David Steinberg, CEO of Zeta | Basis Points

Thank you. Very excited to be here. I love your stuff. Watch it all the time. So, this is really cool for me, too.

By the way, it took me about five to seven minutes to walk here. So, this was easy. I was trying to read a contract on the way and kept sort of like running into people. But here we are.

So you're calling me old, admit it. I mean, I thought we were going to wait.

I'll take it.

Well, first of all, like most people, this will sound odd. I don't really feel successful, right? I mean, I don't feel not successful, but I feel like I'm doing my thing every day. And I wake up and I joke to my wife, I turn to my wife and I say, 'Time to make the donuts,' referring to a very old commercial for Dunkin' Donuts. And my wife for some reason still gives me the perfunctory laugh, which is great. But you know, you wake up every day and you do what you have to do. And I would say first, you know, Zeta is my, I think, sixth or seventh company. It depends on how you count it. I've sold four, taken two public. I'm chairman of another. And you know, I've had wins. I've had some not wins. And the funny thing is, like a lot of us, I've learned far more from the ones I didn't win than from the ones I did.

And I always sort of joke, you sort of either win or learn because fail is such a negative connotation to it. I mean, it's certainly not fun, but at the end of the day, I think it's something that I've learned a lot from. And, you know, my first few companies I bootstrapped. The next few companies we raised large amounts of outside capital and it's been sort of an interesting evolution for me to go from entrepreneur to sort of like manager of a bigger company.

And the thing that most entrepreneurs don't get right, so let me be clear: for the first 20 years I was an entrepreneur, started my first company at 21, I was the youngest person in the room. For the last 20 years I have been universally the oldest person in the room that I walk into. When I was a young entrepreneur, I thought I could do anything better than everybody in my company. And I was totally wrong. As an older entrepreneur, more experienced, I actually believe the single most important job I have is to hire, recruit, train, and empower the world's best people. Hire the world's best people. Let them do their jobs. Now, I'm good at my job, which is sort of leading. I focus heavily on strategy, as you would imagine. I focus heavily on customer relationships. I call it relationship management, aka sales. And I really focus on M&A. Steve Gerber, who's my partner, president of the company, he's really the biggest rock star I've ever worked with. And by the way, Steve would have never worked with me 20 years ago because I would have never let him do his job. You know, he's been in company. He was at Digitas, was part of the group who sold that to Publicis, and you know, then took another company, sold it to Veronis Suhler, and is just such a great thing. And my CFO, and then my general counsel, and our CTO, and our chief data officer. It's really the six of us, I think, more of as a partnership than sort of, I mean, obviously we're a corporation. We're public. We have, I don't know how many corporations at this point on how many continents. But the reality is that that's how I think about operating. I could have never done that as a young entrepreneur.

As a young entrepreneur, I had to be the guy. Now I don't care at all. Like it's almost funny, like I find it funny. Like we were just in a meeting and I walked into the meeting and I could tell everybody was like super nervous that I was there and they had to like report stuff. And I asked this one guy this question and he was sort of like, uh, uh, uh. And I was like, dude, there's only one of two answers. There's, you know, the answer or you're going to go find out and you'll tell me later. Both are acceptable. The only thing that's not acceptable is giving me the wrong answer just because you think you have to say something.

The truth of the matter is I can't imagine that this guy would have actually known the TI we were getting from a landlord on the number multiplied by the number of square feet for a new spot we're looking at. But it's just something that my mind naturally understands and goes to as we think about that. I'm still very involved in our, you know, offices and how we design.

Well, and of course, you know, the landlord might say, 'Oh, you're getting a buck 50.' But they're really getting it for a buck 30. So, I always say, 'Great. We'll take it and we'll bid it out.' And you know, sometimes people who are like really like, I shouldn't say professional, but people who have been like in their jobs for a long time, they just don't think like that, right? They're like, 'It's not their money.' I think this is my money. I mean, it's my shareholders' money. Like, we've got to make sure we're doing the best possible job. It's just an example of it's a little weird for me now that people are nervous around me because I just don't think of myself that way. But that's sort of the journey.

Said the famous, they, we call this the infamous board meeting. Yes.

Just throw everything out and start new. So it wasn't even re-engineering. It was starting from scratch.

Yeah. Which is AI.

