Good morning, good afternoon, good evening, good whatever time it is, wherever you are in our amazing planet. I’m Malcolm Hawker. I’m the host of the CDO matters podcast and I’m thrilled to be joined today by Christian Ward. Christian is the chief data officer of Yext.

We’re gonna talk today all about marketing related data serving the marketing organization as a CDO. Marketing is one of our biggest customers and for many of us our biggest customer and they have a voracious appetite for data. They have a voracious appetite for insights, demographics, you name it and we’re gonna talk all things marketing, maybe a little marketing analytics. We’re certainly gonna talk about third party data.

We’re gonna talk about the things that are changing in that world and how is we as chief data officers can best serve our marketing friends in an era of change. Christian, thank you for joining.

Hey, Malcolm. Thank you for having me.

Awesome. When Christian and I did our prep call a few weeks ago, we’re kinda going through the bios and the and the experience and and we we we have lived very complimentary corporate careers.

Christian did a long stint at at Thomson Reuters. I did a long stint at Dun and Bradstreet, Kind of flirted around the online advertising world. I I worked for a decade in helping build out the online advertising infrastructure at America Online, this little startup many have now forgotten.

But given given given our shared tenures in the space, we had a lot of overlap and a lot of commonality. Let’s let’s start let’s start with what’s really kind of the biggest change factors that you see, Christian. What’s what as a CDO should I be thinking about in terms of what is changing in the marketing world? What’s front of mind for CMOs? What are some of the things that are keeping them up at night from a data perspective?

So, look, I think this is probably the most dynamic time we’ve seen in probably fifteen to twenty years. So when we think of from a marketer’s perspective, so most CMOs for the last fifteen to twenty years have a playbook that they’ve been playing for a long time.

They know their tech staff. They know what they like. They know what they don’t like, and they have very strong opinions on this is how you do these things properly.

What’s happening now is they are waking up to not that they they’ve always liked data. Like, they they’ve always they’ve I mean, they they can be the the largest consumers of dashboards in in in in in the business itself. However, they’re starting to get a different understanding of the role that data plays to achieve all the objectives they want in terms of the customer journey. So from awareness to discoverability, all of that could basically be boiled down to how they dealt with Google or paid for paid media spend. And now it’s sort of flipping. Right? That that whole world is now moving into this different area where the AI is stepping in and becoming a much more knowledgeable middleware to the consumer.

And so they want new data. They know they have to change their KPIs. They know they’re not tracking all the right things, and it’s really cheap data officers that have to help them understand what’s available, what can be gathered, what should be gathered, what shouldn’t be gathered is probably one of the biggest questions, and then build off of that. So for the marketer, I think what we’re seeing is is we are now in almost every c suite conversation with CMOs at the companies that we support at Yext because everything has changed, and it’s so critical to understand how do I prepare for an AI first world when it comes to marketing.

So how much of that role is or should be consultative on our end? Meaning, you know, historically, I I would say that we haven’t really been really great consultants to our marketing organization because they kinda knew the world of Google. They knew the world of SEO. They they knew what they needed to do in order to make sure that they showed up in the first three or four results in Google search. How much do we play a role in consulting to CMO organizations about the things they need to do in order to influence AI in the right way like they used to be able to influence Google?

Yeah. It’s so it’s a great question. I actually think CDOs could be taking a much more front and center role in this because the the real difference here is classically, marketing, they think about things like SEO and showing up organically, and then they use paid as a a speed pedal. Right?

So they can speed things up if they need to. What’s happening here is marketers are starting to get a better understanding of it’s not just knowledge about your business that works for SEO. It’s that the AI is going to rapidly take over the the top and middle of the funnel. And and to some extent, some of the deals we just saw with Google Universal Commerce Protocol and that Shopify has their own with ChatGPT with Stripe.

Even the check-in is now happening via an agent. Right? So what starts to happen is they didn’t I I would say many CMOs struggle with when an SEO or the marketing team is saying, hey. We need more knowledge about what we do on all of our web pages and everything else.

The CMO has not necessarily partnered with the chief data officer to get all that data from the organization. And and that that I I am seeing that change every day. It’s actually really exciting. A lot of CDOs were sort of in the corner, and they’re charged with managing or the security of the master data management.

That makes sense. But that was a lot of times like the CRM world and maybe the financial world.

They’ve all got to move towards this marketing department and say, listen. Did you know that in our CRM alone, we have so much knowledge that would help if you used it in the data that was fed to an AI? For example, in the CRM or in, let’s say, in the CDP or in some of these other core data assets, we might know different times of the year that different products are set selling more to different audiences in different regions, and we can actually help them amp up that knowledge. But the marketer has to then structure it through their tools to get it into the AI.

So so because that is the that is the highway by which that data can get out there. So I think what you will see is many of the chief data officers should take a more consultative role with the marketing departments to help them understand what is the library of assets. Like, I’m always amazed. Most CDOs, they don’t have a data asset library with every field, the speed, the the accuracy, the consistency, the fill rates.

If you can sit with your marketers and even use an AI for this and say, what are all the pages on our domain and all the content we’re putting out there? What are we not talking about that we actually know about our business that would help the marketer? So I really think it’s a it’s a a pivotal time for those two groups to come together.

