Speaker A
This is very exciting. So, I'm asking you more directly. Is this something that you're interested in pursuing? Am I thinking along the right lines?
Zeta Global CTO Christian Monberg discusses AI, identity, data privacy, and the future of advertising technology.
Key Takeaways
- Identity is more than cookies; it is a durable, integrated profile built from multiple signals.
- Zeta’s competitive advantage lies in its proprietary data assets and deep domain knowledge.
- Advertising outcomes and attribution are prioritized over simply providing identity data.
- Generative AI is a transformational tool but requires quality data and collaboration to be effective.
- Privacy concerns and evolving technology require flexible yet persistent strategies in marketing.
Summary
- Christian Monberg shares his engineering background and founding of a marketing AI company acquired by Zeta Global.
- He explains the difference between cookies as identifiers and identity as a comprehensive, durable profile.
- Zeta's identity graph integrates multiple data sources to deliver measurable marketing outcomes.
- The company’s proprietary data assets and domain expertise create a competitive moat difficult to replicate.
- Generative AI and evolving privacy regulations are transforming the advertising and identity landscape.
- Zeta focuses on outcome-driven solutions rather than just selling identity data.
- The internet and customer behavior are irrational, requiring nuanced understanding beyond technology.
- Collaboration with OpenAI and other partners is shaping new advertising opportunities.
- Challenges remain in aligning data definitions across teams and integrating operational data for enterprises.
- Zeta remains committed to innovation while maintaining a clear vision in a rapidly changing environment.
Chapters
- 00:00Introduction and Guest Background
- 01:09Founding a Marketing AI Company
- 02:26Generative AI and Identity Challenges
- 03:36Difference Between Cookies and Identity
- 04:44Zeta’s Identity Graph and Customer Outcomes
- 06:05Competitive Moat and Replicability
- 07:23Complexities of Building the Business
- 08:43Data Comfort and Attribution Challenges
- 11:07OpenAI Collaboration and Future Outlook
- 14:35Closing Remarks and Thought Leadership
Full Transcript — Download SRT & Markdown
Speaker A
Uh, well, what I'll say, Nick, is after reading you for quite a long time now, your capacity to think horizontally is impressive, and you tend to be right more times than you are wrong.
Speaker A
All right, guys. Welcome. I have a very special guest today. It is the CTO of Zeta Global, Chris Mongert. Chris, thank you for taking the time to do this and joining the podcast.
Speaker A
Nick, thank you so much for having me. I love the work that you're doing online looking at retail stocks, especially ours. I always enjoy reading what you think is happening in the business as our earnings come out. Um, as well as
Speaker A
the way you and all your followers are thinking about the market. Thank you. Thank you. I kind of want to take a moment here to give you the floor. Can you tell us a bit about your background before we get into some of
Speaker A
these questions? Tell us about where your expertise is and maybe set the floor for the audience so they understand where you're coming from when you answer some of these more technical questions today.
Speaker A
Sure. Yeah. So, my background is as an engineer. I went to school for engineering. Um, back in 2013ish, I decided to start a company looking at marketing specifically. Um, we noticed that all the challenges in marketing were perceived as very hard by
Speaker A
marketers. They were spending a lot of time going through data, needle in a haystack kind of investigations, and I had built AI nanobots and, um, using artificial intelligence to automate the work of robots pretty early
Speaker A
in my career, and I thought, geez, we should be able to apply this to marketing use cases. Let's give it a shot. So in 2013, we went after it. We had a bunch of patented work. We served
Speaker A
marketers. It was off to the races. It worked really well. In 2017, Zeta bought our company, and we were focused on SMBs.
Speaker A
Zeta was focused on big enterprises, bigger challenges, more sophistication to the challenges. Um, but the same mindset still held true. And I'd say the pitch deck that we used back in 2016 could still be used today in a lot of
Speaker A
ways, which is bonkers to me that we're still challenged by the same things that we were 10 years ago or that marketers are. Um, that's really our pursuit today at Zeta, is solving those challenges. We have new tools. They're
Speaker A
constantly coming out. Generative AI has been transformational. Um, identity's been under attack. You know, things are always changing, but we've remained very stubborn in our vision and flexible in how we execute against it.
Speaker A
Okay. Wow. So, you're more or less an expert in this space, you would say comfortably.
