From Data Silos to AI Strategy: Andy Frawley of Data Axle

In this CEO Series conversation, Leonard Murphy sits down with Andy Frawley, CEO of Data Axle, to explore how data quality, identity resolution, and AI are reshaping marketing, insights, and analytics. Drawing on four decades in data-driven businesses, Frawley explains why breaking down data silos and prioritizing accuracy over scale has become even more critical in the age of GenAI.

The discussion spans privacy, synthetic audiences, one-to-one marketing, and how AI is transforming everything from customer activation to internal workflows. A must-listen for insights leaders navigating the shift from data collection to AI-powered decision intelligence.

Transcript

Leonard Murphy: Hello everybody. It's Lenny Murphy with another edition of our CEO series. thanks to you for taking time out of your day to join me and my guest. And today I am excited for the first time myself to meet Andy Frawley, the CEO of Data Axel. I think we were just kind of chatting before we went live and there's some cool stuff. I think you're going to think that as well. But rather than me kind of spoiling it, Andy, why don't you tell us your origin story?

Andrew Frawley: Great. And thank you Lenny for having me on. so I've spent my whole career which is over 40 years now in data analytics mostly for marketing purposes sales and marketing purposes. and I've done it as in managed services companies, I've done it in consulting companies. So, a lot of different sort of modalities. But the sort of core thread is, how do we use data to enable brands to create better connections with their end customers, whether they're businesses, consumers, or some combination of the two.

Leonard Murphy: Okay. So, I've always saw the marketing life cycle as engage, understand, and activate, right? And with data being the connective tissue, obviously as things have progressed through the years from digital standpoint that seamless integration has gotten easier, but still seems like a lot of companies do not leverage that effectively. and it seems like that is primarily because of the silo of data and if I understand what data axle is doing effectively is you're bringing down those silos. Combining data sources to create a unified kind of data graph that makes that connectivity easier. Is that a fair assessment? Yep.

Andrew Frawley: Yeah. No, that's an accurate assessment. And at the core of that is identity. It's like how do we understand identity in a digital world which with cookies kind of being deprecated now has become more complex but also how do we understand the terrestrial world where you live and where you work and all those things and so we build that sort of common key across both worlds. we sort of continually sort of, allow brands to sort of manage and curate that through a bunch of different techniques and then allow them to activate not from a channel standpoint, but from a... Sorry about that. people talk about being multi- channel

Leonard Murphy: No worries.

Andrew Frawley: Which to me means they operate in multiple channels but they're separate to omni channel where if I open an email and then I see an ad the ad has the context of the email so to  He's good.

Leonard Murphy: What about concerns? I'm sure you follow this as well as I did, mean, post 2016 election, Cambridge Analytica, all of that stuff, And the explosion around data privacy concerns, GDPR, C was the CCPA in California. Yeah, and I've seen lots of folks, myself included, back in the day, trying to take a run at creating a consumer centric system to manage their integrated data.

Andrew Frawley: All right.  Yeah.

Leonard Murphy: Nobody's knocked it out. Nobody's ever hit scale. I've often thought that maybe if Meta or Google or somebody had done it, early on that maybe that would have changed things. But it just continues to seem to be almost an afterthought and particularly for consumers right some obviously very concerned about privacy and understand the scale and scope of the data and how that data is utilized in an ethical way by brands. Have you thought about that conundrum right is there if you're amassing this incredible data view of consumers what if sudden something changes and suddenly all the consumers go nope nope I want my data but just perspective on  Yep.

Andrew Frawley: No, it's a great question. I've also seen a number of companies who tried to sort of give the ownership of the data back to the person or the consumer and have some sort of a value exchange there that made sense and to your point the interesting ideas none of them had got to enough scale to really matter.  So at the end of the day we obviously spend a lot of time adhering to and participating in the various standards one of the challenges is that I think there was a time when we thought there'd be a federal standard but that has not happened and so now we have multiple state standards which largely modeled CCPA but all have their own little nuances that we have to manage. And keep track of and then obviously we're very focused on consent so we do not want to provide data to somebody who doesn't want to get messages I think a lot of the research is as people continue to experience the sorts of digital marketing we can do today that it truly now can be one to one I mean that's a phrase that's been floated around for years…

Leonard Murphy: Yeah. Yeah.

Andrew Frawley: But with targeting and gen AI, we actually can create content just for you in real time, in the context of a channel. most consumers actually value that if it's relevant. so, we think the answer is as long as we're helping, brands deliver high value content, offers that you're actually interested in, the consumer actually likes that. and everybody kind of understands the internet's a free thing, right? If you have to start paying for the internet,…

Leonard Murphy: Right. Yeah.

