Q+A: GroupM’s Brian Lesser
A candid conversation about the future of advertising.
In 2011, Lesser started a company within GroupM called Xaxis, which processes all of that data and figures out how best to target individuals with the right ads at the right time. In five years, he built it to a $1.2 billion annual business. Late last year, WPP — GroupM’s parent and the world’s largest advertising holding company — promoted Lesser to his current job. Today, he oversees six media agencies and specialist companies within GroupM that handle $27 billion in annual ad spending for its clients, or 25 percent of the North American market. (Globally, GroupM controls $103 billion in annual ad spending and accounts for one in three ads worldwide.) Lesser spoke to Traffic in July from a surprisingly nondescript office in GroupM’s to-be-vacated midtown Manhattan headquarters — in two years they’ll be moving downtown, as the anchor tenant of 3 World Trade Center.
Jesse Oxfeld: You’re six months into your new job. How’s it going?
Brian Lesser: I haven’t broken anything yet. I’ve been with WPP for 10 years, so GroupM is not new to me. To be honest with you, I never would have guessed that my path as a product manager for an ad-tech company would evolve into running GroupM North America. But I think that it’s a statement of where the media business is going and where the agency business needs to go.
GroupM in North America is six media agencies and another six specialist companies plus the GroupM entity itself. The four big global media agencies I manage in North America are Mindshare, MediaCom, Maxus, and MEC. Then we have Essence, a digital-only company we bought in the fourth quarter of 2015 that has Google as its largest client, and Metavision Media, which is a three-year-old agency we built for Nestle, servicing all of the Nestle brands in North America.
And then we have the half-dozen specialist companies. Xaxis is the biggest one of those — and of course a company I know well and love. Most people think about GroupM as a holding company that sits above these media agencies and specialist companies, and that’s accurate. Over time, it has to evolve into a platform company that sits beneath the agencies, that provides services that are largely dependent on data and technology.
And you’re a data guy.
The reason I’m now in this role is that my background is in building platforms to facilitate that ongoing conversation with consumers through the use of data and technology. I built Xaxis, which is now the largest programmatic audience platform in the world. And that relies on systems — there are people involved, too, very important people — to reach and engage with consumers at scale, not just across channels and devices, but across countries and markets and regions. The agency business needs to be more reliant on data and technology. It needs to be more automated, more platform-based.
How would you define a “platform” as you’re using the word?
A platform is several applications joined together into one system that allows us to gather data on consumers, amplify that data so that we understand more about them, and unify the data against our own version of the truth — meaning that every time we see a consumer, whether that’s on Facebook, Google, or Snapchat, etcetera, we know it’s the same person — and then activate that data through paid media channels.
So the goal is to be able to identify me as an individual and talk to me wherever I am?
Yes. But from a paid advertising standpoint — and this is an important point — I don’t know it’s you. It’s an anonymous you. But for our clients, they do know you. It’s a first-party relationship versus a second-party relationship versus a third-party relationship. Take you and American Express, for example. If you have an American Express card, then you have a first-party relationship, meaning they send you emails and they understand what you buy. A third-party relationship is a catalog business that sends mail to your house and knows that you like J. Crew versus Banana Republic. Our relationship is second-party, in the sense that we can join up what we know about you from our clients with what we know about you from the rest of our data sources, and we can model that so that we can be smarter about who you are. There’s a lot of data science, and there are a lot of systems, but the objective is to know every consumer in North America and to use that information to help our clients have more effective, more engaging conversations with those consumers.
How close are you to knowing me that well?
We’re very close. GroupM sees almost every consumer in North America, anonymously. And we have literally thousands of data points for most of those consumers. Think of a giant Excel spreadsheet, where the rows are consumers and the columns are things that we know about them. We have that massive spreadsheet within our data-management platform, but it has holes in it.
We might know that you’re shopping for a mobile phone, or that you’re shopping for a vacation, or that you like red ads or blue ads, or that you like certain types of content or you’re on certain websites. What amplifies that is that we work with our advertisers to fill in that spreadsheet with the information they have. We work with third parties to fill that in with the information that they might have on what movies you go to or what car you have in your driveway. And within that, as we fill out that spreadsheet, we can then model the data so that we can look at people who look like you and make some assumptions based on them about other things that we don’t know about you.
