Episode 5: Andy Frawley: How A 50-Year-Old Company Retools For AI Success

[MUSIC] Greetings and welcome to 30 minutes with a podcast about business transformation and the people that make it happen. I'm Anders Ekman, founder of Demandbright and Demand Ventures and today we're talking with Andy Frale, CEO of Data Axel, a marketing data and solutions company that's 50 years old but is entering a new era in the age of AI. Andy shares what's changed in customer influence, the buyer journey and how the need for a solid data foundation is more important than ever. Thanks for joining. [MUSIC] And I am pleased to welcome Andy Frale, who's resume and bio goes on and on and on. But Andy is a CEO, an entrepreneur, a board member and CEO today of Data Axel. So Andy, welcome. Thanks for having me, Anders, and looking forward to the chat. Learn a lot from you as we always do. Just to start and set the stage here, help us out with kind of charting your journey because you've seen a lot of chapters in data and technology and how organizations have evolved particularly, era of AI. So give us a rundown of how career has progressed if you would. Yeah, so my career looks a lot like the evolution of data over the last 35 years. I started in marketing services before the internet existed. So it's this frightening statement when we were doing database marketing but it's really driven around mail and telephony. And we had limited amounts of data that were managed in a very structured way. Access to that data was predominantly by people to write code. And then over the years into the data warehousing days where we built more flexible data sets and had tools that could sit on top of them through the beginnings of AI, which something that's been going on actually from 25 years with neural networks as such. And then into the last couple of years or three years where AI is today is a foundational difference. It's as big as the internet was when in the mid 90s when that started. So we're still in early days in my opinion of how AI will be deployed both to improve marketing and to improve organizational form. And data axle has had a bit of that progression too and I know you're celebrating anniversary. So tell us a bit about data axle and how the company has changed where you sit today. Yeah, so data axle is actually 50 years old. It was formally known as info group. And the company's evolved. When I joined a couple of years ago, I would say we were still kind of in that provide data built a model on top of data, activate that data. Whether it be first, second, third party data. And we really have doubled or tripled down on AI where now our approach is different in a couple ways. One is we view the key is marking to a person and part of that understanding the person is to understand their professional profile and their consumer profile and be able to link the two together. And so that's a product that we've launched last year called profile views that using AI sort of builds those nonlinear relationships between data. And second, we really view data as infrastructure for people to be successful. It can't be multiple data sets that are unconnected, have questionable quality. It really needs to be infrastructure in the same way a database is infrastructure. And then the really exciting thing is the ability to start to democratize that data through AI. So the traditional business person asks the animals the question analyst goes and gives back an answer. Business person says that's almost right, but go ask these three questions. That process is going to be improved by an order of magnitude from a speed standpoint. It's really going to change the efficacy of how people think about both audience targeting and activation, but also messaging and creative. And so are you changing the role of frontline account teams given this capability with AI? Are you supporting them with tools that they might not have had otherwise? >> Yeah, so next week we're actually launching a new product that basically using a chat GPT style user interface allows business people to go in and ask questions, do analysis, create a narrative. And we've been beta testing that with clients over the last quarter where our account people are actually sitting down either in person or virtually with a client. And spending an hour to interrogating the data and trying to find audiences that might have a greater propensity to respond, try to understand things like buying groups where we've done a lot of work with our V2B data over the years. And in an hour they can do what might have taken weeks to do before. >> Fantastic. Tell us the role of unstructured data in your world because to me, that's really what AI has been able to. To use your word democratize. Have you been able to connect unstructured data to the more kind of traditional structured data that data axles always had? >> Yeah, in some ways we have and in some ways it's difficult. If you're going to have a tool like the one I just described, the data has to be very well formed, very well structured, very high quality and have a semantic way that describes the data. That is how the AI interrogates it. So there are cases where we can ingest the unstructured data. This certainly we're using AI to do things like convert PDFs in the structured data, convert historical campaign data into structured data. But of course one of the other extreme exciting applications is using structured data to create unstructured data in the form of content. So we focus heavily on quality data axles. You'll find companies that have more data than we have. You won't find companies that have more data that's high quality than what we have. Again, that was just sort of bedded by Forrest of naming us as a leader in the B2B data wave last week with the most AI ready data in the industry. >> Congrats on that. That's great. And you bring up something I'm curious about. We've talked a little bit about B2B and the different verticals that I know data axle operates in. I know you're in nonprofit and certainly in consumer is the AI agenda, the kinds of data, the kinds of activation you have to do. Kind of the themes across those verticals differing, consistent. How does the business sort of break down? >> Yeah, so we have three groups we go to market with. One is our nonprofit group, incredibly. It is empowered by a great second party data set where we've built a co-op of donors and use AI to generate all kinds of predictive indicators of most likely cause, size of donation, things like that. Actually, I have a group where the target user is actually small businesses and we have a platform called sales genie that allows a small business to do a lot of what a larger business can do. And then I put a very packaged way. And so there, AI will