Andrew Frawley: that's not going to be a popular thing.

Leonard Murphy: I've helped my kids understand this isn't necessarily a bad thing, but you're the product, right? you like this app,…

Andrew Frawley:Right. Right.

Leonard Murphy: That's fine. you're not paying for it. It's because you're being monetized. You're the And I think,…

Andrew Frawley: Right. Yep.

Leonard Murphy: Who knows if that changes things at a macro level. it's always a concern, especially from in traditional market research where we have really focused on consent and anonymity and those walls have come down to a great extent. The kind of Chinese firewall between research and marketing is really not there any longer and hasn't been for a very long time because companies data axel have been able to drive such insight from so many different data sources that are I think I was kind of implicit right I mean it's just and it's amazing and then along came AI so Let's talk about data axle from the standpoint. How long have you guys been around? What have you been doing post AI? And what do you think the future looks like as AI continues to just transform everything?

Andrew Frawley: So I mean data axle is actually 50 years old. we've been around for a long time.

Leonard Murphy: I did not know that. Okay. Okay.

Andrew Frawley: And businesses like data axle have for years used advanced analytics to let's say statistics for lack of a better term to predict Lenny's more likely to buy than Andy or…

Leonard Murphy: Yep.

Andrew Frawley: Things like that. And so we have a deep heritage in doing that sort of work. about four or five years ago we started using machine learning to sort of supplement and much more sophisticated algorithms could be built to predict various sunundry sort of things. two years ago we challenged our engineering and data science team to build a new product to link consumer profiles to business profiles with the idea that we really want when most companies still today they either market to you as a consumer or…

Leonard Murphy: Right. Right. Right.

Andrew Frawley: as a business customer but you're one person you may buy IT services but you like to play golf and so we built to sort of build non-intuitive linkages between data sets. we call that product profile fuse today. so you sort of work we've been doing a long time. We're getting more and more sophisticated. I think what's happened in the last couple years is with Gen AI now we're sort of into it's not like a technical tool we use we're changing…

Leonard Murphy: Mhm.  Yeah.

Andrew Frawley: How we run the whole company whether that's in our agency groups we have an agency that does activation them using those tools as I said do really one-to-one marketing Lenny gets a different message than Andy in  real time. but we're also using it just to change how we, do work period. You know, how the workflows in the company run. We process a lot of data for, a lot of different companies and, not just do advanced analytics, but you do data hygiene and, some of the boring things that have to happen behind the scenes to have, clean comprehensive data sets. And now we're sort of stepping into really how do we use what I've used them as productivity tools. So we just ran a pilot last month with our engineering team and the generation tool we use generated w 500,000 lines of code in a month terrifying.

Leonard Murphy: Yeah. Yeah. It's kind of crazy, isn't it? I wrote I did my first GPT about two weeks ago and…

Andrew Frawley: Yeah. Yeah.

Leonard Murphy: I went into man, I'm not a programmer. I don't know how to do this. And I was like, okay, that was easy. Yeah.

Andrew Frawley:So we, want to move, so we shifted our business philosophy and said, here's our product strategy for 12 months, but we actually have to do it in three months because the world's moving that much faster now.

Leonard Murphy: Yep. Yep. Yeah.

Andrew Frawley: And that's a big change, culturally, organizationally, strategically. we've been lucky to add a couple senior executives over the last few months that sort of spent a lot of time not just thinking about from a tech standpoint, but how do you change the way the company runs. so, it's fascinating times, we're working on governance things like that right now because you got to have an AI governance layer, particularly in a business like ours. so it's going to change the pace at which we can do things. It's going to change how we work. it's obviously changed the sort of efficacy of our data solutions. because we can clean data better. We can identify data better. and obviously we can target and activate better.

Leonard Murphy: Are you I would imagine over the past few years that a growing client segment for you has been the AI development companies for training data.

Andrew Frawley:Yeah. Yeah.

Leonard Murphy: Fair enough.

Andrew Frawley: No, I mean, it's one of the things that we talk a lot about is, AI is amazing, but it's only as good as its foundational data. Yeah. garbage out. Yeah. And that's been true my whole career. It's still true arguably at an accelerated rate.

