Zeta Global Holdings Corp. (ZETA) Earnings Call Transcript & Summary

Good afternoon, everyone. Thank you for joining us at the Morgan Stanley TMT Conference. My name is Elizabeth Porter. I'm an analyst on the U.S. software equity research team. And I am very pleased to have with us today Zeta's CEO and CFO, David and Chris. Thank you so much for joining us.

Good to see you. Thank you for having us, Elizabeth.

Welcome. So we are taking audience Q&A at the end. Mics will go around. And before I begin, for important disclosures, please see the Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures. With that, thanks again.

You're welcome. I like the disclaimer.

Of course. I got to throw that in.

So I just want to start it off with a bit of a high-level question because Zeta is in a very unique position between the marketing and advertising ecosystem. So for those that might be less familiar with the story, can you just start off with an overview of the business? And what was the problem you were trying to solve when you founded the company?

Thank you, Elizabeth. So when we founded the company 15 years ago, the real vision was to provide everything a marketer needed in one solution. I have been running a consumer-based distribution platform for wireless phones, and we had 17 different vendors to run our marketing. So the vision was, how do we do everything in one place? So today, we have a data cloud, we have a marketing cloud. Our platform ingests the vast majority of the Internet every moment of every day. We have 240 million opted-in Americans in our data cloud that we're seeing across 5.2 million publisher platforms on a first-party basis. We ingest all of that information, and we use artificial intelligence. Yes, it took me 4 minutes to get to it. We use artificial intelligence to synthesize that into intent-based scores. What does a consumer intend to buy next? Do they intend on churning off a platform? Do they intend on buying an additional product that our clients sell? And what the platform does is it builds marketing to target those individuals to help very large enterprises and now agency holdcos, which is a new addition to our strategy, create, maintain and monetize customers at a substantially lower cost than they could without our data or our software.

Great. And so you mentioned kind of really focusing on some of the larger enterprises. You have a long tail of customers, but really it's the super-scaled customers that account for over 95% of revenue...

Scaled customers.

Scaled customers.

Yes, yes, yes. Superscale is the next one....

Oh, superscale. Yes, sorry. Scaled customer.

That's okay. Chris would chastise me if I didn't say it now.

And so when we're looking at just the scaled customer cohort, there's about 450 customers in there, which is a little bit unique for a software company. So just give us a little bit more of what does that target customer really look like? And what is their spend with you guys?

Well, I'll let Chris get to the spend. But the truth of the matter is the 452, which is very close to 450 approximate, at the end of last quarter, is actually not the full scope of our customer base. So it's interesting because we work with some very large enterprises, and we work with a number of the very large agency hold corporations. In some of those cases, 1 agency might be 15 or 20 enterprise clients that would also individually equal scaled clients, but our auditors have asked us to report that as one superscale client. So we have almost 700 customers if you look at the sort of full scope of customers that we interact with. As it relates to sort of marketing, we look at it as how do we help these very large enterprises solve problems. And one of the biggest problems they have is customer acquisition and customer retention. Our software, which plugs into our data, utilizes the data to figure out who are the individuals who might not just be in market for their product actively, but because all of our data is deterministic, meaning we know who the individual is, it's not probabilistic, we're not guessing, we can layer stuff like a credit score on top of that data layer. We can layer on top of it a propensity to that particular enterprise's brand. And one of the ways we're able to lower a marketer's cost to create, maintain customers by an average of 50% is by taking out of the marketing funnel people who would never be approved or people who have no propensity to that product. As it relates to revenue per user, you're better at the ARPU numbers than I am.

