Bryan Donovan:

I don’t typically like talking about myself, but in this case I want to share a little bit with you. So starting with a real interest in business, but not knowing where I wanted to go, I got into the investment business for a short while enough to learn how companies are valued, how markets are sized, how to forecast growth in markets, and I was very interested in the research in spite of being attracted to stocks and bonds and how companies are capitalized. And that led me to a fairly long role at IDC, a technology market research firm known for forecasting technology growth by market, market size, market share by competitors.

And that was very intriguing to me. I actually served a lot of the folks who were blowing up the internet bubble on Wall Street with research. So I got a real insight into how equity research and investment banking worked. And that led me to really wanting to understand how to harness the power of data for businesses because data about businesses was interesting, but how does that business use that data to make decisions? Part of my work there was with a smaller part of the company, which is dedicated to consumer technology, and that positioned me for a role later with compete. Compete was a clickstream panel of about 2 million US consumers.

They actually had better data than comScore and net ratings and the analytics that we could do on that was really incredible. So I had a sales strategy role for them really helping to package their data for sales, pre-sales to help bring in clients for subscription revenue and build customization for delivering data to them. And that was my first experience with big data, Compete, innovative grid computing and their CTO later went on to found a number of other companies. This really turned me onto the power of big data and what can be done with that and the business models that could create by taking data that each company couldn’t afford to invest in but then being able to leverage that through a recurring research.

That led me to go to Forrester. I helped Forrester to reinvigorate their consumer technographics business. Forrester was really ahead of the curve in understanding how consumers motivation and attitudes for technology drove their purchase and rate of adoption. And that was very intriguing, I worked with a lot of really impressive companies and I had no idea about the direct marketing space. So at that point when I joined Forrester, I didn’t know about Acxiom or Experian or Epsilon or Harte-Hanks, which was a competitor at the time.

But Acxiom had a shared client with Forrester and HP was very interested in how they could take Forrester’s segmentation data and go direct to market with that because they found that research really helped explain customer buying behavior channel preferences, even down to what creatives appeal to these people. And so I formed a partnership with Acxiom. So Acxiom was able to leverage Forrester’s technographics data to create what is now called Audience Propensities. That was an over 10-year agreement with Forrester that netted a lot of revenue for both companies. So it was really profitable.

It was Forrester’s first data partnership and my work with Acxiom, I was really attracted to Acxiom as a company. I first started working with the legal team, which in my experience in the industry is the best legal team in the industry. Kelly Killing, who’s now in a privacy role was actually my legal counterpart at Acxiom working the deal. And without somebody with her knowledge, that deal wouldn’t have existed, those customers wouldn’t have benefited, Forrester wouldn’t have been able to fund their surveys. Eventually I was given a job opportunity to join Acxiom and my role at Acxiom initially was the individual contributors sign up new data partnerships.

So I helped sign up some of Acxiom’s first partnerships and TV data, email inbox behavioral data, and a few others including online as well, bringing from my past with Compete. I was then later asked to lead the overall data acquisition team. And so that included beyond data that was used for propensity all of Acxiom’s core info based data as well as the cookie pool that we were building at the time, pre-LiveRamp acquisition. And there’s been a really interesting trajectory since then with a LiveRamp acquisition. Managing all of those data partnerships, helping our team was a key to helping launch LiveRamp’s data store.

In fact, we signed the agreements for a lot of their early larger contributors to the data store before LiveRamp built a team to take that over. And we trained that team on how to set up those deals. So really proud of that even though Acxiom didn’t get much credit. And then later to the sale to IPG and the integration of IPG, we were able to take on all of the data sources that IPG licensed and managed that on behalf of the enterprise. And so that’s the role we play today. As our team manages the third party data relationships, big and small data for all of Acxiom’s products, for client resell and for our partners at IPG.

Kyle Hollaway:

It’s awesome, Bryan, and one, it’s an honor that you’re on the show. I’ve always really respected your voice inside the company as just someone who always seeks the truth. So it doesn’t surprise me that you come from a place like IDC or Forrester, that kind of research side. It seems to me like the common thread is that you are looking for facts. And I’m just wondering, have you used that research side of the IDC-Forrester in your role in data acquisition or how do you use that to decide what data to acquire and what’s going to benefit our clients?