Yeah. I mean, the decision to throw out our existing platform, which we had spent a fortune to build and buy, to rearchitect an entirely new platform that put this technology that nobody had heard of called artificial intelligence as native to the application layer. We also made our data cloud native to the application layer. It was a bit crazy at the time. And I'd like to say I knew everything was going to happen right the way it was happening. Here's what I did know, Steve. I knew that at the time our systems were not fast enough to make the right decisions for our clients to create the best return on marketing spend possible because we were ingesting even at the time billions of signals. Today it's trillions of signals. And I went down to an engineer and I was like, 'I need to run a report on this thing.' And the report was a complex report. And I went down like two or three hours later. I was like, 'Where's my report?' And the guy is like, 'Oh yeah, I'll have that for you next week.' I was like, 'What do you mean next week? We got to run the campaign. Like it's this has got to...' He's like, 'Yeah, yeah, we can't crunch that type of data.' We were sending it to our office in India and we had our analytics people there and they were building it. And I had been reading and it really started with machine learning, right? And we'd been really focusing. So let's step back again. I don't sleep a lot. I read, I'd say five to seven hours a day. Books, articles, just a lot of white papers, a lot of books. Nowadays I probably waste a little time on Instagram or X, but a lot of ingestion of information. I read five to seven publications cover to cover every morning and sort of all this different stuff. And I had started reading heavily about natural language processing. And I started looking at it and the ability to process billions of data points in seconds was just really exciting to me. And we started looking at the ecosystem and we found this company called BoomTrain which was being run by our now CTO and chief data officer, Chris Mberg and Nish Gore. And we started working with them and they had built one of the most incredible products that nobody would ever buy possible. And they hate when I say that, but the truth is they had built a platform that ingested everything you read and recommended what you should read next. Think of it as like the Netflix recommendation engine for publishers.

Yeah. So I said we should buy this company and we should convert your natural language processing into signal synthification for marketing.

But we needed a deterministic individual and we needed the ability to track them on the internet. So simultaneously we had been introduced through, I don't know if you guys know Fred Wilson who runs Union Square Ventures.

So Fred's a good friend.

I mean, Fred's a legend and by the way, as good a guy as anybody you'll meet.

He called me up. He said, 'You should look at this company in our portfolio called Discuss, DISQUS, which at the time was the world's largest commenting platform. And they were selling publishers this tool to comment, which once again, they didn't really have the money to buy, but they had, I don't know, 200, 300 million people logged in a month. They were tracking because every publisher that used the technology, they had their first-party tracking pixel and their JavaScript on the page. So, I'm triangulating all of this, right? So, now I'm reading articles about machine learning. We're learning about natural language processing through the BoomTrain guys. We're then learning about how the open web works with the Discuss guys and we're trying to get smart on all of those things. And I didn't just go into the board meeting and say, 'We're going to go from making 50 million this year, 60 million next year, and 70 million year after to zero.' I went in and said, I actually didn't know exactly what I was going to do the day I walked in a board meeting. Because the big question was which of the two deals could we afford to do? Could we do BoomTrain and then figure out the data part later? Could we do the data deal and then figure out the NLP, natural language processing part later? And I was simultaneously trying to sell the board that we would buy these assets and in some way go from making all this money to nothing for three years. It actually was longer than that but then we just told them three years, but we didn't know that at the time. I mean, I thought it would be two or three. And by the time I got halfway into the argument with one of our private equity partners, I just looked at him and said, 'You know what? We need to do both of these deals.' And he was even more pissed off, right? You think his head would have exploded, this young guy. And you know, the good news was I controlled the company. So, we were able to get there. It's the first board meeting I have run in over 20 years that I didn't have unanimous vote. And by the way, I have tortured the guy who voted against it since that day.

Yes. You know why? It's important to have board members who disagree with you.

And by the way, he hasn't disagreed with me again, smartly. But the point is that we went into that board meeting not knowing exactly what we were going to do. And then I had to call the BoomTrain guys and I had to call Fred who was the big investor and say I want to do this deal. I'm going to do it in conjunction with this deal. So I need to restructure the offer and we're going to do more earnout, a little less cash because I don't have the cash. I'm going to give you a little more stock, but we're going to put this whole thing together and we're going to build a multi-billion dollar company. And they thought I was nuts, but they didn't have a lot of options. So, we bought BoomTrain, we bought Discuss, we put them together in San Francisco into one office, and it became, you know, game-changing for us. It really changed the game. So, today we are the only marketing cloud that has artificial intelligence and data as foundational. It's literally native to the application layer.

But the problem is they can't become AI native.

Because they have a legacy architecture. So let's pretend this is a marketing cloud for Salesforce, Oracle, Adobe, SAP. They bought ExactTarget, Responsys, Neolane, and they still use the same architecture. So if they want to do an algorithm, they have to step out of the platform through an API to an algorithm to query it. The algorithm then does a data dip into a data repository back to the algorithm to create the intelligence and then it informs the platform what to do. That latency destroys return on investment. Because it's native, I can make a decision in a millisecond and look at a thousand attributes. They can make a decision in seven to 10 seconds looking at five. So when you look at the targeting capabilities for either churn reduction, monetization of a customer or a new customer acquisition, you can make a decision so much faster. It allows to create much better return on investment. By the way, recently Forrester put out a study that for every dollar a client spends through our platform, we return 600 to 700% return on marketing spend. The other massive moat in our business is our data cloud. We have 555 million opted-in individuals who have given us permission to track them and market to them.