It's interesting, we have a very -- David talks about the value we provide to our customers. It creates a very sticky customer such that the longer they're with us, the bigger they actually become. And there's a very interesting cohorting analysis where we look at how many have been with us less than a year, 1 to 3 years and then 3 or more years. And what we find is most of the time we start off with a pilot or proof of concept that might be around $100,000 on average, it takes us less than 12 months to get them to around $600,000 in spend. But that second cohort of 1- to 3-year customers grows from $600,000 to $1.3 million, and then the 3-year-plus customer goes from $1.3 million to $2.1 million. But the most important cohort is the 1- to 3-year one. They were the very first set of scaled customers, very large enterprises that got on to the new data marketing platform in 2021, right when all this capability was released; the same set of customers that went through a hunter/farmer motion, so they were closed through a pilot and grown by adding channels and use cases and the same customers that didn't reduce spend or hold spend, but grew their spend with data through a pretty choppy marketing backdrop.

Great. And kind of going to a little bit about a choppy marketing backdrop and a lot of noise that we're seeing in the industry right now, there's just a lot of change going on, whether it's cookies, IDFA, Google, Yahoo! e-mail changes. We can keep up with a pretty long list. So I want to just kind of level set the conversation, is data technology impacted by any of these changes?

Well, it is impacted, but for the positive. So to put it in perspective, we do not use a third-party cookie to identify individuals or track. We use the Zeta ID number, which we're able to identify across 5.2 million publisher platforms, which is around 1 trillion pages of content on any given day. So as the third-party cookie goes from hypothetically disappearing to actually disappearing, the efficacy of our platform stays consistent. So what's happening is you're seeing other ships drop while our ship continues to sail in the same direction. The Yahoo! and Google e-mail changes went into full effect in February. We saw an increase in open rate and an increase in click-through rate. I think a lot of that is that we've been partners with those guys for many, many years. And we've been adhering to the rules that they put into a place for everybody else already for like 2 or 3 years. People ask me, how many changes we had to make to effectuate not being affected? And the answer was none. I mean we really had already done it. Now we had a war room set up. We had hundreds of our employees up 24 hours a day between San Francisco, New York, Paris, Czech Republic and Hyderabad, India for our network operating center during that period. We experienced effectively no disruption. We had 1 client of 1,000 that had a disruption for about 35 minutes that we were able to right by getting on the phone with the male manager of one of the ISPs. But that was a net positive for us because the less bad messages that get into the inbox, the more attention, good messages get and just sort of works that way. And then as it relates to the elimination of the IDFA, for those who don't know, that was Apple's tracking mechanism inside of the iOS environment. And it hit a number of the mobile-focused platforms very, very hard. We've never used the IDFA. It was not something we had used. We really see mobile, which is, call it, 2% of our revenue today, as one of the biggest revenue drivers going into 2025 and beyond. And I think that's because we have now figured out how to track and identify an individual using the Zeta ID inside of the iOS and Safari ecosystems. So that's now created, we think, a very unique selling opportunity. Most of the mobile platforms that have come back are doing it on the 8% of people who are opting it. So they're taking 8%, they're using it as a sort of cohort. They're building statistical samples, and they're saying the deviance is low enough as they extrapolate out to the rest. We prefer 100% attribution and we tend to do that. So the Zeta ID is now uniquely able to do that. We believe that our new mobile platform will be widely available and in production by the end of the second quarter. We think it will be heavily utilized going into next year. I want to point out that the changes to the cookie and changes to e-mail and the changes to our mobile platform, none of which are baked into our guidance for this year. Those would be upside to guidance if they happen sooner, and we expect most of them begin to really impact the business in the years to come.

I do want to come back to mobile in a minute. But before I do, just given -- it sounds like Zeta is pretty well insulated, not impacted by a lot of the changes that we're seeing in the industry. And while the industry is going through broader disruption, how is that changing your conversations with customers more directly and their willingness to spend? You've talked about some major RFPs hitting records, are those things related at all?