Bryan Donovan:

That’s a good question and thanks for the kind words. I think that… So how does that help us with the data that we choose? Acxiom wants broad coverage. We have a lot of clients in different spaces, and so we want to be everything to everybody [inaudible 00:05:51] which is hard for a business in the data space. I think market research will always have a role, and maybe this goes into the future of third party data, but market research had a role before there was cookie-based data. It’s going to have a role after there’s cookie-based data, but the role doesn’t solve all those problems.

And so one of the things that I observed in competing against comScore is that they would use a panel to understand great level of detail and then they would use a census view to understand how to project that. And those two things together are really valuable. And I found that Acxiom kind of had that census view info base and that was what Acxiom had tried to build, but you can’t get really deep insights from that demographic data, lifestyle attitudinal data, sometimes harder to come by. And so the research piece was really used to create seed data for that.

So to understand from a representative population, what are people’s answers, how do they behave and how do you extrapolate that to a larger population to help Acxiom’s clients get insights and help drive better audience selection? And I think one of the themes to my current data is I want to use facts and intelligence to help businesses grow because that helps consumers and that’s a nice virtuous circle of capitalism. [inaudible 00:06:58] We’ve always talked about data being a fuel, but in almost every place I’ve been including Acxiom, data is underutilized. And so that’s what I like to do is find not only good data that helps turn the corner on things, but also to help find ways to get that data to be used.

Kyle Hollaway:

Certainly data such a huge topic, like I said, everybody’s got it. And now certainly with a lot of happening in the industry around GDPR, around even state level legislation and then tangentially the whole cookie deprecation thing, but it feeds into the mindset and how the industry is grappling with the use of data. You’ve had a storied career there of working through a lot of different change and now we’re accelerating that change. How are you dealing with that? What are the key pillars of change that you’re dealing with in the strategy around Acxiom data?

Bryan Donovan:

That’s a good and probably pretty complicated question. Definitely. The privacy landscape is definitely making it more challenging. And so you think about it like going on a hike, it becomes a more technical hike all of a sudden. It’s still a hike, there’s still work, but that work tends to take a little more talent, a little more conscientious decisions about who we’re going to work with, how we’re going to vet the data, what data we’re going to accept. State laws will certainly make certain data go dark, but Acxiom has a lot of data about people that we can extrapolate from. And we talked about using research data as a seed to extrapolate things.

We have lots of data points about people that extrapolate… A lot of that data will continue to exist. Yes, we will see, and I predicted cookies going away many years ago, like a lot of people, I was very wrong. I didn’t understand the resilience of that, but it was almost too good, right? All of this information accrued, but it’s about a device, it’s not a person. We don’t know the person behind that. I like to use an example of Facebook, which I don’t use anymore for a lot of reasons, but Facebook thought they knew a lot about me. There were a lot of things they knew nothing about me because I chose not to put it on Facebook.

And the more controversial your views or your hobbies or your other things, people hold that back, right? There’s a lot of self-censorship. So for these platforms to think that they know me is very wrong, but how do you know me? We know me through all of these different businesses that I interact with, that I allow through value exchange to take my data. A lot of that data could end up with Acxiom. And as a consumer, I’ve decided or maybe I’ve ignored the opportunity to opt out, but as I engage with these businesses, I’m giving off various signals. No one group, whether it’s Facebook or anybody else, can have enough data about their customer.

So I always feel like there’s always going to be an opportunity for third party data. There’s always going to be more you can know about somebody, whether that becomes less economically viable over time, we’ll see. We’ve got a lot more efficient about how we access and process data, but the cost of acquiring that data and dealing with some of the gaps goes up. So I am not great at making predictions honestly, as much as I was in the technology market and forecasting things because some of these things are unknowable. There’s too many dynamics here. And like we saw with cookies, there were a lot of bets on that, and a lot of those bets were wrong.