Well, it’s it’s interesting you should say that. I totally completely agree. I I I share I’ll share a little story, a little anecdote here. I went for the first time to a conference called Kilometers World, knowledge management world. It was in DC in December.

And I walked around the vendor hall and I felt like and I was giving a keynote speech about about the importance of integrating data management and knowledge management. So so obviously, I’ve kind of already tuned to this world. But I walked around the vendor hall at this conference and a whole room of vendors I’d never heard of before. Like literally a hundred plus vendors that I’d never heard before.

But I walked around and every single one of the banners was all was a was a problem statement or a value prop. We would say in the data world that they’re saying in the knowledge world Yeah. From a vendor that I’d never heard of before saying, hey, you know, you need to get your hands around, you need to create trusted, more curated data. Right?

We need we need to we we need to help make sure that we limit hallucinations in AI. We need to do all of these things.

And they were from vendors that I’d never heard of at a conference that I’d never been to, full of literally thousands of people that I’d never met before, all saying the same things that I’ve been saying for thirty years, they were saying it through the lens of knowledge.

What I’m doing what I’m saying here is I’m paraphrasing is what I’m hearing you say is that the world of knowledge management, content management servers, knowledge management servers, even even all the stuff sitting behind SEO, all of that needs to come together with the world of the CDO. How is that all gonna happen? Knowledge management and all of that marketing is is a slightly contained world, but there’s like learning management systems out there that are completely separate. Content management systems that are completely separate, totally outside the purview of of Aceto’s where there’s all of the product videos and sound clips and all. How do we what’s how do we start to kinda bite this elephant off one one one bite at at at a time?

So so one of the best ways to do this is actually using AI. You you basically give a problem statement. So so what I often try to show people is because they’re sort of like, how do I get started in this? Understanding what is the difference?

For example, you might have seen all the articles in the last two weeks about context graphs. Yep. So there’s knowledge graphs. Now we’re talking about context graphs.

Trying to understand why did we decide to do that? We have the data on record. We have the CRM. But why did that customer get a thirty percent discount?

It’s like, well, because it’s the CEO’s nephew. Right? Like, there’s there’s things that are not in the data that need to be stored as context. So I I think that is a little bit of a hint of where this is going, which is all of those different systems, we have experts at most companies that manage each of them, then you have a CDO who’s sort of responsible for knowing where they are, what we use, what’s available.

But for the marketing team, like you said, they’ve had their own metrics. It is somewhat contained. What we have to start to understand is in the customer journey, the customer, if you really think of it as a connected connection of nodes, let’s say I say, I’m a surfboard company. The best thing to start is you go to the AI and say, what are all the things that a consumer might consider about buying a surfboard?

And it gives you fifty questions. And then you say, that’s great, but can you tell me the next hundred and the next this? Now what systems is that data typically stored in in my type of business? And what starts to happen is it will actually build you a knowledge grant.

It will say, well, finance will have this, and this will have this. There’s a great paper I advise everyone we can maybe we can put a link onto it. There’s a great paper by a professor named Sean McMahon, who I runs the entrepreneurial program at Elon University. And the pay name of the paper is Lean Data.

And what he tries to explain is most businesses don’t have a good grasp because the data’s buried in twelve other systems. What is the leanest data necessary to make the best decisions both as management, but then also as the consumer?

And so what starts to happen is you can kind of ask like, if I ask you, do you know are you a pilot? Do you fly?

Okay. Love you. So I’m not. But I’ve I’ve I’ve given speeches on lean data before, and I had a pilot come up to me and goes, you know, it’s pretty interesting.

If every other instrument on a on a plane went down, wasn’t working, what’s the one metric you need to fly the plane?

And I I was like, there’s only there’s one? Go there’s one. There’s one lean data element for everything.

And I I said altimeter or the the thing that does the pitch in the yaw and the The horizon.

So so I was like, is it the altimeter or the horizon? He goes, nope. Nope. And I’m like, what is it then? He goes, airspeed.

The one thing the only thing you need if all the other experiments failed, you could fly a plane if you have the airspeed because you know he goes, you can’t know that. You’ll be in a plane. You don’t know if you’re doing two hundred and fifty miles an hour or a hundred and fifty miles an hour. But if you’re not going fast enough, you’re gonna flare.

You’re gonna have a or you’re gonna go down. And he’s like, but with airspeed, you can. It’s a great demonstration of what I mean by lean data, which is there’s too much data. It’s like data gluttony in all of these systems.

The best thing a chief data officer can do is help extract the lean data that makes the best, fastest, most efficient decision for the marketing department on where to focus, what to focus on, and how to do it. Because remember, many of these systems in their world are gonna lose most of the data they have once AI becomes the all the cookies, all the surveillance, all the privacy stuff Right. That’s gonna be AI. And so they won’t have half the metrics, and they’re gonna have to rely on bringing the lean knowledge about what they do, how they do it, what their service is, where they deliver, all those things that they don’t really think about.

They think about the brand. That’s not gonna work when you have a customer talking to an AI that just wants the answers before they ever consider your brands. And that’s that’s a that’s a big change. I wanna say, like, I think for many marketers, it it it’s got them scared, but it it’s a good thing for everybody if they follow that sort of mean data approach.

Well, I love it. And that’s something that I’ve always talked about over the years when I was an analyst is something I certainly espoused. Right? Like, take an MVP approach to everything.