usually need to generate emails, generate content, find ideal customer profiles, the kinds that are constant meet ideal customer profiles. And then our enterprise group, I think, where we do both B2B and B2C where in B2C it's really more about understanding all the signals, integrating intent data and to do that sort of micro targeting and then micro messaging, hyper personalization. In B2B, sales and marketing has changed a lot of the last several years, not just from an AI perspective, but just how companies buy in a B2B sales cycle. And one of those changes is the notion that a person buys from a company is sort of obsolete now, it's all about buying groups. We've done a lot of work with our foundational data sets to understand titles and seniority and intent signals and really look at that. Not just as a single sales funnel, but almost a knowledge graph of where the people are in the buying cycle, where you are in your interaction with the different people in the buying cycle. The ability to sort of do send different messaging to the marketing person versus the procurement person versus the analytics person. And that's obviously incredibly complex and really couldn't be done without the answer is AI. Right, one of the things that we hear all the time is particularly B2B complexity, too much data, too many platforms, lots of disintegration and things like that. Are you sort of hearing the same things and are you able to kind of help guide clients toward simplicity or is it still a hard road to hope? Yeah, it's a great point, Anders, I think over the last seven or eight years B2B sales and marketing groups just acquired a ton of different technology. And so you'll see 10, 11, 12 sort of platforms, quote unquote, involved in the marketing and sales funnel from sequencing platforms to CRM platforms themselves, chat, etc. And that's created a problem that we've all seen before, which is we've got a lot of different data sets that don't connect. We have different levels of granularity from an identity perspective in those data sets. And I think in many cases what I'm seeing is enterprises, they're trying to skinny that down and get to from 12 platforms to five platforms. However, there's still, you know, is a need for the data integration. So, you know, again, we spend a lot of our time taking our third party B2B data set, linking it to a consumer data set. So we get that consumer context and then, you know, integrating it. it with first party data and getting it into a data lake of some sort that we can then apply to AI analytics too. Got it. So zoom out for me for a second. I know you talked to different kinds of marketers all day long. How is AI sort of almost outside of your business? How is AI changing the environment at marketers? Are teams acting differently? Does the infrastructure stack have to change? What are the big themes that you're seeing and your clients overall? Yeah, so certainly I look at across two dimensions. One is how you do the work and then what the work product can be. And so AI certainly provides opportunities to do the work differently. Whether that's using it for content creation, using it in a jack environment for workflow management analysis. We encourage our own marketing team. Every person on the team has to have an AI employee that can do some of their work for them and then think through that happens. That's hard because it creates a lot of organizational change and some people will resist that in my experience. But massive efficiencies can come forward. And that needs to be the case because what the work product is, it's really data isn't about data. It's about the ability to improve the decisions you're making and drive the outcomes that the business needs. And because we've been able to do hyper targeting for years now, but now we can actually do the hyper personalization that goes with it so we can deliver very fine, great messaging, content in an on-me channel environment where everything's sort of connected. But that creates new work streams. And if you want to do that sort of work with the same number of people, you need to be more efficient than historically marketing teams have been. Yeah, so you bring up something interesting. It seems to me that the traditional frameworks of the purchase journey, the funnel, call it whatever you will in this era of AI in the context of committee purchasing decisions, things like that. These frameworks seem to be kind of not as relevant as they were. We're in many ways back to a concept that got launched about 20 years ago, but never really got realized I didn't think, which is the segment of one. And try to follow that individual through their journey, whatever that journey is. And it's probably going to be different based on that person. Yeah, absolutely. So the notion that we can lay out an eight-step customer journey that goes from consideration to purchase is really obsolete now. And the reality is brands don't own that journey. The person owns the journey. And everybody's journey is going to be different. How they consume information is going to be different. And then you'll well be the fact that it's not Andy Frawley making a decision, but it's six people. And Andy Frawley making the decision, who are all going to have their own customer journeys, you can't map that out in a flow chart. Right. You have to be able to understand. And again, that's why we think of it more as a knowledge graph infrastructure that can capture all those signals. And then try to put them together into some logical order. And that's been tried for years from campaign management products, customer journey products. And they really just weren't built to handle the level complexity that is our reality today. So it sounds like you guys are playing a role in helping your clients transition to AI. And I guess I've seen, and I'd be curious as to your opinion, a lot of AI initiatives sort of get started, but peer out. They fail somehow. They don't really realize the benefits that the bosses might have. What happens there? What's going on under the hood that sometimes stalls these things out or makes them less effective than they might have been? Yeah. Well, these initiatives have to be connected to the outcomes, otherwise the science project. And those will always peter out. And so our frameworks in terms of helping people think about activation schemes are very focused around using AI to lead to an insight that we can link to an outcome, too. Well, and sometimes that's simple, sometimes it's super complex. And all this still needs to be wrapped in a test and learn framework. So you can understand both the incrementality of the outcomes as well as the cause of the communications. And so, maybe testing is like an idea that's been around