Speaker A
I've been at it for a while now. That's interesting. Okay. So, I have a question for you to start off. There seems to be this misunderstanding or miscommunication about how Zeta collects its data and how they utilize their data. Some people seem to think
Speaker A
it's very similar to third-party cookies or browser cookies. Can you delineate how Zeta is collecting their data and how it's different than those traditional cookies in a sense and why that's important in today's world where privacy seems to be at the foremost of
Speaker A
everybody's mind? Sure. Um, you know, you nailed this. It's a huge misconception. We get asked the question all the time. Um, and I think it might be missing the mark a little bit. It's a question born of headlines, not born of the way
Speaker A
technology is really moving. Um, so first of all, I want to say that Zeta is not really a third-party cookie business. Uh, and I think that's where people want to shove us into like ad tech or something. Second, I want to say that
Speaker A
cookies are an identifier, and that is very different than an identity. Nick, you have an identity. I hope to get to know it a little bit better. Who you are, what you're interested in, how your parents raised you. Um, that's different
Speaker A
than your phone number or your email address or your multiple email addresses and your WhatsApp. Those are identifiers. Those are little signals that we can go grab and get in touch with you through them. Some of them are
Speaker A
more important than others. Some are more durable. You may not have changed your phone number in the last 10 years.
Speaker A
You may not have changed your primary email address in the last 10 years. Those are durable. Your third-party cookies are not very durable, but they're still part of the trail you leave throughout the world you're in. Um,
Speaker A
so Zeta is very important technology that we spend a lot of time and have a lot of great engineers. A lot of them have PhDs around the world working on is our identity graph, how we stitch all
Speaker A
of those sources together from Zeta's proprietary assets, whether it's through signals we pick up from Discuss or LiveIntent, which are properties that we own and operate. Um, if we're folding in partners, how that data folds in and
Speaker A
how we not only create a graph, but we continue to make it better. Um, because our customers aren't uniquely just buying identity from us. Um, they're buying outcomes, and those outcomes need to result in a approved bottom
Speaker A
line. B, they need an attribution model that actually makes sense. When they're running attribution, they have to say, "Oh, I actually got to Nick. This makes sense." Um, so we're operating and activating through our pipelines. We're using our identity graph. All of this
Speaker A
has to map up. And if it doesn't, they're going to walk away. Um, that's in a little bit a direct conflict with some other businesses that want to sell you an identity graph, but they're not responsible for the outcomes. They're
Speaker A
not responsible for measurement. They just say, "Good luck with it." So the biggest misnomer in our business is that we are all about cookies. While cookies are part of the signal that we see across the internet, I'd say it's the
Speaker A
underlying integrated identity graph that really sets Zeta apart. Okay. So during that reply, you mentioned data assets and some of those owned assets you guys have, which is quite a robust roster. Now from Discuss to LiveIntent, Boomtrain as you know very well
Speaker A
and now Marold Enterprises. How difficult is it for a competitor today, let's say, to go out and replicate that infrastructure layer? Would it be easy for them to do that or would this be something that's extremely hard? These are rare assets that nobody else can
Speaker A
touch. It depends how you look at it. If you look at it through the eyes of an engineer, replicating a code base is simple. It's getting more simple. If I want to go build a Facebook competitor tomorrow, I kind of could. Um, if you
Speaker A
look at it through building a business size, it's impossible, and I have to look at it through both. Um, I believe it is next to impossible to replicate Zeta's business with the equivalent of vibe coding because of our domain knowledge. Uh, because
Speaker A
the internet isn't rational and maybe for your readers, like you know the stock market isn't rational. We think that we should be able to read the signals every day and create the perfect algorithm to short the right trades, right? Or create
Speaker A
the right profit margin. Um, but the stock market isn't rational. Neither is the internet. Neither are our customers.
Speaker A
So, our customers in APIs, you might give somebody a response and they send back a 200. Sometimes that 200 means a 200. It means, "I got it. I'm good to go." Sometimes it doesn't because
Speaker A
the internet is imperfect. And it's all of that nuanced understanding of the way that enterprise customers work that is encoded into the way that we think about the business. Um, I've been asked this question before and I thought, well,
Speaker A
geez, I'm an expert. Could I go rebuild this business quickly? Uh, say quickly is in a year.
Speaker A
Um, I couldn't do it alone. I would need to bring a lot of the best people at Zeta with me. Um, and getting them to join me would be challenging. Not impo
Speaker A
missing from this ingredient uh list. Second, it's easy to build a tool, very hard to build a system even today, and this is getting a little bit better with time. We're years away from it. Um so, could somebody build like a I don't
Speaker A
know, an email editor that's as good as ours? Yeah, probably. could they get that integrated into the entire workflow of our system that gets uh disproportionately harder to do. Um third is and this is the one that a lot
Speaker A
of investors miss is asking that question assumes that we are standing still and Zeta is doing anything but standing still. I can say with hand to heart that our evolution in AI native development is at the forefront of the
Speaker A
entire industry. We are inventing new tools and building our own internal platform to accelerate us faster than any other CTO, CPO, CDO, CEO that I speak with. Um, and so it's one thing to say, okay, if they start standing still
Speaker A
at the starting line, could they keep pace with us? If we were standing still, I'd say no. But we're not. We've got a finely oiled machine that's running at Mach 10 across that start line and blowing past them. Um, and I think
Speaker A
that's true for most sophisticated businesses today. Um, I'm not part of the SAS apocalypse. Uh, I don't believe that all SAS is going away. I do believe that those SAS companies that present just as simple tools, um, are going to
Speaker A
have a harder and harder time of competing out there. There just going to be more competition for these small micro tools. Buyers, on the other hand, are looking for more in integrated tools, more integrated systems.