What’s the minimum I need in order to solve this problem? Yep. However, I think the biggest challenge here is not not a systems problem and not even organizational problem. Right?

Like, I’ve been I’ve been saying that CDOs need to go and and kind of wrap their arms around knowledge management functions, go find those folks, try to find a way to integrate them to their organizations, whether that’s physically or dotted line. I it doesn’t matter. Right? Like, if there’s somebody out there that is managing a content management server with a whole bunch of data related to your customers or whatever they’re saying in their customer service interactions, whatever it is, You you need to go and partner with those folks and get closer to those folks.

So I don’t think it’s necessarily a systems problem, and I don’t necessarily think it’s an organizational problem. I think the biggest challenge here is organize or I should say, more of a mindset problem because what you just said is is that a few different things and it’s all Pareto related, Pareto adjacent, know. Yes. Eighty percent of value is gonna be in twenty percent of the systems or databases or processes, but that eighty percent is gonna be good enough.

Yes. And for marketers, it often is. Eighty percent for a marketer would be, hell yeah. Sign me up.

Yes. But for a CFO, maybe not so much. Yeah. Herein lies the problem is that many of us get hung up on these ideas of perfection and manage to exceptions and try to to to manage to the highest risk, which is almost always the CFO, and in the process kind of distancing the the CMO and the CMO is like, you’re not helping me, you’re not serving me, I’m gonna go do my own thing and I’m gonna go buy a CDP, I’m gonna go write the contract with D and B, whatever it is, I’m gonna do my own things because you’re driving me nuts because your your data quality level’s up here and mine is down here.

Yeah. How do we how do we get to a point where CDOs take more of a context driven view of the world?

Yeah. Was well, number one, think CDOs for years have sort of done themselves a disservice because they have stepped back to more of being that systems or data relationship management, and they have not focused on the front line. This is this is their opportunity. To me, if you’re a chief data officer and you have access to all of these systems, your job right now is illuminating what those systems can do to help the marketing team.

And so and I’m not saying people don’t do that, but to your point, a lot of times it’s very focused on point a, point b, how do I go from database a to b, but it’s not really understanding the problem for marketers. You you take a step back as the CDO, and you say, listen. What is what are the five biggest drivers of how we market and we gain our customers, and then how do we keep our customers. You have to put yourself in their shoes.

Then you relook at all your data assets and say, I’m gonna rank all of these super easily. Like, I mean, literally like like baseball cards when we’re kids. Like, got it. Need it.

Got it. Got it. Need it. Really fast. You do that, and you say, okay. These work, and the fill rate is good enough.

Get used to that phrase. It’s good enough, and it’s accurate enough, and it has a way to passively update. That’s really important. If we’re waiting on other humans to enter data and that’s what you’re gonna use, you’ve already built in a dependency.

Look for data that’s sort of passively gathered and is always available. That’s usually the best to start with with the marketers because marketers, generally speaking, they’re not gonna stand for something that updates every three weeks because Paul and Sally finally got around to up CRM. You’ve got to work with data that’s much more, you know, passively gathered. It’s always there.

But if you get that, then as a CDO, also look at external data sources. So I I always use the weather as a joke. But go pull down the weather channel’s API, start analyzing it, build correlation studies, and start to show marketing that, you know, you’re not using this in how you’re targeting and how you’re explaining what we do, I really think there’s an opportunity there. Because as you point out, a lot of CMOs also split off from the IT stack.

So CMOs now spend more on their own technology than the IT team spends on company technology. So that I think that happened two years ago where that budget switched. So you’re literally talking yeah. Sorry.

Go ahead.

No. That’s wild. Like, that’s that’s that’s wild. Yes. Continue. Yeah.

Yeah. So I think what we’re saying here is is the chief data officer is meant to be an ambassador between these departments who can help them crosswalk to more assets. And it’s again, like I said, I I think it’s you have to look at each dataset, and I definitely recommend AI. It’s very good at this.

That will show you interesting correlations or corollaries between data that you probably aren’t thinking about. The other piece of the puzzle, which a lot of people this came up in the context graph blog wars over the last couple weeks, is you can’t just look internally. A lot of CDOs are so worried about what’s happening internally, they don’t realize that your company’s direct competitor just opened a store next to every one of your stores in the last six months. They’re not thinking externally, and so that data also affects marketing.

So you’ve gotta pull in external competitive data and organize and get lean on the internal data, bring those to the CMO, and work together to have a plan of how are we going to leverage this. And, again, it’s not just AI. It helps in search too, but it might help even in product development, what what areas you really market as your differentiator versus the competition. All of those things are in the data.

It’s just the CDOs a lot of times aren’t thinking that way, and so we’ve got to change our mindset too.

Yeah. For sure. That mindset thing I think is key. Right? Just embracing and and getting comfortable with the idea that close if for a marketer, close enough is going to be good enough.

And for a large company, a one percent lift is millions and millions and millions of dollars. Right? Yes. So so we don’t need ninety five percent.

We don’t even need ninety percent with a lot of these processes. We just need a few percentage point gains.

Yeah. And I think if we as CDOs can build these partnerships, get a little more consult consultative, can start exposing and maybe even opening up the data that we manage or control access to back into the marketing organization, which you may not have been doing, we can start to show some of these kind of moments that are gonna help lead us to some of these high value use cases that were probably always sitting in front of us all along and didn’t know it.