Yes, 100%, and very intuitive of you, Elizabeth. So yes, we announced that we were at a record pipeline, record RFPs. And miraculously, we had record revenue in the fourth quarter and last year. And it's -- they're directly correlated. What we're seeing in particular, in the agency holdcos, they are getting very, very nervous about their existing tool set around programmatic, around CTV, around other components of the marketing ecosystem with the impending coming of the elimination of the third-party tracking pixel or cookie. And we're seeing those cohorts of our business growing at an accelerated pace, as I'm sure you know better than anybody, Elizabeth. Last year, we entered 2023 and we gave guidance to 17% growth for the year, we finished the year at 23% growth. We always expect to beat and raise. The fourth quarter was our 10th quarter in a row, which is the number of quarters miraculously we've been a public company, we beat our guidance and raised guidance. We expect next February or our goal is next February to be able to announce our 14th straight quarter of beat and raise. So we always come in with a conservative outlook and a conservative guidance, but I don't think even we expect it to go from 17% to 23%. And I think a lot of that was what's happening in the industry is fueling accelerated growth for Zeta as a company.

And I want to move over to product and competition a little bit more. When we're thinking about the Zeta Marketing Platform, kind of your ZMP, who else is out there when you're talking about competitors? Who are you seeing in these [ bake-offs ]? And really, why does the ZMP win?

Well, it wins because it's a superior product, to be frank, but any CEO is going to say that, right? I'm not going to say we suck. I think the statistics do speak for themselves. Last year, we won over 50% of the RFPs and engagements we were invited to participate in with an average of 8 to 12 vendors showing up. In a 100% of the cases, to my knowledge, I have to legally say, we beat or displaced Salesforce, Oracle, Adobe, The Trade Desk or Epsilon or SAP in 100% of those RFPs. So the platform is working, and we are beating companies that there -- we're in the middle of what we see is the largest replacement cycle of marketing clouds to date. Most of the main marketing clouds came online when large tech holding companies started buying e-mail assets, I'm looking at ExactTarget responses, Neolane and others about 10 years ago. And most of these very large tech companies have not invested heavily in those assets, right? I sort of joke, they started as their side hustle. Now in most cases, they're the side hustle to the side hustle for those large companies. And these are great companies, and I sort of joke about it, but they are exceptional at their core businesses. But when you look at their marketing clouds, our marketing cloud is purpose-built. And according to Forrester, which named us the #1 marketing automation company in the world 2 years in a row, interestingly enough, in the same year, they moved us to #1. They removed Salesforce, Oracle and Adobe from the leader category. In that same year, I like to point out, we didn't even spend $200,000 on services with Forrester. So we're not one of those companies that miraculously spends $10 million with certain rating services and end up at #1, which is not to say they do that. I'm simply saying we didn't spend that much money with them. And if you look at the results, they said 2 things that were like headlines that even my mother could understand, and that's saying something. Zeta makes the most complex marketing problem simple. And that's our biggest goal, right? Now a little more complex is they also said that Zeta is the only marketing platform -- and as Chris pointed out, it's because we launched it in 2021, we didn't launch it 10 years ago. We're the only marketing platform to put artificial intelligence and data as native to the application layer. Why is that important? Because we can make a decision in a millisecond. Our competitors have to step out of their platform to an algorithm, which then has to do a data dip into a third-party data source. It just takes more time and it's too slow. So those are the reasons, I think, we are able to lower our enterprise clients' costs to create, maintain and monetize customers fairly dramatically. That comes from the technology's capability and the data's capability.

And I'd love to just dive a little bit deeper into that data layer that's embedded kind of natively into the app. And clearly, the data is really a key differentiator for you guys. So when we think about kind of what's in there, can you kind of walk us through what are some of the main identifiers? How have you collected these data sources? And really, how did you combine such a big data asset?

