They're like, Dad, you were alive before the internet? I'm like, Oh yeah. Yes, I was. I mean, not technically ARPANET, which started in 1959, but I wasn't born until ten, eleven years later. But the point of the story is everybody said that the internet was going to put everybody out of business. JP Morgan Chase, Walmart, FedEx were all going away. In all cases, companies that had deep and meaningful relationships with their customers, controlled proprietary data and or distribution that adopted the internet became juggernauts. Companies that did not have those characteristics or did not adopt the internet went the way of the dodo bird, AKA GOP, right? And I think in the three revolutions that I've lived through, and listen, there's a cycle every eighteen to twenty four months, but the revolutions that I've lived through, I would say the biggest from my vantage point of my lifetime were first, the internet. Second, mobile. And third, AI. And cloud was important. You could argue cloud doesn't exist without the internet. Then outsourcing was important, but that doesn't exist without real time communications. And you could go through all the derivatives, right? I mean, you go back to the Gutenberg printing press if you want. I was not around for that.
00:03:24 · Guillaume Moubeche, David Steinberg
[00:03:24] S1: And talking about proprietary data, because I think it's extremely important. It's also a hot topic. I know you have worked a lot on it. We're seeing, you know, like the OpenAI and Anthropic with their closed models, a lot of people, you know, like start realizing that their data is also like going away when using like this kind of providers. [00:03:47] S0: As my friend Alex Karp likes to say, Data sovereignty is the key to maintaining your company's alpha. And we had Zeta Live the other day, and Alex was sort of my surprise keynote. He came out, he was unbelievable, as you would expect. He doesn't do a lot of those things. So it was very, very cool for us to have it. Of course, Palantir is one of our biggest partners globally now. And we quite frankly are a very important partner to Palantir, as he said on stage. But companies that are feeding all of their proprietary data into frontier and large language models are going to ultimately be giving away what's their differentiator in business, because ultimately that data feeds models that learn that can feed your competitors. Right? And that's one of the reasons we believe that no data exhaust and maintaining our client and our proprietary data outside of the frontier models is something that's been mission critical to us as a company. And I would tell you for our 800 plus global branded clients, very, very important to them too.
00:03:47 · David Steinberg, Guillaume Moubeche
[00:03:47] S0: As my friend Alex Karp likes to say, Data sovereignty is the key to maintaining your company's alpha. And we had Zeta Live the other day, and Alex was sort of my surprise keynote. He came out, he was unbelievable, as you would expect. He doesn't do a lot of those things. So it was very, very cool for us to have it. Of course, Palantir is one of our biggest partners globally now. And we quite frankly are a very important partner to Palantir, as he said on stage. But companies that are feeding all of their proprietary data into frontier and large language models are going to ultimately be giving away what's their differentiator in business, because ultimately that data feeds models that learn that can feed your competitors. Right? And that's one of the reasons we believe that no data exhaust and maintaining our client and our proprietary data outside of the frontier models is something that's been mission critical to us as a company. And I would tell you for our 800 plus global branded clients, very, very important to them too. [00:04:54] S1: I want to go back a little bit to 2017 because obviously Zeta has like a long journey. And at some points, your company was highly profitable and you went to your boards and said, okay, like, I think this is the time we need to totally change where we heading. Can you maybe walk us through the story and how you convinced the board to let you do what you wanted to do? Yeah. I mean, to
00:05:20 · David Steinberg
[00:05:20] S0: set the stage, we were a good, not great customer life cycle marketing platform. We sort of had a CRM engine. We had a customer acquisition engine. We had a customer monetization engine, and we operated in silos. And one day I had to run a report. We had built this huge data repository. Back then you called it a data lake, where we had millions and millions of opted in consumers. Today Zeta has five fifty million globally opted in individuals in our data cloud. But back then it was sort of smaller, but it was a value to us and it helped us with the marketing. We were going to make in 2017, call it 50,000,000 in EBITDA, which was a lot of money back then. It's still a lot of Yes, it is. Yeah, know. I know. I was going to say that so I don't sound like too much of a dick. But now it's a month for us in some months for profit. But the reality is back then, it was all the money in the world. And I had to go into a private equity backed board and explain why we were going to make no profit for the next two or three years, because I was going to take 400 engineers and put them full time on building an entirely new platform that would put this new technology they thought was science fiction called artificial intelligence and data as foundational to the application layer. Now, the catalyst was I had to run a report for a very large client. And back then, I'd need a very typical startup. We're in this sort of building on the corner of 34th And Madison Avenue in New York. It was a really cool, but smaller building. And we were on five floors. And I don't mean we had five floors. I meant we were on five floors, right? Because we kept running out of space and then we'd have to get more space in the building. And then this guy guy was leaving on that floor, we'd take that space. So I had to go down to the floor where the data scientists were and say, I need this report. And they were like, Great. So I go back up to my floor, I'm doing what I'm doing. Hour later, I'm like, Where's the report? I go downstairs, like, Guys, where's the report? They're like, Oh, don't worry. We've got it. We sent the data over to our India campus. To this day, our biggest employee group is in Hyderabad, India, and they're going to work on it. They're going to have the report back for you next week. Now, I needed that report in minutes.