And a lot of people spend money, maybe they didn’t have to. A lot of people spent good money and now are ahead of the curve and now are smarter about their first party data even though cookies are still there. And I’d rather be in a position as a client where I’m doing that now, and I’m not waiting till they’re gone and I understand how am I going to collect first party data. But we say that the ripple effect of cookies going away is publishers and content owners are going to have to find ways to engage consumers, and maybe there’ll be more logins. Maybe you’ll be more incented to provide your personal data in return for that.

We’ve seen this ad driven model with cookies where it’s not very transparent to consumers. Yes, I know I’m getting free content, I’m getting a free service. I’m getting things that I normally would have to pay for, but that’s being funded. I’m the product now because my data is being used for that. Some people understood that. Other people didn’t understand that, but if they all of a sudden had to pay for that service, they might understand the value of that. And our litmus test when we look at data is, “This is my family’s data. Do I feel comfortable about this? The data in our systems, it could be my data, my wife’s data, my parents’ data.

Am I okay with that? Does that feel right to them? Does that feel right to me as a consumer?” And Acxiom, I believe, as a company has tried to live that way. There’ve been some challenges. I say you don’t have ethics if there’s no cost, right? You can say you have ethics, but it you never cost you anything, if you ever lose anything as a result of that, you probably just talk. So ethics costs, it costs to have a higher value on that. And Acxiom has tried to differentiate the firm from the rest of the market by doing that. And there’s some trade-offs that come with that.

Kyle Hollaway:

Let’s just do a little PSA real quick because we’ve talked about first party data, third party data. Dustin, why don’t you give our listeners a quick definition so that they understand the context of those?

Dustin Raney:

I’m so glad you asked me that question, Kyle. We like to think about first party data as data that a consumer consents or gives to a customer directly. So I always use J.Crew as a [inaudible 00:12:01] reason. If I visit jcrew.com or if I go to a retail store and either information collected from my device on their website or me giving them my name, address, email, whatever. That becomes data that is kind of J.Crew, it becomes their first party data. And typically I’ve consented that to them. Whereas third party data, in our case I like to say it’s really data about humans, about people.

It’s about everything that happens after I let J.Crew, I’m changing houses, maybe moving, I’m getting another job, I’m changing behaviors, lifestyle changes. And that comes from more of a census level view that I’m not necessarily going to give to any one brand, but it just represents more of my personal preferences, behaviors that any brand could use. So that’s going to be my perspective of that. And I would say that it’s important to note that third party cookies and third party data are two different things while third party cookies are definitely half gone, third party data isn’t going anywhere. And maybe that’s a good [inaudible 00:13:06] Kyle, any different view on that response?

Kyle Hollaway:

No, I think generally I would agree with you. I know there’s a little confusion in the general marketplace around zero party and first party. Zero party being stuff that is mine that I’m willfully just handing to you directly as a brand. And then there’s all the data the brand just collects organically as part of doing their business more on the first party side, but collectively, all that ends up being first party.

And so I think that is the key differentiator is what does the brand have the right to because they are actively part of that collection. And then what is third party, meaning it’s coming from another source. Now we can always… Bryan, we always throw a second party in there. You want to comment on second party as part of this?

Bryan Donovan:

I will. And my simple view of third party is that somebody else’s first party data and maybe the nuance of that because it does make a good point. When I buy a property for example, and that gets registered in town, I’m not really interacting. That’s not really interaction. But there’s a lot of companies I do interact with where their terms and conditions say that they will use my data with other parties for marketing purposes. And in most cases, that benefits me. When I did use Facebook, I actually had some pretty good product recommendations that I would’ve never known of.

So to me, that benefited me as a consumer and obviously the people using Facebook marketing looked at my data, looked at my behavior and made the choice that, “Hey, he’s going to serve that ad.” So I think that’s pretty interesting. Second party data, there’s no legal definition of second party data. I always think of it as first party data that’s being generated by a partner, and maybe that’s a simplistic view, but let’s say I’m a brand and I place advertising on Facebook, the exposure data that Facebook may or may not give back to me would be second party.