Well, it took us 15 years. I mean this was not something that could be put together overnight. We bought a series of tools that support publishers on a first-party technology basis, a series of networks. I mean we own some of the largest job boards in the United States. We own Disqus, which is the world's largest commenting platform, which is embedded into a million of the platforms. We own ArcaMax, which is one of the largest publishers of opted-in newsletters. All of these businesses as stand-alone were small and not that interesting. And quite frankly, we didn't buy them for their core business strategies. We bought them because we like the data generation, and we were able to combine them. You put that together with one of the largest DSPs where we're literally ingesting every ad impression bid on everywhere, every moment of every day, and we're able to synthesize all of this down to a deterministic individual. Now to protect that individual, we layer a Zeta ID number on top of their personally identifiable information, and we remove that PII. So they're protected. Nobody knows who they are. They just know what Zeta ID number they are. That has a multifold purpose. One, always be on the right side of regulation. The 240 million Americans in our data cloud have opted in to be there. Two, we never sell our data to anybody at any time at any price. Three, we never expose the personally identifiable information to any enterprise client at any time. And four, there is no primer to decode who that individual is without the Zeta Marketing Platform. So when you put our CDP in place, we're importing between 2,000 and 4,000 data elements, some are data in motion. What did you read yesterday? How long did you take to read it? What did you search yesterday? What were your credit card transactions? Where did you go? Some of it is data at rest, your demographic information. That doesn't change often. All of that sits in there, and the algorithm gets smarter inside of the CDP. They can see that these 2 million people are not only actively in market to buy a wireless phone from them, they have a high level of propensity to our clients' brand and they will meet our clients' credit approval rates. Would you like to market to them? If you want to market to that cohort, there's no way to decode the primer. There's no primer. You have to go to the Zeta Marketing Platform, which is one of the reasons I think our net retention rate has been between 110 and 115 over the last number of years, and Chris is guiding it to the higher end of our range this year.

Great. And more on the portfolio side. You alluded to this earlier, but you have a couple of really interesting opportunities: one, kind of being the mobile; and two, with AI Intelligent Agent Composer with a new product of your...

Say that 3 times fast...

Seriously, it's a tongue twister. So on mobile, kind of we first look into that one, kind of what are you guys really trying to build here? And why was this channel the right one for you guys to go after now?

Well, I mean, quite frankly, we are -- we tend to be very opportunistic as a management team, go where the money is, right? And the elimination of the IDFA, which iOS is the dominant platform in the United States of America, which is the dominant marketing country in the planet, just created a Wild West to marketing. And once we were able to crack the code on tracking the Zeta ID through the mobile ecosystem, it became a no-brainer. I wanted to make sure we could do that before we slotted the investment dollars to build that out. And so that made a tremendous amount of sense to us. As it relates to the agent builder, one of the things we see for artificial intelligence -- and by the way, I'd like to say we've been talking about AI for 7 years, not 7 months. We have a meaningful patent portfolio around artificial intelligence and machine learning. And it's been something that we've been heavily invested in for many years now. We see it as a revenue opportunity, not just as an efficiency opportunity, although we're certainly using it for both. When you think about the agent builder, which is what we love and we call it because it's easier than the name that somebody made up for it, when we think about it, to me, it becomes really a virtual data scientist. So if you're paying a data scientist $250,000 a year, to answer questions, it might take them hours to do. Why wouldn't you convert that into a $50,000 a year subscription service with Zeta to become a virtual digital -- I'm sorry, data scientist who can give you answers in milliseconds. And that's where I see the real opportunity. A very intuitive question today is, isn't that a different sales motion when you start to sell that? The answer is no, when you look at our existing client, right? So if you've got 452 scaled clients, which represent 700 enterprise clients of scale, turning that on becomes very, very meaningful. And we'll probably give it away free for some period of months to get people sort of addicted to it and then start charging. But we're very, very excited, like many companies are, at the prospect of utilizing artificial intelligence as a product versus just building intent and efficiency.

And so I have to ask on monetization since you left the door open a little bit. So it sounds like the strategy is more driving adoption before trying to pull the levers on price. But any way to kind of frame for investors across agent and also mobile, how we should think about that impacting the financial model?

Well, right now, we don't have it in the model. So it's upside to the model. I do think both those products are extremely high gross margin. And we like high gross margin, much to everybody's disbelief. But in all seriousness, our goal is to continue to get stickier with our clients. Our goal is to continue to drive 20-plus percent revenue. Put in perspective, this was our fourth year in a row of 20-plus percent growth. We've guided to 20%-plus growth this year. Last year, we started at 17%, ended up at 23%. Our goal is to continue to beat and raise. And part of that is driving new products to existing customers that are very high margin and very sticky. Both of these are things that our clients are already paying for. Both of these are cost savings to what our clients are currently spending with competitors. I like that from a sales motion perspective. It gives you a unique opportunity to get large adoption rates quickly. In many ways, it reminds me of our CTV business, which we launched a few years ago and said we would scale to a $100 million business, which we're doing; and high gross margin, on platform, move that money from linear to our CTV business, and it's been a big windfall for us.