Leonard Murphy: And…more impact the implications of bad data in era of AI I think are far more significant than we were

Andrew Frawley:Yeah, it could just explode. And so, we have always prided ourselves on being the highest quality data set. we may not be the biggest all the time. but we do a lot of work from a verification standpoint. we make 30 million calls a year to validate the businesses are open and where they are located etc. we obviously use some AI to do some of that as so we focus on quality that's sort of our hallmark. We almost always win when people want quality. And increasingly that's the battlefield we want to fight on. more is not better. high quality is Y.

Leonard Murphy: Yeah. So there's, we're seeing all of in the market research space specifically the kind of two big multiple trends, but there's a few that are interesting relevant to this conversation. One is panel companies leveraging their data to start creating synthetic personas, we call it synthetic sample. although it's not a particularly accurate term,…

Andrew Frawley: Yeah. Yep.

Leonard Murphy: And then there's another version of nonhandle companies but building synthetic personas off of client data. and where a year ago that's cool. It's already a commodity. Everybody's done It table stakes now. but you think of a company with the scale and the scale of data assets and the variety of data assets. I would think that you guys would be front and center in building this customized synthetic persona segmentation models. is that a product that you have or are you you

Andrew Frawley: Yeah. No, so we have audience data on all of our data sets. Some of which it's deterministic, some of it's probabilistic. but we've created a whole, Geni audience development engine that has all of our data in it as well as panel data. so we can go and, very rapidly, in a day, create new so we do a lot of work in the insurance industry and they're very interested in social determinance of health kind of lifestyle stuff. That's not something that you can go get on a deterministic basis. but we're able to use the panel data, with our third party data to some come up with those audiences. And we do that both in a syndicated fashion where we build them and they're out in the ecosystem. You can buy them or we do it, for a brand will come and say, look, I'm really trying to target these people. I don't know how to find them. can you at scale, go find, 10 million of these people that look like,…

Leonard Murphy: Yeah. Have you gone as far as …

Andrew Frawley: Active, lifestyle, healthy, young, whatever the sort of characteristics are. And those audiences perform very well. we've had a lot of success with those audiences.

Leonard Murphy: what I'm talking about is it actually replaces the research process. So rather than going to a live human and there's enough data basically to create a digital avatar with a very deep profile of attitudes, values emotions etc etc. I do not think that is a replacement for high value high impact strategic decisions. I definitely think that it is a potential replacement for more lowv value tactical decisions.

Andrew Frawley: Damn it. I think that's right.

Leonard Murphy: Yeah and so people are embracing that. so have you gone so far to create a tool that actually says don't worry about doing an ad test with, whatever. Just test it against our synthetic personas because we've profiled them so deeply and…and understand this depth of audience.

Andrew Frawley: Yeah. So I mean a couple thoughts I guess there. So certainly have we have sort of performative benchmarks of sort of how things work and don't work in different channels etc so that's something we have. Now I will say one of the industry challenges is with all of this data with all the GenAI content how do we measure its effectiveness so we used to do an AB test the green envelope versus the red envelope that doesn't work when you have a million colored envelopes and…

Leonard Murphy: Yeah. Yeah.

Andrew Frawley: So I think there will be some interesting advances over the coming years and how to really think about testing and methodologies in a omni channel media environment. I don't think there's a kind of an obvious answer yet.

Leonard Murphy: Yep.

Andrew Frawley: But it will probably involve sort of synthetically creating customer journeys or media exposure whatever the case may be. so it's interesting times that right? Yeah.

Leonard Murphy: It is the omni channel attribution, I mean, that's been the holy grail. I've been chasing forever. It is interesting what I'm finding is there's players Walmart data ventures which people know that I'm very friendly with but also now Instacart announced the same thing of course done Humvey's been doing this for years 8451 yada right of leveraging their first party data particularly their purchase data but not just I mean they sit on a lot of different data assets …

Andrew Frawley: Yeah, absolutely. Yeah.

Leonard Murphy: And as they get into instore media, right, and media networks, they're get pretty darn close to being able to look at maybe not omni channel attribution yet because they're still somewhat limited to but within their ecosystem, they damn sure know, right? …

Andrew Frawley: Yeah. Yeah.

Leonard Murphy: You heard this ad and you bought, this product. Yeah.

Andrew Frawley: Yeah. Within an ecosystem we can do it. It's when we sort of get to the omni channel world where the reality is the wall gardens provide limited data back around what actually happened at a person level.

Leonard Murphy: Yeah. Yeah.

Andrew Frawley: All the sort of display advertising and that sort of stuff. you get some data but it's somewhat cookie dependent which creates holes. yeah so I do think somebody will come up with a tool which sort of takes what data is there synthetically fills in the gaps and says all right this is a representative media exposure for Andy Frolley probably yeah it won't be deterministic and…

Leonard Murphy: All right. Yeah.