Absolutely. Yeah. And the look. The I I think the I think the CDOs that really do well, they have they have business experience as well.

They didn’t come just out of IT, and they’re not just a data analytics dashboard person. It’s that you’ve either ran companies, started companies, and you ended up really loving this area. People that do that, they you have this unique time. Like, again, a lot of data has been trapped in these systems, and then you have, you know, people not talking to each other.

You really have the chance to almost build your own understanding of what this your your company can do. And and the reason why I say that’s maybe now more than before, anytime you wanted to use data in the past, you either have have all the analytics, the team members, the dashboards, the tools, the BI, all that stuff. I’m saying take a step back and forget about it for a second. That stuff’s great.

However, from the marketing perspective, where a ton of the budget sits, what what is the customer seeing? What is changing in the customer journey that marketing is absolutely worried, concerned, or excited about? And and talk to them. Find out what like, for example, a lot of times we sit with our our clients and we say, how much are you using AI?

Have you tested how your brand is showing up? Have you tested, like, what is it showing up for? Why are people talking to it? Because what’s starting to happen is search, which is one of the biggest areas of this business for the marketer, that’s becoming a a dialogue.

And dial you know, search comes from the Latin circus. It’s just circle. And when you encircle that’s why search actually doesn’t wanna give you the answer because it makes more ad revenue if it kinda gets you in the circle of an answer, but it doesn’t give you the answer. Dialogue is dialogueos, the Greek, through words.

And what that is saying is is it s I say, hey. I want a surfboard. It goes, were you a beginner or an expert? I’m a beginner.

Well, what what type of board are you looking for? Oh, and how much do you weigh? Those three questions alone get the person way down the funnel in the customer journey. That’s not search.

That means the product catalog, how much each surfboard weighs, whether it’s beginner classification, those data facts become the actual search process, the dialogue. And so, again, I’m not saying they don’t have that data on the website somewhere, but marketing doesn’t necessarily think about that. The commerce division might, but that’s not the same way. They might need to market their beginner surfboards way more in the northeast based on the finance data, the product data, this data, and they’re they’re not getting that message.

They’re still branding and building this big brand. It’s like, you really should be different things to different regions or different product specifications. That’s a huge opportunity for the CDO to walk them down a new way of thinking about how data is gonna be used to bring the customer through a dialogue to the proper answer.

Well, so I’m having a what’s old is new again moment. Right? Because everything that you just said, as as transformational as it could be, are things we were talking about ten years ago, twenty years ago when we first twenty years ago when we first started talking about personalization online. Right? Yeah. And and getting away from these kind of extremely broad demographic cuts of the world into more hyper personalized, know, led to the cookie and a whole bunch of other things.

Ten years ago, we started talking about, you know, three sixty degree view, single view of customer like that and and here we are again, because everything you just said is is tied off of the to me, the beating heart of all this is identity.

Yes. Which which is super super hard. Yeah. And in the midst of all these great conversations about context, which which I love and and absolutely true and some of my contemporaries are talking about twenty twenty six being the year of the ontology and Yeah.

You know, all of this stuff and and Again, my goal is new again.

Yeah.

Yes. Yes. Because twenty years ago, it was, you know, it was the World Wide Web or thirty years ago, the World Wide Web Consortium and the Semantic Web and all of these things that we were talking about thirty years ago. Crazy.

But here we are again talking, I think, indirectly about identity. Because you can have all the context in the world, but if John Smith is not the same thing as Jonathan Smith, you got a problem.

Yep. Yep.

Right? How how do you how do you do you see this playing out in the AI world? Do you see AI playing a role here in identity management? Or I mean, you basically just said you can’t put a human in the loop.

I mean, you Yeah. I’m I’m paraphrasing you now. You’re not gonna if you can’t wait for John S. Eli to put something in the CRM, you certainly can’t wait for a data steward to to say yes or no to a transaction.

How do how do you see kind of identity playing out here in a world of AI?

Well, so look, we’ve had many situations throughout the history of humankind where the technology wasn’t ready, but we could already envision what we wanted. And so we used other means to try and achieve what we wanted.

And and this is a case, absolutely, where everything we tried to do for personalization I will I’m just gonna throw it out there. Bit of a privacy walk. I hate cookies. I hate surveillance. I hate tracking. I hate all of it.

Because what we were trying to do and I’ve told clients this. I’m like, please don’t buy a CDP. In the world of AI, it’s not going to work. And and here’s why.

We’re all trying to go, oh, and non anonymously then reidentify them. Do this. It’s surveillance. And by the way, I’m not saying as marketers to to marketers that you should want that.

I’m saying the way you went about it is pretty creepy. Like, for example, if if if put it this way. Let’s say you’re a marketer, and your grandchild comes to you, and they’re like, you know, hey, papa. What did you do for a living?

And you’re like, well, I I I found out where people were, and then I’d see what they read. And then I’d see who they talked to, and then I’d follow them around because I could track them. And then I’d I’d gather all this info and then the kid’s gonna go, grandpa was a stalker? What what what is it’s when we do it online, Malcolm, when we do it online, it’s okay.

It’s not okay. And what’s happening is the AI will deliver personalization. It’s just that the marketer at the corporation won’t own the data. And it’s really this is the other reason why I think CDOs have a really big opportunity right now.