Great. So we've heard a lot about the product. So I do want to dig into some of the go-to-market motion because you guys have also been making changes there. First, on whether it's technology partners and then also the agency partners. So first, just to focus on kind of the go-to-market partnerships with Snowflake, AWS, kind of systems integrators, what do these partnerships really allow you to do now that you couldn't have done before? And what's the opportunities to accelerate kind of the customer add growth from you guys?

I feel like I'm answering all the questions, so sorry, Chris. Once we get to the numbers, I'll let Chris answer everything. Just...

Don't worry, I have some numbers questions...

Because I do have -- I circled those at the bottom and said, Elizabeth, jump down there as early as you can. So let's put it in perspective, right? Anytime you've got a channel partner, no matter who it is, you're able to do more with less. It's just that simple, right? So it started with the agency holdcos, which have gone from 0 to 100 miles an hour pretty quickly. And we sort of talk about the fact we've gone from 0, 4 years ago to -- we just executed our fourth of the big 6. So we're getting pretty meaningful penetration there. To put it in perspective, the 4 agency holdcos we work with today manage over $100 billion a year in marketing and advertising spend. So there's a meaningful opportunity for growth there, one of which I call mature, although it's still -- the mature one is still growing faster than our core company rate. So that's good. One is on board and scaling. The third is coming on board and what will probably be the most meaningful onboarding we've ever done. And the fourth, we've just executed. So we're working on that. So anytime you can sign a deal with one enterprise and pick up, I don't know, 15, 20, 30 of the world's largest clients, we like that. It's funny because people I'm not sure really understand that even though a lot of those guys start on our integrated platform versus our direct platform and if you think of the mature one, when it started a few years ago, it started at 93% on the integrated platform and 7% direct. Today, it's 70% on our platform and 30% on the integrated platform. So they start off and then they migrate on for a whole host of reasons that we expect all of the existing ones to do. But the other really interesting thing about it is they're scaling faster. We're seeing those clients start more quickly. Yes, the gross margin, if you think of our business, the on-platform business, which is 75 -- 70% to 75% of our revenue, call that this year a $600 million business with mid-70s gross margin growing above our core rate is doing very, very well. What people are not realizing as they look at the integrated platform is even though that's a 30s -- mid-30s gross margin, the incremental expense to operate it is the commission we pay the salesperson, which is about 5%. So the incremental margin on that business is, call it, 30 versus our operating margin guide this year to 19. So even though you're seeing gross margins sort of flat; operating margins, we added 1,000 basis points to operating margins over the last few years. We haven't had a year yet where we haven't added at least, what, 200 basis points to gross margin. And we've certainly not guided to that this year, but certainly something we've done in the last few years in a row.

And kind of with those agencies coming on, so it moves it a little bit more towards the integrated platform. But what have you seen customers at scale, kind of what's the opportunity to shift that back to direct? And then for any investors that are new here, like why is the direct platform kind of the preferred channel?

Chris?

I think you have to start with where does the customer get the highest impact, right? So they choose Zeta, the agency does, to start with integrated because we have the ability through our Zeta ID to be able to point back and match to a meta ID, which is unique to the industry. So they can use Zeta's data, use Zeta's measurement and attribution to know exactly who is in market at that moment in time, but inside the walled garden, we'll go purchase that inventory for them. But what's really difficult about marketing is being able to keep the fidelity of the audience to create no matter the channel you choose because until you were working with Zeta, you'd have to work with a different probabilistic audience if you're going after display video or e-mail or connected TV. By starting on the social with Zeta, you can keep the fidelity of that exact same audience, and now all these brands are running omnichannel marketing strategies anyway. So now I can keep that same audience to go after CTV. I know the marketing dollars I've deployed there. I know who I'm trying to engage in through Zeta's data, you can see whether if an action was taken. You can follow through on display video, that exact same audience and then send an e-mail. So what's unique about Zeta's attribution is the marketer, real time on our platform, can see dollars deployed, decisions made or not. So you have the ability now as a marketer to go to the bad guy CFO and say, here is the actual hard ROI I'm getting on my investment.