Andrew Frawley: prescriptive but it will be representative but if you're a company that is spending $100 million of media  It's going to take a while for them to trust that

Leonard Murphy: So do you worry about I worry wonder as the value of data becomes more and more important and more than that I think anyone who owns data in the era of AI now recognizes that that is their oil field right.

Andrew Frawley: All right.

Leonard Murphy: So is how they engage in an ecosystem and in commerce with their one core asset that cannot necessarily easily be duplicated.

Andrew Frawley:Right.  Yeah. Heat.

Leonard Murphy: I mean I think of it as the moat, that they stop working potentially with companies like data axle or anybody else that is in the business of connecting data or they price themselves out of the market because they're trying to protect the only at least potentially AI resistant I don't think it's proof asset they have. Is that any rumbles of that? I've seen some publishers say, "Nope, this is ours." Reddit famously did that as well. Any concerns around that that data scarcity may become a challenge because of distorting behavior? right.

Andrew Frawley: Yeah. …

Andrew Frawley: I don't really worry about that too much. the reality is that there's more data out there now than there been before and is how do you curate it and make sure it's, accurate and actionable. I do think there's a trend and this is something that we do in the sort of first party data management side of our business where the first party data and our third party data needs to live inside their virtual private cloud and all the processes need to live there as well. So it used to be somebody would send us data would do something and send it back. Those days are going to go away like everything we have has to live in brand X's VPC never has to leave those virtual walls so to speak and that's something we've enabled over the last year is sort of moved a bunch of both data management analytic capabilities but also our data cleanliness sort of the non-glamorous backend necessary things so they can run in a VPC  right?

Leonard Murphy: Yeah. Yeah. So, I want to be conscious of time because I think you and I could go on and get really geeky getting into the nuances here. so we've kind of touched on it but now as we're moving into this currently a world of data ubiquity and ease of utility obviously that impacts the insights and analytics traditionally the traditional market research industry that fundamentally was a feed. It was collecting data. and that's still obviously important. but there's changes that are happening. what do you see the role of data axle potentially playing in that space specifically? Not just that marketing where we've been, kind of driving revenue, but potentially replacing some aspects of market research.  Yeah.

Andrew Frawley: Right. Yeah, I think the data collection part has to happen and will be there. what's changing radically is sort of the insights and analysis part of it. So, we've just literally in the last couple months exposed all of our data into sort of a, pilotish environment where you can write, natural language questions and it just doesn't give you answers. It gives you analysis. it's very cool. we built it initially for our people to use. So, when they go to visit a client, they can walk in and say, here's six things, about your company you didn't know." and…

Leonard Murphy: Yeah. Yeah.

Andrew Frawley: so there used to be big team and particularly in ad agencies there big teams of people who did that sort of work and created presentations I mean that's all going to be done by the machine I mean they made human augmentation but I think that is going to be transformational to consulting companies advertising agencies market research firms where that sort of business analysis step is going to largely be done by the machine.

Leonard Murphy: And real time and agentic, right? I mean there are companies in the research space very large companies that are fundamentally shifting to basically a subscription model with kind of a Bloomberg terminal approach of always on data that is connected and…

Andrew Frawley: Yeah. Yeah.  Yeah. I mean, yeah.

Leonard Murphy: exploratory to answer specific business issues. It's not, for everything. but we are definitely seeing that transformation happen now. and my understanding from brands, buyers, the CPG, etc., is that's what they want. Yep.

Andrew Frawley: I mean because you think about the billions of dollars that have been spent building business intelligence dashboards reporting it just completely changes that game. I mean you don't have to do that anymore. You don't have to anticipate every question that somebody's going to ask.

Leonard Murphy: Right. Yeah.

Andrew Frawley: So it's very powerful. I think the work there is actually to put the right guard rails around the data so that it's correct. a couple of the tools I've played with are very competent shall we say absolutely what's going to happen in your market next year.

Leonard Murphy: Yes. Yeah. Yeah.

Andrew Frawley: So that tech has evolved rapidly in the last few years and will continue to. Obviously there's a lot of people investing aggressively against it. So I'm excited about that just because I think you can get to sort of nonlinear observations fast. so there'll still be people gathering data.

Leonard Murphy: Yeah. Yeah...

Andrew Frawley: There'll still be two, doing people using analytics to create data. but the ability for, a business person to actually go in and identify insights is radically going to change.