Marketers today think that they’re tracking their surveillance, their cookies, their reidentifying all works. You and I both know it doesn’t. It’s directionally accurate, but it’s not great. The new world is be gonna be my AI. My personal AI has memory about me.

It knows me. And so it goes to the surfboard website of ten surfboard websites, and it says, Christian is this tall. He weighs this much. He’s a terrible surfer.

He doesn’t want a he wants a foam board, but not that like, it’s gonna walk through, and it’s gonna remember those conversations, and it’s gonna go, and it’s gonna want answers from that website and from your brand. And if you don’t provide the answers, you’re out of the conversation. You’re out of the dial. And so for marketers, they think they should know where I am and all that data about me.

You don’t get that until I buy from you. Then you get it. Now, again, this is very uncomfortable for marketers because they think they’re entitled to know that I visited the website. I looked up this.

I did that. That stuff’s all going away. Walmart’s a great example of this. You saw the article with Walmart.

They said they they did the deal just before Black Friday, that ChatGPT via Stripe, could check out with Walmart just by talking inside the chat. Right?

I do not.

Oh, so so Walmart said twenty percent of all their traffic now comes from ChatGPT referrals. And before Black Friday, they did a deal where you could buy anything from Walmart right through OpenAI. You never had to go to Walmart’s website. Now here’s what’s crazy about that.

Oh, interesting. Yeah. Agents of Walmart on the back end okay. Interesting.

Yep. So Walmart. Wow. Largest so if do you remember Jet dot com? They bought Jet dot com for six billion dollars back in, like, twenty sixteen.

It was a Hoboken company, a great great little town, great little company. They bought it six billion. They probably put in hundreds of millions of dollars since then into their website, their tracking, their survey, all that. Everybody did.

They just said, you know what? It doesn’t matter. As long as our website gives the accurate data to the AI, we just want the customer. Because remember, they only get a transaction.

They don’t get any of the dialogue. They don’t get to see what I asked AI to get the Walmart that might get for my kids. So they’re giving up that because they want the business, and and that is look. Amazon has not done that yet.

A lot of retailers have have blocked AI. They wanna do it themselves. I don’t think that’s gonna work because my AI knows me, and it knows what I’m going to wanna buy next in a way surveillance could never deliver. And I think that’s actually net positive for the consumer, and it’s gonna it’s gonna feel very uncomfortable to marketers.

So they’re gonna need CDOs to help them understand what’s the data I can have, and how do I leverage that to show up more in these dialogues.

So my head is kinda spinning a little bit because what you just described is a complete disruption of several multi billion dollar businesses. Yes.

Yes. Yes.

So if I’m, you know, Epsilon, Acxiom, insert extremely uber large aggregator here, data on boarder, whatever.

If that’s my core business, what you just described is a massive disruption to that. Yes. Not to mention a massive disruption to how I as a marketer sell my wares. Right? Yes. Because now I’m no longer trying to appeal to the, you know, whatever, pick a demographic thirty five year old housewife or the fifty year old semi retired nerd who lives at the beach in Florida.

What you’re trying to market to are those you may be how do you even market to an agent? How do you even like, what what what how does that even work? What sorry. My head is spinning a little bit.

I haven’t I haven’t figured this out market you market to agents knowledge.

That is the whole point. You provide the agent with completely consumable, in-depth, highly accurate, and ridiculously consistent knowledge about your business.

Yep.

And the AI agent is gonna learn that you didn’t cost it a ton of effort to get the answer it needed. You are the lowest compute, highest return on dialogue continuation. Remember, the AI agent is optimized for continued dialogue and transaction, you know, in a way. Right?

It’s not like search. Search is optimized for an ad rendering and an ad click. This is very different. So if I’m if I’m those big data companies, number one, don’t don’t persist.

But you and I, think we talked about this last time. What’s the most terrifying thing for anyone that says profile sell big profile data? And you remember this.

Yeah.

Yeah. They would only be, you’re selling to me Dun and Bradstreet. I’m like, cool. Pull up my record.

Pull up my No.

And I’m like, it’s a it’s a table. You know? And it’s like, that’s wrong. That’s wrong.

That’s wrong. The data was never that good at that aggregated level. The AI, on the other hand, has both context, continuous memory of everything happening. And so the the real battle isn’t gonna be trying to reidentify me or target me as a group.

It’s gonna be, does the AI have the knowledge necessary about your business, your brand? What are the ingredients? What are the allergens? Like, what what charities do you support?

I don’t care what it is. We all make decisions. We’re human. So we we do not function economically perfectly.

We make decisions on emotion, irrationality, all these other things. And so if that’s the case, then you have to give it all the knowledge. So I’ll I’ll give you a good example. I asked my AI the other day demonstrating to a financial service firm.

I said, hey. I’m looking for a new financial adviser. This is one of the largest companies in the world. I’m looking for a financial adviser.

ChatGPT comes back and says, well, I recommend I think it was Charles Schwab in Palm Beach.

And I go, okay. Why? And he goes, well, they offer this, this, this, this. I’m like, yeah. But everybody offers those. Why? And he goes, well, remember you told me that your daughter goes to the University of Florida and that you also go to the University of Florida.

Go Gators.