Speaker A
So, correct me if I'm wrong. Kind of extrapolate what you just said. You believe that it's not only the own data rails that you guys have, the data assets, but it is the integration of every single one of those assets into an
Speaker A
ecosystem that gives it an edge over the competition that makes it very hard to replicate.
Speaker A
Absolutely. And everything I said about domain knowledge and deeply understanding that data is still true.
Speaker A
Uh when I started at Zeta, somebody used these funny words with me. They said, "Oh, well, I got to get comfy with the data." And I was like, "What does comfy with the data mean? I've got a bunch of
Speaker A
algorithms that will run over that data and tell me exactly how to feel about that data. Uh but it's true. You have to get comfy with the data. You have to understand how it works and how it feels. And there's a certain janicequa
Speaker A
in all of our businesses. The way we approach uh trading for your audience uh the way we approach writing if we're a writer or um or engineering if we're an engineer. And it has to do with that little something that we understand
Speaker A
natively and deeply about uh the domain we're working on, our subject m matter expertise that creates um a small margin but it when it compounds it's disproportionate outcomes from the work we're doing. It's why the greats are better over time. That benefit compounds
Speaker A
and I have seen it happen firsthand at Zeta where our understanding of our own data, the value we get out of it, the way our data scientists continue to compound gains creates more disproportionate leverage over our own business and better outcomes for our
Speaker A
customers along the way. So absolutely both data and technology are uh are steaming forward at an unprecedented pace and um for big complex ecosystem plays like Zeta uh there's no vibe coding startup that's going to take them overnight or I would beg to I would I
Speaker A
would say inside of a 5-year envelope and these days outside of 5 years all bets are off. You know I used to be a a 5-year guy. I make my bets and they're pretty accurate. It gets harder and
Speaker A
harder with a rapid change of pace, but I I can say with safety inside of five years, I at least understand this dynamic. Well, Chris, let's pivot a little bit and talk a little bit about partnerships and potential. Recently at the JP Morgan
Speaker A
Tech Conference, David, the CEO, came out and said that. Well, uh, I'm super excited to mention today that we have now executed an agreement to partner with Open AI and help them run their advertising. We're going to bring our enterprise clients
Speaker A
into the open AI ecosystem and we'll begin to serve ads there on behalf of our enterprise clients.
Speaker A
Can you talk to us a little bit about what do you think the true potential of this partnership is? Where do you see this potential? Talking about 5 years, maybe not exactly 5 years, but maybe 2, three years down the line. Where do you
Speaker A
see this partnership blossoming into? What do you think it has in the future ahead of us?
Speaker A
Yeah. Well, um, good question. you know, two or three years down the line for something as new as this OpenAI, I'm probably not going to comment on, but I can comment on it for now. Uh, the ability to serve ads inside Open is
Speaker A
really exciting to us and our customers. It's a new channel, and we've all read the stats of how much um how much time and space Open AI is taking up in people's uh brains, in their daily activity. So, being a part
Speaker A
of that is a natural opportunity for where Zeta wanted to live. Um, again, Zeta is all about identity and intent and intelligence. Open AAI is a great partner for all of that. So, it's a more intelligent channel than most. Um, I am
Speaker A
really curious about the intricacies of where people are going to spend their time and how it impacts brands specifically. So, how they're thinking about generative engine optimization, how they're thinking about share of voice inside of places like uh OpenAI.
Speaker A
um and how it's driving bottom line for those brands. How the shifts are going to happen from traditional SEO to GEO strategies is a big part of that as well. Um so, uh I guess I'll say this about it. I'm really excited that our
Speaker A
customers get this new channel. I think we're all learning, including OpenAI, uh in the coming months and quarters as to how this ecosystem is going to evolve.
Speaker A
Uh we have a front row seat to it. So, not only are we there today, uh, their engineers are working directly with our engineers, their business teams are working directly with our business teams to help understand where new
Speaker A
opportunities evolve. If there's anything I know about OpenAI as a company is where they start is nowhere near where they are going to end. Um, so I'm just excited to be part of the ride with them as well as with the brands
Speaker A
along the way. That's quite the cliffhanger. Now, I want to shift gears a little bit away from OpenAI. There was another partnership that was recently announced by the Zeta team with Snowflake. Now Zeta is now officially becoming a
Speaker A
partner in developing the OSI, the open semantic interchange. Can you talk to me about the potentials here, the capabilities that you're pursuing for data transfer, that data bridge between enterprises and that Zeta ecosystem?