forever, but it's still, it changes in this world, obviously, but it's still really important to feel and demonstrate and prove performance. Yeah. And I think that we're sort of entering the post theoretical world of AI where there are a lot of marketers and business people in general that are becoming comfortable with it, understand the various puts and takes. And what that's doing, I think, is driving AI closer to execution. That sounds like it's where you guys are taking it. Yeah. I mean, it needs to be there from insights through execution. Right. And there's different types of AI and different approaches to doing along the way. I think the other thing that is a foundational change has happened over the last couple of years is it relates to marketing specifically is, you know, for years, brands and you and I saw this, we worked together. More data is better than less data. And what we're seeing now is kind of a return to quality because if you don't have high quality, foundational data, AI is just going to do a really good job of making the wrong decisions. You're going to make them a lot faster. And so we're now, you know, we're in the middle of an evaluation cycle right now, the, you know, fortune 50 company, the limited evaluating data for 18 months, it's a B to B data across 10 or 12 different, you know, variables. So we see the return to quality is a good thing because if you build your AI on low quality data, it's not worth. Right. Well, particularly in the era of agentic operations where you're entrusting an agent to make a decision or advance a process or whatever it is, that agent has to be operating on a knowledge base that is tested and true. So you mentioned that your teams are using AI. Maybe that's a genetic. Maybe it isn't. But what role do you see agents playing in both your own work and then that of your clients? Yeah. We use a genetic AI sort of foundationally under the cover of our natural language, the fire off different, you know, models or different approaches. So you go, go ahead and have a conversation. So you know, give me all the born and people with blue eyes and Boston and then build a look like model of those people in New York, things like that. So that's all empowered and can be done by certainly analysts and I think likely increasingly business people. How do you take that, those audiences and the content and then sort of push them out into the channels is where I think the future of a genoc happens and again, it's the concept of multi-channel marketing, but omnichannel marketing where every interaction is informed by the prior interaction. And I think that's where you'll see agents sort of talking to agents and then talking to display a platform or an email platform or telemarketing platform. That's sort of where we see the genoc stuff going and again, getting to that as you said, segment of one approach where we really are trying to speak to people. Not as an audience, but as a person. Well, yeah, and I am starting to see we are that AI is playing or is capable of playing the integration role between this cacophony of data and platforms and things like that. In many ways, if you can design your AI motion correctly and have agents kind of take this inference and turn it into an action, maybe that's getting us toward an integrated world that I think particularly be to be, but across every category, a lot of marketers wish they had and don't quite today. Yeah, no, I think that's right. And with the integration layers where AI can talk to AI and GCP servers and stuff like that, my AI will talk to your AI and then make a decision that they both agree with. So and going on the genetic framework, so obviously what sort of powers all that. Yeah. And a guest on a recent podcast of ours, Guy named Santos, Sharon has designed an application that knits together buyers and sellers in B2B. It's an agentic buyer led application, if you will, or AI capability. So all of that sort of down your path of your AI, talking to my AI, I think that's going to be here within a year or so. The Santos is going to launch his company in a couple of months here and it's ready to go. Yeah, I think it's here now, at least the early days of it and you're going to see a lot of different approaches, but they're not going to be silos anymore. They're going to be able to work together. And from a data actual standpoint, our approach is, you know, we want to be able to deliver the algorithm in the payload into whatever environment needs to go. That's an execution platform or an agentic platform. And we'll be able to surface the uniqueness of our data in a way nobody else can, obviously, because we understand how it's built and all of its strengths. So I think it really starts to change the playing field and level the playing field to some degree. We work with a lot of, you know, Fortune 100 brands, but we also work with a lot of kind of lower enterprise brands. They're going to be able to do lawless as well. It's not just going to be the land of the big companies that can do the sort of AI work. It's going to be democratized. Right. So as you reflect back on 50 years of data axle and the development of the company and data and marketing and all of that, we're going to be able to do that. What do you look or what do you see for the next 50 or the next hundred? What sort of era are we entering and give us your sense of what the road ahead looks like? Yeah, well, I think I'm not sure I'm smart enough Anders to trick out that far. Come on now We're in the first inning of the AI era right now And so it'll go through generations like all other technology and business waves But I think you know, there's definitely a future where you know the robots would be gonna be talking to robots and making decisions We still think it has to be human participation in that so to speak because the AI is not perfect at least now It'll get personally better but from a marketing standpoint the real manifestation is that it's gonna be much Easier to give a highly personalized value prop and probably products in the future and do it at a rapid pace So it's gonna speed everything up. It's gonna make everything much more granular and Yeah, I think the winners are gonna be very profitable should be very cool to watch and it's great to celebrate Data axles 50th and economy plate the road ahead and to hear your views on the future So Andy Froly, thank you for your perspective and your time today really enjoyed it and we will all be watching and Appreciate it again. Yeah. Thank you. Anders is fun. I have a conversation and should revisit it in a year We will do it again. All right. Thanks again Thanks for listening to 30 minutes with to learn more you can visit us at demandbreak.net or Demand Ventures.co. We'll see you next time