Actually as a first party, I’m causing the action to cause that data to be generated, but it’s not being collected by me. Legally it’s the same as third party data. So Facebook would still have terms and conditions that allow them to share that data back. Most first party data is to allow the company you’re interacting with to better service you, but they cannot always better service you based on their interactions and it’d be weird for them to ask certain questions that may actually really help them because it may be out of context or might make you uncomfortable. We can help them understand more about Dustin so that they can take what they know about Dustin plus what other parties have collected and shared about Dustin to make a better J.Crew offer to Dustin.

Dustin Raney:

Absolutely. And you mentioned one word I think super important even as it relates to the third party data in its collections, is the value exchange really. Whenever I’m consuming content or I’m receiving some value, think about Facebook and the newsfeed, being able to show images of my family or experiences very quickly and consume the same from others. All the comedy that you get to see and all the video feeds now in Instagram I get lost in. There’s some value there and it was enough for me to actually consent to them to know me. And I think the same holds true from a third party data perspective. You want to touch on that?

Bryan Donovan:

I think that Facebook example is a great one because you get used to that. What if Facebook pulled that all away and said, “You know what? Here’s the price for you for a monthly subscription. We’re not going to do advertising, we’re going to charge you for this service.” What would people pay for that? There’s certainly some value there. Absolutely. And if you’re running a small business, there’s a ton of value there. So yes, that value exchange, and as I said before, it’s hidden. We’ve become the internet economy. We’ve become used to getting stuff for free, but whenever you get stuff for free, you’re the product.

And that doesn’t always feel good unless it’s very explicit and open and they’re like, “Hey Dustin, we’re going to use all this information, but you’re going to get this benefit back from it.” I think the industry as a whole probably could do a better job being clear about that. I think there was a push at one point that kind of fell off where we’re going to use more standard readable language and privacy policies in terms and conditions. There might need to be the legalese, but for a consumer to be able to simply say, “This is the data we’re capturing from you and storing. This is how we’re storing it and this is what we do with it.”

It’s nice to know. I’m probably one of the few people who actually reads privacy policies before I sign up for something. Part of it is curiosity, part of it is this a value exchange that I want to engage in? And I do that before I’ve given them my name, email or other things, but I’m probably a little bit of an outlier on that I would think.

Kyle Hollaway:

Just to drag back then into third party data in particular because you’re really responsible for Acxiom’s third party data strategy and how we’re collecting and partnering with other companies to bring that in to help inform our data products and our solution offerings, things like GDPR, all of the different privacy regulations currently at the state level across the US. I know there’s a little bit of movement within Congress about coming out a committee with maybe a national legislation, but that’s still way far off if ever. How do you see that impacting third party data strategy in particular like at Acxiom?

Bryan Donovan:

I think it’s going to make it more costly. It’s going to make it harder for companies to compete in that space. And whenever regulations go up in history, it shows that the companies that are larger and more established actually benefit because they have the scale to manage those. And I don’t know if that’s great for the industry overall that the big companies in that case benefit, but if you’re a small third party data company and now you’ve got to deal with all these changing state privacy laws, you’re either paying a whole lot to outside counsel or you’ve got to build a pretty big legal privacy team.

We actually advise some of our data partners on that because they’ve been good partners with us. We will have calls with our legal and privacy team and tell them what our positions are and what we think is going to happen with the industry. And we’ve actually influenced partners to change their privacy policies to make it more transparent to consumers or to allow them to do business with us because we’re like, “We can’t do business with you until you fix that.” But it will get harder, there will be more things that go dark, and I think that’ll put more pressure on analytics.

We’ve definitely leveraged a lot of analytics at Acxiom over the years, and that’s a core competency of this business, but I think there’s going to be even more analytics, AI machine learning required that kind of gets into maybe the topic as synthetic data, but I think it’s going to take more analytics, more manufacturing and just more tactical expertise to navigate through that and to understand what data we can and can’t accept and how do we still provide answers that marketers need even though certain data might go dark in certain states.