We then take that information, and we put it into an activation platform that allows us to buy media against that data. The easiest for us are the programmatic platforms, online display, online video, Facebook advertising. Where it gets a little bit more complicated is when we try to execute that in broadcast channels, but we’ve made significant strides where we can actually buy television inventory off of that data as well, if it’s through a streaming-TV device or an otherwise connected television.
Part of the shift to programmatic that you’re talking about — from publishers’ point of view, making buying more efficient also means making it cheaper. They’re getting lower ad rates.
All “programmatic” means is automating how an ad is served to a consumer and how we make that ad relevant using data. It doesn’t necessarily mean a publisher has to put all of its inventory in ad exchanges. We’d rather find great content and great publishers, and then work with those publishers to make sure that the pipes are automatic — which means programmatic. But we want to have relationships with those publishers where we offer them fair prices for their audiences.
But display rates and revenues are falling.
Display advertising is now the slowest-growing part of the digital advertising ecosystem, because display advertising was always abused — used too much by publishers and not welcomed by users. Display advertising is still growing, but it’s growing at a much slower rate than video and a slower rate than native. And as a result of that, you’re going to see publishers start to take display ads off their pages and replace them with more engaging advertising.
The use of ad blockers is growing — doubling every year. What happens as ad spend shifts from TV to digital, but no one actually wants to see digital ads?
Consumers fundamentally understand that content has to be paid for. So they’re either going to pay for it, meaning paywalls, or they’re going to have to see advertising. I think ad blockers come not as a result of people not wanting to see advertising, but rather as a result of them not wanting to see bad advertising — ads that that are interruptive and not relevant.
People are annoyed by display ads, particularly in mobile environments. However, people don’t want paywalls. And there still is a desire for good journalism, for good content. So these things have to be reconciled over time. And I think the way we reconcile them is to weave advertising more elegantly into content and make it more relevant to consumers. We think that good advertising that is relevant and delivered in an engaging way ceases to be advertising and instead becomes content.
I think we’re going to get there. We’ve spent the last 10 years using data and technology to figure out how to serve you a relevant message, but we’ve spent almost none of that time figuring out the composition of the message. And so I think you’re going to see a lot of investment in what’s called dynamic creative optimization, which is to say, “How do we assemble an ad in real time based on the information we have so that it’s engaging?” If we can do that, we will see consumers open back up to advertising and to content.
So how does this data go back and inform the creative, help make creative that consumers actually want to see?
I think it’s very early, and we have a lot more work to do to get there. But I don’t meet a lot of creative people who are immune or allergic to data. They want more data. They want to understand how to use it — not just to generate insights, but also to create better advertising. I think we need to take all of the work that we’ve done in developing messaging frameworks and models around what the ads deliver, and put that energy behind the assembly of the ad in real-time.
Talking about native — that’s been growing exponentially. Publishers see native as what’s going to solve everything for their businesses. Is it?
No. This concept has been around since advertising began — it used to be called advertorial. If it’s done well, consumers really appreciate it. If it’s done poorly, they just look at it as another irrelevant ad that’s masquerading as content. I think publishers have to be aggressive, but also careful that they don’t run into the same problem they’ve had with other forms of digital advertising — whether that be display advertising or email or other things — where they’re just overwhelming consumers and therefore turning them off.
How does all of this impact where GroupM will go over the next several years?
This all impacts the people we hire and how we train them. It impacts the way we work with our clients, both in terms of the services we provide and the models we use to engage our advertisers. It changes the way we think about investment and acquisitions. And it changes our partners.
We have in GroupM North America 5,000 incredibly talented people who work on client business every day. We already have thousands of people who are engineering majors and mathematics majors and statistics majors, but we need more of them in the future, because this business is increasingly data-, technology-, and math-driven.