And it goes and that you’re a big Gator fan. Well, as it turns out, the four principals at that office are huge bull Gators, and they love they’re always so I thought maybe you’d hit it crazy. Now if you what by the way, this is a little tidbit for the financial service listeners. If you are not predict what college you went to, what what golf club you belong to, all the facts, that is how people pick financial advice.

It’s insane. But the AI is mimicking that or using the knowledge where you went to college for that answer. Again, I’m not saying that’s how you should pick your financial adviser. Please do more research than that.

I am saying it’s amazing that because those with Schwab Schwab had put that data out there, the knowledge of where everybody went to school, that affected the customer journey through the AI. So knowledge is the way you market, and it’s gotta be, like I said, accurate, very rapidly updated if it changes, so real time updates, and as consistent as possible. That’s what the AI is looking for. Remember, AI is just a math solution to a probability problem.

The more something has the same instance of the same data, it gives it confidence that if I say to you, kill two birds with one Stone?

You know that that is your your brain, from a probability perspective, is like a large language model. It just goes, the probability, because I’ve seen this so many times, is that’s accurate. That’s how you win. It’s it’s really simple.

What you described is a major pivot from a data perspective in how we as CDOs would best support that marketing organization. Because what you just described is to me, I wouldn’t even call that a data management solution. That’s a content solution. Right?

That that’s content. I’m aggregate, you could call it knowledge, fine. Knowledge, content. But I’m aggregating and creating and and curating this corpus of information that would help differentiate us.

You’ve gone down the path of kind of transparency and, you know, saying where our executives went to school and our charities and all those things are all are are also going to be important. But this is a very very very different model. I’m wondering, you know, what are what are some of the baby steps taking that lean kind of MVP approach to start to migrate towards that model?

I think in my head, what I start to kind of come to is, you know, could be best described as as one click, Right? How do I give the content that the AI wants right now, right, to make Gemini happy and to make Claude happy? Is that a good starting point to kind of start to think of the world through that lens and where where it’s not SEO anymore, it’s it’s it’s these it’s, you know, it’s these LLMs and how do we get what we want out of them through a content management strategy?

Yeah. So one of the best things I can recommend to people is start with where you are. So if let’s say you’re a CMO or a CDO, one of the best exercise I tell people to do is go ahead and scrape just control a, control c. Take everything on one of your pages on your website.

K? So you just go to your own website, control a, control c, and then paste it into Claude or Chachiopty or Gemini, the premium models, please. Stop being cheap. Buy the models.

When you do that, what and then you ask it. Show me every structured piece of data that could be structured with, like, schema markup or whatever on this page showing every piece of structured data that would help a customer understand what we do and how we do it. You basically can ask it sort of where are we today? You would be amazed.

So once you’ve done that, you then say, what are all the other pieces of structured knowledge that would help an AI help a consumer choose this page? And it will literally tell you, and it’s really interesting what it comes up with. For example, I did this with one of the large telcos, and they had their mobile phones. They had their channel lineups because they also do, you know, whatever, streaming.

And then they had and, basically, what what the AI said was, you know what’s missing?

Where’s all the information on this page for how bundling these things together in in this ZIP code would save you more money? And where’s all the it it literally started offering other things as to this data is not present, and that’s the beginning step. Because then you go, okay. We never thought I’ll give you another example. Was talking to a gas company. So a gas company that they don’t have their pricing. Most gas companies, by the way, do not have the pricing at the stations on the station level page.

Did you know that ChatGPT recommends gas based on what it thinks the price is? Because that’s what a human would do. Gas is an almost exclusively price driven purchase. And so, literally, I was like, the AI just told you what you have to do.

Go go to your gas prices and put them on the page. Because what what what what you’re lacking is you’re getting in this argument of, does that belong on the page? Does it not belong on the page? Should we put it over here?

You’re missing that that is not about you. It’s about the AI being able to interpret the data easily and quickly with little compute so that the answer gets to the consumer.

Well, so instead of guessing what this the AI wants, ask it what it wants.

Yes. It it sounds like good advice for my marriage as well.

Yes. Yes.

I’m I’m wondering how governance, small g, maybe big g, just just governance in general. I I’m this is something I’ve been thinking about for three years now. I’ve I’ve got my ideas. You talked about, you know, what’s out there for me to structure.

I when I when I think about structuring things, you know, marking marking things up, getting things into more semi structured that may be in more text form today, I start to think about governance at scale or how to enable governance at scale. But the old war way of doing it, like the old data quality rules, the the old kind of deterministic rules don’t seem to fit very well in this world. It’s something I talk about often on LinkedIn. How do you see kind of governance playing into this, particularly through the lens of when something might be fit for purpose or not fit for purpose?

Well, so it it depends as always on what we’re what type of dataset we’re talking about. So at for example, Yext, we mostly work with data, like I’m talking about knowledge that a consumer needs to know about your business. So I don’t I don’t actually manage the the transaction files, the consumer records. That lives in a different area.

So probably the one of the fastest ways you can do this is kinda separate the the datasets into this really is no risk. This is mid level. And most people have this through their data audit and their data privacy attorney or their DPO. They’ve already categorized this.

And so one of the things that I think we start with is these are fit for purpose in that they are all about the customer journey choosing us. That’s like that’s like free to air, free fire, go. And you there’s a lot to be done just there. The next level is pre login or pre identification.