Dustin Raney:

I do want to put a pin in that synthetic topic for a second, and I want to go back to that. This is real talk about marketing. So typically our listeners are, in some form or fashion, probably wearing a marketing hat or playing some role that relates to marketing. With all the regulation that it’s hitting and honestly, some of the states that legislation starting to look more like GDPR is seemingly heading more to explicit opt-in to use data.

When brands or marketers are deciding who to partner with to maybe augment their first party data with third party data, a company like Acxiom, what are things they should consider as it relates to the collection or privacy on the data side?

Bryan Donovan:

Well, I think with third party data, there’s always a risk consideration and who you do business with is really important. We’ve seen over the years there’s been some bad actors in our industry. They’ll represent one thing, but they’re actually doing something different. Part of our due diligence process when we choose partners just like our clients should do, is to really understand how the data is collected, under what terms and conditions, what use cases are approved, how they do a privacy impact assessment. We do a privacy impact assessment, as you guys know. Every step of the way from the data itself, how it was collected, what data we can take in, how we store that data, what use cases it’s available for.

So a client really should make sure that whoever they’re working with has a lot of rigor with third party data and doesn’t take the attitude that, “Oh, we’re just helping you get access to somebody else’s data without actually taking responsibility for the provenance of that data.”

Kyle Hollaway:

And we’re the controller of that data, right? So we have an obligation to CCPA and others to offer up the ability to actually opt out at Acxiom.

Bryan Donovan:

And I think people might forget because we did it so long ago, we were so ahead of the curve, but Acxiom was the first third party data marketing services company to put up an opt-out for consumers and allow consumers see what data we have on them. I thought that was pretty powerful, but as much as there was a hype, people didn’t take Acxiom up on that offer in most cases. So we give them the option to opt out. You can go dark at Acxiom, you could log on our website, you could see what data we have about you. What you’ll find is that they probably have a lot of interesting stuff about you. Some of it’s probably wrong, some of it may be dated.

There’s nothing that’s probably going to shock you. We’re not talking about sensitive information. The activists like to talk about dossiers and portfolios and things like that kind of make you think of the Soviet Union, but that’s not the type of data we keep on people. We’re interested in helping the marketer understand what are people’s interest, hobbies, lifestyles, life stages, demographics so that they can not waste their marketing dollars, which helps the economy. When you think about the impact this has on GDP and efficiency in the economy, if our government wants to take data out of the economy under the protection of consumers.

And to me I’m hearing this conflation with hacking and stealing of data, with collecting and using data for reasonable purposes. These things are totally different. I think we talked earlier about one of the biggest tax of data was the government office of personnel. That’s our government. All federal employees, all of their data was hacked. That is really scary. Acxiom has never been hacked, thankfully. But if Acxiom ever was, the risk of that data is minimal where they’re going to know to go on hikes and have kids, but it’s very useful for marketers.

Kyle Hollaway:

No, that’s great. So into another section now around AI, because AI is fueled by data. Ultimately all the LLMs and everything that’s going into these AI brains, they’re coming up with things including recommendations on marketing that you’re seeing more and more used within the marketing ecosystem. Talk about your view on third party data, AI, how those are, I don’t want to say converging about how they’re coming into play with each other.

Bryan Donovan:

AI is pretty big, but we keep it limited to our space, which is marketing. How do we use AI to do better marketing with third party data? Right now it’s training these large language models. It’s helping them understand how consumers behave, what they do. So a ton of data is needed for that, but it has to be also good data. You just don’t want to throw bad data at there, data that hasn’t been properly formatted, data that hasn’t been gone through hygiene like you guys do for your identity products. So you want clean data, you want accurate data, and if that data is bias, the algorithms are going to be biased.