And we're also the only platform I know of that has automated the onboarding of social media marketing. It's a very cumbersome platform and it's a very cumbersome thing to do, which is it's a big pain point to the holdcos, which is why we started by solving that first. And then when they start, to Chris' point, using the direct platform, not only are they seeing the fidelity and the efficacy, the return on investment goes up exponentially because it's all connected. So I opened a message this morning. I didn't click on it. We showed them a CTV ad that evening, and then they see a video ad online the next day and they purchase. Our platform is the only one I know of in the world that can track every one of those into one attribution model.

Great. I'm going to have a couple -- 1 or 2 more questions and then open it up to audience Q&A. So first, Chris, I want to get you involved kind of a little bit more here.

Yes, I threw it to him last...

You did, you did, you did. I wanted to touch on kind of the guidance that was laid out for the 2025 target, and you did that back in 2022. It's a plan that you have been executing very well against. So can you just walk through the levers of the 2025 plan? Are there any areas where you think you need to be investing incrementally behind? And where do you feel the most confident in achieving that plan?

I think we're confident across all 3. What Zeta did that -- well, I guess, not so unique is having a long-term model. What becomes a little bit more difficult is when you put a win on it, my favorite question is when is when. But I think what is unique about our model is when you put the set of KPIs, the same ones that I use as a CFO, that my planning team uses to forecast the business with our President and COO, in the hands of all your shareholders and your analysts so that you can have an objective conversation whether you're tracking to them or not. For our business, we've got 5 levers. How fast we should be adding scaled customers. Our model is to grow year-over-year between 8% and 12%. I can't think of a quarter where it's been less than 12% growth in recent history. The second is ARPU, how big should we be making them. Our model is between 8% and 12%. I think we've had 3-plus years of double-digit ARPU growth. Net revenue retention, how sticky they should be, the model is between 110 and 115. This year, we said we'd be at the high end of that range. Direct mix because we want to make sure that they're profitable over time, the model is 75%. And then quota carriers, right, how quickly and efficiently we should be adding quota carriers to the model, and the model there is around mid-teens growth. So we've got a number of levers. And as you said, we're fortunate to earn in each one of those dimensions, we're executing to them, and we feel very confident. In fact, we said at our Investor Day in September, all of those are net leased and we think we'll get there early across the board.

Yes. People keep looking at us and saying, is that what you're going to do? And I keep saying, we say a minimum of, which is the case. We would have to slow our growth rate down dramatically to only hit our current 2025 plan.

Got it. And on the expense side, your guidance for fiscal '24 is about 19% margins, implying about 120 bps of expansion, that's after you did around 300 over the last year. So still driving that impressive expansion, and you mentioned you're already tracking well ahead of those 2025 targets. So when we think about the incremental leverage in the model, where is it most coming from and what opportunities do you have to push?

Primarily from G&A. So we've made really good inroads on G&A over time. Although the beauty of what the company has set up now for a decade plus is being so globally distributed. So we have the ability to invest in very, very high capable, expensive resources in certain places in the globe; whereas in other places of the globe, not lose efficacy, but get it at a much better price point. So we can do that in R&D. We can do that in G&A. And then on the sales side, we've really evolved having to go from hiring many people to the right people. We have a hiring strategy around tenure. The average tenure of our sales reps, over 10 years, so very experienced. They have verticalized expertise. It's very difficult if you're a seller to walk into JPMorgan or Morgan Stanley 1 day and then Frontier Airlines on Thursday. You have to have the ability to go deep on their business model to be relevant with your product. So we have a hunter/farmer designation. We really test each of those, and we're hiring the right people. In fact, we have a new staff that we shared at our earnings, where if you just look at this year's class of sellers against the prior year's class of sellers, this year's class is closing their first deal within 5 to 6 months. And as you know, Elizabeth, in software, that's pretty fast. But they closed 4x as much TCV in last years than the class prior, and they're winning at a higher rate. So we're getting better and better at training them and wrapping a really good certification and L&D around them.