No, this is all 100%.

No. They either, they're subscribed to Discuss, they're using our ESP, they're subscribed to one of the billions of newsletters we send out. All of this is, this is misunderstood. I'm glad we're talking about this. 100% of our data cloud is first-party proprietary data. Now we license data to append to it. So like, I'm not a credit, we don't do credit scores, it's Zeta, but we're an agent of the bureau for all three credit reporting agencies because I can build an audience that's not just of the highest level of intent, I can build an audience of people who will absolutely be credit approved for my client's products before we spend a dollar.

So let me take you through the process. It's actually super cool. So, we set up a consumer data platform for our enterprise client. The huge enterprise puts all of their first-party data into that CDP. We then match it with our data. We match in the US at greater than 92%. We have 245 million Americans opted into the data cloud. We then remove the personally identifiable information to protect the consumer. There's another big benefit to us and I'll share that with you in a minute. And we import 5 to 7,000 data elements per person, demographic, psychographic, every credit card transaction they do, location-based data, and tons and tons of web behaviors because we've removed their personally identifiable information, the enterprise never even sees it. But then what happens is our proprietary algorithm inside of the consumer data platform starts saying, 'Okay, their best clients read articles on these three topics for 3.5 minutes. They had a credit card transaction where they complained about renting a car in the six months prior to signing up for their service. They entered these three retail locations prior to making the decision to buy our client's products.' We then do two things that nobody else can do. We then take 100% of our client's data. We layer it against the 245 million Americans and remove their existing customers before we spend any money on marketing. How many times a day do you get an ad for products you already own?

Yeah, that's a waste of money.

We eliminate that waste. Then let's say it's a, we're trying to issue credit card. We say these two million people are actively in market for our client's products, but this 1 million won't be credit approved by their criteria. Don't waste any money on them because all of our data is deterministic. So now let's go back to the Zeta ID number. We go back to the consumer data platform and we tell our client that this 1 million people are not your customers. They're actively in market to buy your product and they will be credit approved with a 90% or greater propensity by your credit score for your products. Would you like to market to them? There's no primer to decode the Zeta ID. You can't take that data anywhere. You have to market through us. And if you terminate us, you lose 100% of the learning in the consumer data platform. Which is why I think last year our net retention rate as a company was 120%. 100% is considered great.

We got several companies with a higher net dollar retention rate.

Yeah. Just like two of them. Palantir, Nvidia, maybe.

But you know, the reality is that that's because our clients love us and they keep spending more money and it's because we deliver them a superior return on investment. I know that's a really long answer to why native AI is a differentiator to the marketing cloud. But let's look at the SaaS apocalypse for one second. Let's just, I would love for somebody who actually works in this...

Runs a software company because I have my own thoughts on this that I've been saying for quite some time. I've been the one that's actually negotiated the contracts on the other side of the table.

So let me start by saying that any software company that is solely focused on workflow management and is not creating intelligence and return on investment for their clients is going to have a massive problem in an AI world and is going to have to heavily pivot their business.

Yes, I think that's going to happen. It's already happened. We're seeing it in the slowing down of numbers shrinking. You know, Salesforce Marketing Cloud shrunk last quarter. You know, TradeDesk is not growing, you know, the way they used to. I mean, these are great companies, but in an AI world, they're challenged, right? When you think about, and by the way, last quarter, our segment grew 8 to 10%, we grew 50%. 30 of that was organic. So by definition, we're taking meaningful market share. We're now, you know, well above a billion in revenue and still growing. It was our fourth consecutive quarter of accelerated growth. That's because of AI native. That's because of controlling our data. But let's go to what is going to be valuable. Proprietary data, intelligence, and return on investment. So let's look at Zeta. Let's go back to the CDP example. The first part of the CDP example is a Fortune 500 company gives me 100% of their first-party data. You think they're going to give 100% of their first-party data to a third-party large language model?

Zero chance. Two, the 550 million opted-in first-party data people in our data cloud, we've never fed them into a large language model and we never will. They're solely used to train our models. So even though we partner with OpenAI, which I assume we'll talk about in a few minutes, they never see our data. Our data is 100% proprietary. So now all of a sudden you've got this massive layer of intelligence where we become the operating system for our clients' marketing ecosystems. This is the one thing nobody understands about our business. The reason we're at 120% net retention rate is because almost nobody fires us. It's because we create this massive layer of intelligence. Now then what do we do? We use that to create a 600% return on marketing spend which by the way with Athena I think will get to 1,000% return on marketing spend. So when people start looking at us and saying, you know, you're going to be disintermediated, I'm like, dude, we grew 30% organically last quarter, 50% total. We are the disintermediator, not the is. And I think that like I'm old, right? So back to...