What are some other things? That might be discounts. That might be pricing strategies. That might be other things that you’ll still expose, but you don’t necessarily want them everywhere.

That might be a little different in how you handle that information. But by categorizing them as the CBO, what you’re sort of doing is is you’re you’re almost enabling these other groups to get a little more creative with how they merge the data together in an outward facing way. But when it comes to big g, look, I I think Europe is a good example. I think it’s over g and and regulation wise.

Look. There’s there’s look. There’s only one model in Europe right now, Mistral, and and it’s it’s really not much from them in the last eighteen months.

You know, you’re we have to be very careful about sort of the the devil we know, and then they think the devil this devil’s worse. I don’t know I don’t know that that’s true. I look. At least with AI, you can switch from one AI to another almost at at the drop of a hat.

You you literally, as the consumer, control what you tell it, what you don’t. Think of all the bad data records that are affecting who markets to you and who doesn’t, and they don’t even have the right information because it’s gathered, you through surveillance. Like, I still have Casper mattress pads following me. I bought a Casper mattress, like, twenty years ago.

Like, it doesn’t make any sense, and it’s it’s like, this we know, and now we’re gonna overregulate this. I actually think the AI memory agent for the consumer relationship is a far better thing to let happen than than another cookie banner privacy pop up that drives everyone nuts that does nothing. So I I’m hoping regulation doesn’t stop us.

Well, yeah, that’s just probably a separate episode talking about the connection between investment and regulation, venture capital and regulation, all of these things. Probably don’t need to go there. I I think I think I I generally agree with you that this genie is out of the bottle. And I don’t tend to look at the world through this kind of this, I don’t know, bipolar, we gotta get there before the Chinese get there.

I I mean Yeah. Just don’t see it working that way. I think it’ll be a little more nuanced and a little more subtle and if we ever get to AGI, I’m not even entirely sure we’ll know because the machines will be in control at that point. But anyway, that aside, tell me a little bit more about Yext and tell me tell me the kind of the I I found it interesting when we were talking before kind of the evolution of old yellow pages and and where you are now and where you fit into everything we’ve just been talking about for the last forty minutes.

Yeah. Absolutely. So Yext is all about your brand’s visibility or really all the facts about your business. And, originally, it was really just trying to fix, like, all the broken phone numbers on the web.

Meaning, like, if you moved off if you moved from one place to another I’ll give you a funny anecdote. When I worked on Wall Street, my wife could never explain what I did for a living. I was in economic and financial research independent equity analysis. And she she didn’t she didn’t she didn’t we didn’t talk about my work much.

Then I switched into this world, and she said, do you ever drive somewhere? Like, she was talking to a bunch of people at, like, a Christmas party. Said, do you ever drive somewhere and it’s closed or, like, they they closed early, but it said they weren’t or they the buildings moved. They don’t have any in there.

That ever happened? And everyone’s like, of course. Yeah. You know, this is fifteen years ago.

And she goes, yeah. That’s his fault. Right? Meaning, like, my job is to was to fix all that data in as many places as possible.

As it turns out, that also was a huge benefit to Google because Google, just like AI, multiple signals to understand, is this business at that location? Are they open? What do they offer? What do they?

It uses, again, probabilities. And the easiest way to think about it is what Yext does is if I walk into a new town, if I come visit you and I say, hey. Where should I get a slice of pizza? And you say you should go to Goodfellas Pizzeria.

Here’s their phone number, here’s their hours, you should order the the meatball parm, and then I asked everybody that you work with, fifty other people, and they all said the exact same thing, I will be really confident my like, to go to that business, that’s what the search engines in AI are doing. They’re looking at what’s the knowledge, how consistent is it, when was it last updated, and does everybody agree this is the right answer? That’s what Yext does, and we do it at scale. So the largest businesses in the world use us to fix that data on their own platform.

So if you give me data at Yext, I’m synchronizing to Google, OpenAI, Bing, Apple, Facebook, and even these older directories like MapQuest, Yellow Pages, people don’t necessarily use them as much, but the that’s not the point. AI and search use them as a signal to process, to see the signal strength that your business is there and it’s accurate. So it’s very much a data business, but it’s really very focused on that visibility in all the models and in classic search.

Well, again, what’s old is new again.

What you just described to me is classic MDM. Right? And I and I know it’s a little more sophisticated and I know I know that you’re dealing with probably millions of sources of data and not just a handful of sources and you’re dealing with real time connections into these other downstream consuming systems. But what you just described is creating a reference dataset based on some consistent rules, some explainable rules that that are consistent over time and are trustworthy over time. That to me sounds an awful lot like master data management.

It it it’s so funny because and I it’s like, again, you and I have had similar career lives. Yeah. You know, everything we did on Wall Street twenty five years ago, fifty years ago, moving to real time quotes on the Quotron, right, or or Bloomberg getting real time soccer scores. Like, everything has always been about the more accurate, the more consistent, the faster.

Like, no data asset that is not real time beats when it becomes real time. Anything that that should be real time will be real time. And so for businesses, this is a new way to think if you weren’t on Wall Street. There, we got into high frequency trade.

Like, we knew you had to match the data that way, or you wouldn’t you wouldn’t get an alpha. Here, businesses, it usually takes fifteen years for what’s normal on on Wall Street in terms of data practices to make it into a retail chain of five hundred stores, but that’s where we are. And the the kind of the fun part is AI is even making that easy. So, literally, we have systems now where if I ran, let’s say, a Panera location, and and I I could literally now say, you know, hey, x y z AI.