We’ve already seen early examples of that. And in fact, I feel like there’s subtle attempts to manipulate AI and feed it data to get it to do certain things. So AI becomes politicized in a sense because now it becomes a tool to tell people what we want to tell them instead of telling them what the data says. And I get a little worried because when things get beyond our ability to comprehend it, it’s hard to do that litmus test. Does this make sense or not? We can compare results to real world and get back to it, but historically we do analytics for… We talked about audience modeling and we use standard regression analytics and you could see what are the…

You could say, “What are the variables that are the most predictive?” For example, AI is not going to tell you that, right? Maybe some can output, “Hey, here are the things that were the heaviest weighting, but it’s all of those connections between the things that’s beyond our comprehension and yet we’re going to be making decisions, marketing, probably not huge decisions early on. If you’re a good CMO, you’re going to put your foot in it. Maybe take some small risks before you bet big money on it, but in other areas, the use of AI definitely concerns me and how they’re using third party data power, that worries me as well.

Kyle Hollaway:

I think it is interesting along with, like I said, some predictive nature or being able to ascertain from these large amounts of third party data down into the LLM, the training of it is then the use of that in things like creative generation. Being able to say, “Oh, here’s all these things I learned about this audience, and what are those signals that kind of define that audience and that are the indicators of maybe some kind of propensity from that audience, and then how to translate that into some kind of creative that’s dynamically generated by AI.”

And so it’s interesting flow through model that was now really cohesive, which it does, like you said, it starts all the way back at how is that data sourced and how accurate is it? How confident are you in that? And then ultimately, how predictive is it and then therefore, does it actually generate images or advertisements that then are compelling and then full circle, right?

Bryan Donovan:

I think that’s really exciting. That is using data to drive creative, and I think that’s a big thing for IPG overall and why part of the reason they purchase Acxiom, right? We know that the easiest data monetization play is directed data towards media, right? Data’s value is in its application. How do you apply the data to create value? You can apply it to market research for strategic planning as one value. You can apply it to media targeting that has a different value. It hasn’t really been applied to creative. And when you look at AI ability to generate that and then maybe really creative people to help, feed, tweak or add the human touch to that, I think it gets very interesting.

I know that we have a broader working community with an IPG of their market insights and research professionals from each of the creative agencies. And they’re very savvy about integrating data into their work, but there’s always been a separation between the creative types and the data types, but that is starting to disappear and you’re starting to see people that really understand that data can help at least direct or scope the creative, and then you can play within that space that the data helps guide you towards. Maybe AI replaces some of those folks, but I would still hope there was room for creative humans in the loop.

Kyle Hollaway:

For sure. Bryan, unfortunately, we are running short on time, however, we do like to ask a fun kind of standard wrap up question. Our question used to be, do you see a dystopic future or a more positive future? Changed it up a little bit, made it a little more personal. If you fed all the data about Bryan Donovan into AI, what are the three words it would produce to describe you?

Bryan Donovan:

Or images? I’m an optimist in general for the prior question, but I like this question better. Three words, I would say researched because I tend not to get involved in anything unless I’ve done a lot of homework and digging and talk to people and read stuff and watch videos. Skeptical. I’m just a natural skeptic. I was in sales for a long time. I’d like to think I was a very ethical salesperson. The sales profession can be a little bit tricky at times, and so we’re sold to a lot… Acxiom spends money on data, and so people want me to spend our money with them, and so I’m a natural skeptic.

I think that’s actually served me well in this role, and I would say I’m an outlier. I’m not easy to pin down. I’ve looked at Acxiom data, different segmentations, and I’m like, “I don’t…” Maybe that’s true for a lot of people, and I’m not going to share this on the podcast, but you think of some of the activities, some of the things that are involved in. There are people that are like, “You do that and you do that.” That’s not the tribe that I think of when I think of this activity and maybe this goes back to when I was in high school. I wasn’t in any one of the cliques, but I could hang out with any of them.

So I like to experience a little bit of everything and not be pigeonholed, but who knows? Maybe I can predict me one day.

Kyle Hollaway:

There you go. Thank you so much, Bryan. Excellent conversation. Super relevant to the market today as we look not only today, but even into the near future as data just continues to be at the core foundation of the marketing ecosystem. And so thanks you for your insights and appreciate you being on the show.

Bryan Donovan:

Thank you very much.

Speaker 1:

Thanks for listening to Real Talk about Real marketing. An Acxiom podcast. Find all our podcasts at acxiom.com/realtalk or your favorite podcast platform. Until next time.