Speaker A
What do you look to accomplish out of this partnership with Snowflake in the long term?
Speaker A
Sure. Great. So to start with AI needs shared meaning not just shared data. Um and this solves both of those simultaneously.
Speaker A
Um we know AI is only as effective as the data it has access to. Um the challenge is the way one team inside of an organization or across organizations defines data is always different than another team. Where I see the biggest
Speaker A
opportunity for this is uh inside of a single brand that we work with today. Uh we come up with this every day and the use case of a product's hot, right? So we're doing marketing for some company, some uh
Speaker A
national retailer, and some products really hot and our algorithms are popping off the chart saying, "Oh, we should promote this more. People want to buy this." We should be able to very quickly see how's the inventory doing.
Speaker A
OSI, where were we? OSI. So OSI matters because AI is only as effective as the data it can trust and it really needs to trust that data.
Speaker A
We already know that hallucinations are um you know plaguing the industry at certain times without deterministic understanding of that data. So uh one system defines one interpretation of data for an entire organization. The example that I I like to think of is
Speaker A
inside of most brands we work with, the inventory management system and the marketing system, they're managed by two different teams. And uh you know, we're in marketing, so let's say that a product is really popping off for some
Speaker A
national retailer. And uh the algorithms are screaming, sell more of it. This is this is going to help them hit all of their goals. Yet somewhere there was an inventory management system that was choking saying oh no we are
Speaker A
running out of that product what are we going to do. OSI allows us to anneal those two different systems automatically. So Zeta spent a lot of time um and you know one of our specialties is annealing those systems
Speaker A
through services and for deploy engineers. They go write scripts and write transformation layers. uh OSI allows our brands and ourselves and Snowflake to establish one rule of the road that we can all abide by. This clearly drives better outcomes for the
Speaker A
brands that we are working with collectively and uh for Snowflake I think that OSI was the right way to get into taxonomies across a lot of different verticals. I think it was a very bright move for them so that they
Speaker A
can power businesses holistically rather than just working with small nuanced um teams inside of those businesses.
Speaker A
This is very interesting because me as an investor who's studying Zeta day-to-day, I see this as an opportunity to expand Zeta's knowledge perhaps to create or its ability to create intelligence not just in that consumer domain but also in that operational
Speaker A
domain and marry the two together almost creating a operational consumer ontology layer for that enterprise. Obviously the enterprise would obviously have to opt in to such services and data transfer.
Speaker A
Am I thinking along the correct lanes here when we're we're wanting to expand Zeta tan potentially further into business intelligence? I know that this is very exciting. So I'm asking you more directly, is this something that you're interested in pursuing? Am I thinking
Speaker A
along the right lines? Uh well, what I'll say, Nick, is after uh reading you for quite a long time now, your capacity to think horizontally is um is impressive and you tend to be right more times than you are wrong. Um
Speaker A
I'm a technologist and I think in terms of abstractions as much as often as often as I can. So our data platform uh the streaming nature of our data platform the way we perceive data process data it was built in a way that
Speaker A
was very uh generic uh in many ways and over the last many years people have said well that that approach won't work for us and I say well show me how it won't work for you and surely within a
Speaker A
hour's conversation not only does it work perfectly but we have everything documented and well understood on how to process data whether is for uh retail use cases travel use cases, inventory management use cases, couponing use cases, uh store use cases like where
Speaker A
they are when they're open. And so your observation could that technology um support adjacencies in terms of different data types is right on. That's the way you would build something today.
Speaker A
And you see that elsewhere in the industry as well. People that are building really smart data layers are building it almost precisely in the in the same way. I'd like to say I was clairvoyant, but I think we just kind of
Speaker A
followed the tea leaves and understood uh the right way to build technology um as it was emerging.
Speaker A
Well, I'm super excited for the future, Chris. This has been awesome. Thank you for the time. I I've had a great time talking with you, talking about the future of Zeta. I'm very excited about what what's to come and perhaps I'm
Speaker A
going to see you as a live. Oh, I hope so, Nick. Yeah, thank you.
Speaker A
And again, uh I love reading what you have to write and uh as well as your followers. your followers have a voracious appetite for what you're writing and they're engaging um deeply.
Speaker A
So, watching from the sidelines over here is always a pleasure. Always happy to join you and thanks for the opportunity for including Zeta in all of your thought leadership.
Topics:Zeta GlobalChristian MonbergAI in advertisingidentity graphdata privacygenerative AIOpenAImarketing technologyadvertising outcomesdata integration