Just talk to my phone. We’re out of the the broccoli cheddar soup. Update the menu. We don’t have it.

That updates Yext as a knowledge. We synchronize it everywhere. Instantly, all of those systems now know that location’s out of broccoli cheddar. Don’t go there if that’s what you’re looking for.

That’s the experience we’re going for. Everyone has to see that as the north star of what AI is gonna do.

Well, that’s so that’s interesting. And I think I can’t speak on behalf of large third party data providers, but the idea of what you just said is crowdsourcing for and and I know that may be a bit of a dirty word, but I don’t think so.

No. No. No. Sorry. That’s that’s not what I’m saying. So I’m I’m I’m pretty anti I’m saying the store manager at Panera.

Okay. Not just the consumer.

Yeah. Yeah. No. Yeah.

No. Consumer crowd said it. Maybe fifteen people said it.

Yeah. Yeah. So the problem with that is is very quickly, we learned that your competitor just constantly logs in and goes, hey. They’re out of this. Hey. They’re out of that. It happens all the time.

So it’s it’s a really dirty signal. It’s like using reviews for signals. And with AI making all the fake reviews in the world now, that’s a bad like, you really have to be careful. So YET’s whole model is that the business controls the data.

So if the business updates it but we don’t crowdsource. We don’t scrape and then change it. We’re constantly if we see that what’s really kinda fun is Google and other platforms will literally send us a notification saying, hey. Listen.

A lot of consumers are saying that this store is actually closed until eleven AM every day, but you guys are telling us it’s ten AM when it opens. Dude, we’ve actually had scenarios where corporations learn that the manager at a certain location doesn’t like getting her up early and literally opens the store late. But they, on their books, think it opens at ten because that’s a company wide mandate. They fired them, now they open at ten. But that data comes back to us where the crowd helps us check the data, but they don’t control the data.

Okay. So what you described then is, you know, still the idea of of trust is not equally distributed. Right? That there that there are some people or sources or systems or APIs or something that are trusted above others.

Again, getting right back to MDM. Right? And getting back to the idea of identity because as as much as we’re good talking kind of about context, to me, truth still is imperative here and truth lives at the record level. Right?

It doesn’t live at the segment level or the demographic level or even the object level. It lives at the record level and that’s not gonna change. So how we determine that whether it’s through a third party like a Yext or whether that is through, oh, who knows, maybe blockchain one day.

Yeah.

Sounds like Actually, and and What do you think about blockchain?

I feel like the crypto and blockchain thing has maybe calmed down a little bit. It was it was it was frothy there for a while, but I would say it’s still it’s still a brilliant way to think about how to potentially manage these things because I think there can be value. But a lot of the way we view this and and where I think the AI comes in is, look. We may not even have the same level of connections that we had before in terms of the data systems we use.

So Gemini, for example, Gemini does not use the same data sources that OpenAI does. They have a base to train on, but when they ground their answers, they’re all coming up with different solutions. For example, I’m sure you’re familiar with MCP where, Anthropic made it an open standard. A lot of people adopted it.

It’s very good. But even MCP is sort of a thinly veiled dialogue on top of an API. Right? It’s it’s not necessarily the same Generally.

Yes. Generally. Yes.

Because because APIs are still how changing now.

Yeah.

Yeah.

No. I was gonna say APIs are still how most companies communicate with the external world. So you just slap an MCP over top of a bunch of a bunch of, you know, REST, and you’re good to go.

Yes. And and so to me, when they’re doing that, that doesn’t necessarily guarantee that OpenAI is going to use that layer or that Claude continues to use that layer. And we’ve already seen this. We did a study in October on seven million citation sources on, like, a million questions about businesses.

So, like, you know, where should I get coffee in this region? We analyzed it all. Eighty something percent of all the data that’s powering these AI systems were business questions, not like, you know, why is the sky blue? They’re, like, actual business questions.

The eighty percent of it was the websites of the businesses was forty percent, forty five percent, and the the other were the third party citation sites. So things like Google My Business, Bing Yep. MapQuest. That answers that’s a ton of the the whole game get getting that the basics right there. So it’s not I I think as you as you’re pointing out, MDM was always the right construct.

It’s just that AI is sort of making it far more possible than to use it than than it was in the past.

Or scale I think scale. I mean, it’s to me, it’s really about scale because what you described is old school earlier, which is, you know, waiting for somebody to update a record in a CRM or waiting for somebody to update a record in an MDM hub that is integrated to the CRM.

That’s not fast enough. That’s not customer centric enough. It’s not good enough for marketing that’s trying to do something in real time transaction like right now.

AI is going to deliver scale and speed. Right? And we’ll still be doing some of the same things that we’ve always done under the covers in terms of managing it for identity, creating reference data sets, validating verification, all that stuff. It’s just gonna happen a lot faster and in a ways that are far far far more adaptive, shall we say, or maybe context driven.

So with that, Christian Ward, thank you. I I I could literally talk to you for hours about this We could go in the in the in the way back machine, back to the, you know, oldie early early Internet days talking about all that stuff. But thanks for spending an hour with us today. Really appreciate it.

Absolutely. Thank you, Malcolm.