One of the great statistics I like is in the year before we went public, which was just 3 years ago, we did $368 million in revenue, and we had approximately 1,500 employees. Last year, we did $729 million in revenue. That was almost all organic growth, a couple of small tuck-in acquisitions, and we had approximately 1,600 global employees. So we're very good on automation. We're very good on managing. And as Chris said, we're very good on building a global sales force. We architect our software in San Francisco, Silicon Valley and Boston. We engineer it in the Czech Republic and in Bangalore, and we program at our campus in Hyderabad and in Manila, which allows us to have a 24-hour day, and it allows us to hire staff engineers in San Francisco, while you're using regular engineers in Prague, while you're using programmers in Hyderabad.

You guys are staying busy.

Yes. Well, we don't sit around long.

Do we have any questions from the audience? So I wanted to ask on the scaled ARPU. That's grown really nicely from around $880,000 to $1.6 million in 2023, and you expect that to grow to north of $2 million in your 2025 target. How -- Chris, you mentioned this a little bit that you do the really interesting cohort analysis, but how much visibility do you have on these metrics? And customers are clearly spending more with you the longer they stay, but where are these budgets coming from?

I think -- so David talked earlier about this replacement cycle, right? So if you think about the Chief Marketing Officers and the CIOs of their partners' budget, it comes in 3 different categories: they manage data, and that's expensive, so like CDPs; they are looking to automate their marketing, their marketing clouds; and then their activation strategy. So we're able to cut across each one of those areas of technology spend and offer a product that then allows them to consolidate multiple different point solutions. Think about our ability to grow ARPU in 3 different ways. We serve 3 different use cases. We help brands acquire new customers, keep their existing customers and grow their existing customers. If we're working with a very large U.S. auto insurer, for example, we might be helping them today acquire new policy owners through e-mail as the primary channel. And we can make that bigger and bigger and bigger, just staying within a single use case, single channel. Second most likely way we'll expand ARPU is then having them add channels. So staying within that acquired use case and just sticking with the same U.S. auto insurer going from e-mail to connected TV to display video. The third would be then getting into helping them retain their existing customers and then cross-sell into them. So lots of different ways to grow that ARPU, which, as you pointed out, it's gone from $80,000 to now $1.6 million, on its way to over $2 million.

And one of the things I look at is our existing customers today, Elizabeth, today spend over $100 billion a year on marketing. So our 2025 plan has us getting to 1% wallet share. We have a lot of upside from here. So when you talk about where is it coming from, it's coming from other partners that they are currently working with. We're taking market share, right? We're growing our business, I think, at almost 100% greater than the current marketing ecosystem is growing. Is that about right?

Yes, we're about twice as fast, yes, exactly. Yes.

So when you're growing almost twice as fast as your market, you are taking market share, and that's what we're seeing. Although I will say, coming out of the Consumer Electronics Show this year is probably the most positive I have seen anybody in marketing in 3 years. I met with all the exec -- senior execs from the holdcos, a number of enterprise CMOs, everybody was talking about growing nicely this year. So we think the macro environment is also good. It's also a political year for us, and we have a political component to our business that tends to be very successful through those cycles. So we're feeling very, very good about not just taking share from others, but from our current growth rates.

Great. So with that, we're almost out of time. I think it's a great place to leave it. There's a lot to look forward to, macro starting to get a little better. You guys have AI, agents, mobile, working more with agencies, which would be a new lever. So I'm looking forward to everything that comes in 2024.

Well, especially your new baby.

Thank you.

Thank you. Thank you very much.

Thanks, everybody.