00:07:55 · David Steinberg
But what was happening was data was growing so exponentially that no reporting infrastructure of its time could keep up. Now at the same time, Steve Gerber, my partner, our president, came to me with an idea around this company called BoomTrain, which was sort of like the hot AI technology company of its day. Was too early, but an unbelievable team. Chris Memberg and Nizh Gore, who were two of the three co founders, were running the business. To this day, Chris is our Chief Technology Officer and Head of Product, and Nizh is our Global Chief Data Officer. But at the time, we were just introduced to them and they were using natural language processing to process very large amounts of web data very, very quickly. I ended up meeting Chris. We went for a two hour walk. I like to do that, especially if it's a beautiful Parisian day, beautiful New York day, beautiful Miami day. I'm like, Let's just go walk. And I got back to the office and I said, Jesus, we need to buy this company. And Steve and I set the stage for that. Then independently, I was introduced through my friend Fred Wilson, who runs Union Square Ventures to discuss DISQUS, which was at the time... I mean, now it's the world's largest commenting platform. But at the time, was commenting platform that was embedded into millions of publishers from a first party technology perspective. And the data that it was extracting on a deterministic basis was unbelievable. They weren't even monetizing it. Their whole strategy was selling the technology to publishers, sort of like on a SaaS basis. For better or worse, publishers didn't have a lot of money back then to do it, less now. So we sort of had this conundrum. We wanted to buy BoomTrain. We wanted to buy Discuss. We didn't have the money to do both. At the same time, we knew that the company, to become what we wanted to become, required a complete rearchitecture of our platforms that we had invested hundreds of millions to date to build in 2017. So I go into this board meeting and I'm trying to decide, do we want to buy BoomTrain? Do we want to buy Discuss? Do we want to start with the AI layer and then figure the data out later? Do we want to start with the data layer and figure out the AI later?
00:10:24 · David Steinberg
And in this board meeting, I mean, of the most contentious of my life, let me be honest, probably second most contentious board meeting I've ever run, but the most contentious board meeting by far that we'd ever had at Zeta. And at that point, we were already eight, nine years old, and by far the most contentious even today at Zeta. There was one at a prior company that was pretty ugly, we'll go with this one. And this young guy who was an analyst, I'm sorry, he became a partner in the fund, but he was on it. He just was adamantly opposed to these deals. He pissed me off so much during the course of the board meeting that, Gee, I just said, You know what? We're buying both. And this guy's head exploded. It was just great. Listen, for better or worse, I controlled the company. I still vote greater than 50% of the shares today. My family, through everything, we're still by far the largest shareholders in the company today. And back then I was the controlling shareholder. Still, as I said, vote the controlling shares. I just said, Well, we're going to do this. And to be honest, everybody sort of lined up except for them. That group was not for it. Fortunately, the young person... I shouldn't say young person, in his mid to late 20s at the time, is no longer involved with us. He's off doing something else as he should be. The gentleman who was the partner on the deal is actually still with us and says to this day, He was wrong. This was the best decision ever. Was funny, G, a few months ago, a Harvard case study actually published on Zeta written by Columbia with Harvard on that board meeting, which for an entrepreneur, right? I didn't go to Harvard. I didn't... So having that happen was pretty cool. But that was the moment that launched us into who Zeta is today.