The second thing is the services we offer and how we work with clients. Our agencies are becoming much more focused on how we integrate technologies, how we join data on behalf of our clients. And those services are going to come with different business models. You’re going to see us be more performance-driven and less time- and materials-driven.
There are also the investments we make. You’re going to see us make more investments in data and technology. You’re going to see us make more acquisitions that help propel us into a future where we’re more technology-based.
And then, finally, we need great partners, so you’ll see that how we work with Facebook, Google, Snapchat, the broadcast networks — it’ll evolve over time. It will be less about traditional negotiation and trading, and increasingly about, “How do we integrate systems How do we share data? How do we work together to understand the consumer?”
Do you get the data that you want from Facebook and from Google? What I keep reading is that advertisers don’t.
Yes and no. We get more from Facebook and Google than our competitors do, because we’ve built much of our own technology within GroupM. That’s a huge differentiation between us and our traditional competitors. Our competitors rely on third-party systems for the ingestion of data, for the normalization of data, for the activation of data. And so in many ways we’re in a much better position to get the data we need from our media partners and to leverage that data on behalf of our clients. There is still this notion of the walled garden, where Facebook has one version of the truth, we have a different version of the truth, and sometimes our clients have another version of the truth. We would like that to be consistent. But the nice thing about being in the services business and being in between the media vendor and the client is that we have a view over all of the media vendors, so we can help our clients unify their view of the consumer across all of these platforms.
With all this data and targeting, why do I still see ads that make no sense for me?
That’s a good question. And the other one that drives me nuts is retargeting gone poorly when —
I’ve already bought a product and you keep trying to sell it to me.
That’s exactly right. It gets back to the fact that there are more impressions available in an open ad marketplace than we could possibly fill with relevant advertising. The good news is that we could significantly cut back on the number of impressions and drive an advertiser’s return on investment higher. But many advertisers haven’t solved this connectivity problem. In some cases they’d rather just spray and pray because it’s cheaper to do it that way.
When you say a “connectivity problem,” do you mean it’s a question of connecting the right pieces of data to the right vendors, who are then issuing the right messaging to me?
Yes. I think retargeting has sort of gone off the rails because of that and other issues. Retargeting was so much better than what came before it, because you had this golden piece of information that someone went to a website and was shopping for a new pair of shoes. That was such a significant upgrade over what existed before it that now everybody is investing in it beyond what that can bear in terms of return. And so in many cases an algorithm or an advertiser will say, “Well, I know they put this pair of shoes in their shopping cart and they already bought it, but that’s still a more effective use of that impression than some other random ad I’m going to serve.”
So what’s the next step? As a consumer, what’s the next thing that’s going to wow me — other than not seeing ads for the shoes I’ve already bought?
I think the next evolution of ad targeting is us joining your habits and behaviors across multiple devices. What you’re going to start to see much more of is — you went to a retailer app on your phone and you’re going to start to see advertising that is consistent with that experience on your television, on your PC, on your tablet, in the gym, on a bus kiosk, in-store. We’re getting better and better at joining all of those devices to where we don’t necessarily need to see you in all those environments to know that that’s you.
That’s a little scary, but it makes sense.
But, again, from an advertising target standpoint, we still don’t know it’s you.
OK. But the fact that a bus shelter knows I’m walking past it — even if it doesn’t know who I am — is still a little freaky.
Well, maybe. Like “Minority Report” — I thought that was kind of cool.
It’s what you do for a living.
I geek out on this stuff, obviously. My wife and I are shopping for a new washer and dryer, and I get excited when I’m shopping on that device [gestures toward computer on his desk] and getting ads for it at home and on my mobile device. I’m like, “Wow, somebody’s doing their job.”
That’s welcome information. Right now my lease is up on my car, so I’m trying to figure out what new car to get. Auto manufacturers know my lease is coming up, so the ads I see on my desktop, on my mobile, in the emails I get, are all joining up. It’s information to me — it’s content, it’s useful. This is the experience you should have.
THIS INTERVIEW HAS BEEN EDITED FOR LENGTH AND CLARITY.
We will be publishing our readers’ best feedback and incisive dissent; send your thoughts to [email protected].