Leonard Murphy: agreed the 100%. we could again go on for a long time about that and the role of consultants will still be there to it's going to change,…

Andrew Frawley: It'll be there, but I've Yeah.

Leonard Murphy: Right? But as a consultant I hope that it is primarily there because I hope people still want to know what the hell do we do, Lenny? I hope that's the piece that doesn't go away. But yeah,…

Andrew Frawley:No. Yeah. No, I think the piece where you actually draw the conclusions will stay, but the piece where you've got a bunch of analysts sort of, combing through data sort of in Excel spreadsheets and creating graphs and charts, that's going to change.

Leonard Murphy: Yeah, 100%. And already has what's the old William Gibson quote? future's already here. is not widely distributed. yeah. So Andy in this series would like to try and get wisdom from CEOs. So if you had one bit of wisdom that you would impart to emerging leaders, what would it be?

Andrew Frawley:There's a few things. But one of my mentors early in my career said always remember all the money in a company comes from the customers. You got to optimize the customers.

Leonard Murphy: That was wisdom.

Andrew Frawley: And that's true whether you're a B2B marketer like I am now or a B toC marketer. and I think companies spend a lot of time optimizing other things productivity standpoint or efficiency standpoint but they don't spend a lot of time saying all right here's a customer or cohort of customers I'm getting x from them now there's probably another y out there I could get that they're spending with a competitor or some alternative

Leonard Murphy: Mhm.

Andrew Frawley: How do I optimize that and that's everything from thinking about marketing to customer service to value ad. and most companies aren't set up to do that. Honestly, they're set up to, I can remember, doing I used to be a consultant back in the day and working with an airline that,…

Andrew Frawley: Tell them, the agent screens if they were on the call for more than three minutes would hang but some, buying a $20,000, international first class ticket and,…

Leonard Murphy: Yeah. Yeah.

Andrew Frawley: The rep would be penalized for that because the call went on too long.

Leonard Murphy: you know that if we are truly in the era of one-to-one marketing and I think we are delivering the right message to the right person at the right time excuse me it's also the right experience right to delight and…

Andrew Frawley: Yeah. Yeah.

Leonard Murphy: Engage and retain the customer and that is changing the old net promoter metric probably is not enough at this point that may be an outcome you want to get to, but I don't think it's enough in this era of ultimate fragmentation and choice to keep people coming back. so I'm really glad you brought that up. Yeah, the customer may not always be right, but they are damn sure always the customer.

Andrew Frawley: Right. Yeah.

Leonard Murphy: Yes. Yeah.

Andrew Frawley: And again, I've done a lot of work and research around the concept of, creating an emotional connection between a brand and a and so if I value, trust is the thing that I emotionally connect with, you're going to speak to me very differently than if I value prestige. and what we found is if you can create that emotional connection between a brand and a consumer, they'll stay longer, they'll even ignore, functional things like price in some cases. And again to do that you need lots of accurate data. you understand…

Leonard Murphy: Right.

Andrew Frawley: Who they are and again geni now we can actually build that real one to one content which up until recently you could do the targeting part of it but it was too expensive to actually build 10,000 versions of the content  Not

Leonard Murphy: Right. Yeah. Yeah. I mean, there's a buddy of mine, Ted Tagalakus, with a platform called kind of a marketing automation platform. shout out to Ted but, you create the profile, you create the subject, and instantly you have, hundreds of different versions of the content.

Andrew Frawley:So they can find me on LinkedIn.

Leonard Murphy: I mean, it's kind of mind-blowing with the efficiency gains. Andy, where can people find you? If you want him to.

Andrew Frawley: Yeah. I'm always happy to be found. I'm an active LinkedIn user. So for Andy Paroly and then dataxel data-com Sorry,…

Leonard Murphy: It's too early for our listeners. this has really been a great conversation. I hope we have a chance to chat again. I think that companies data axle will increasingly be if we use that analogy of the oil, right? you're the refineries and…

Andrew Frawley: Yep. It's a good way to put that.

Leonard Murphy: And that the refineries necessary to Yep. Yep. You got to turn into hugely important.

Andrew Frawley: Good. …

Leonard Murphy: Thanks for the conversation. good luck in the future. Is there anything that you want to add before we sign off?

Andrew Frawley: no, this has been a great conversation. Always fun to talk about these issues with people who sort of have lived them, which clearly you have. so I appreciate the opportunity and happy to come back.

Leonard Murphy: All we will try and make that happen. So, thanks to our producers and our sponsors and most of all to the people who take time to listen to this. So that's it for this edition of the CEO series. Everybody take care. Bye-bye.