AdExchanger: You mentioned that, historically, Experian Marketing Services built assembled audiences for CRM or database marketing and now you’re working primarily with digital audiences for display or mobile. Do those two campaigns ever converge?
RICK ERWIN: There tends to be a continuum from pure brand-driven advertisers to pure customer data-driven advertisers.
On that continuum, we tend to have more clients toward the data-driven side. We have clients in the middle that do some of both, but what I’ve noticed in 25 years in this industry is that an advertiser tends to have a DNA makeup around brand vs. CRM. And it’s rare to find the company that’s the perfect blend of all marketing disciplines in one advertiser.
Consumer packaged goods (CPG) is one area where that shift is happening quickly. Have you seen an uptick from that vertical?
Yes, both directly and through partnerships with companies that primarily serve CPG customers.
We have a very vibrant advertiser base. We have a very vibrant partner base. We deal with both and see it as a way to have our services find their way to a broader cross section of the market.
We work with a lot of partners working almost exclusively with brand advertisers.
Who are they?
Mediabrands, for example. Magna Global. And others you would think of in the pure data and analytics space with CPG, we’re working with.
There’ve been a lot of acquisitions recently around companies providing shopper data.
We normally partner with people who are deep into shopper data to provide solutions where our services and their services come together and can perform better than either of us alone. We have some relationships where we work directly with large loyalty databases. That’s not the mainstay of our client business.
What are your considerations when you partner?
It’s usually about the partner needing to offer a service to their end users that uses our services. And they have to have a partner like us.
The simplest example would be any database marketing agency that’s building and hosting customer databases for clients that don’t have their own original source of data. They need that descriptive data to put into their clients’ databases. That’s a very traditional CRM example.
To go over to the display media world, we have partnerships with lots of audience distribution companies. Some are pure-play DSPs (demand-side platforms) or DMPs (data-management platforms), some are exchanges or trading desks. That’s distributing audiences found by our data and sometimes our analytics.
When did you start partnering with the DSP/DMP guys?
Around 2008. We have an interesting route to market there. To contrast our strategy from somebody from the Deep South you hear a lot about, we’re not building one platform everybody needs to come to. We’ve integrated our data, our analytics and our other services with dozens of platforms.
Why is that your strategy?
In my own marketing career, I’ve learned it’s dicey business to try to change the way a customer is buying. If your service makes them change the way they’re buying, it could be hugely successful or a long, tough slog.
We just want the world to be able to use Experian data and analytics and all of our marketing suite services to make better, more effective advertising decisions and campaign buys. To do that, you need to integrate with dozens or even hundreds of entities in that space.
What do you think of the single platform model?
There is a time and a place for platforms, but our strategy in digital media has not been to create one platform that everybody else has to come to. You can see how it’s a viable strategy for somebody, but it’s just not ours.
We’ve seen media adapt data in different ways and places. Direct mail adapted data, telemarketing adapted data, email services adapted data. Digital media will be the same way, and you don’t want to close down access to it and make people go through one tiny chokepoint, you want to do the opposite. You want to broaden the ecosystem access to your data.
How many DSPs or DMPs do you partner with today?
It depends on your definition of them, but well over a dozen. The only thing it says about the future there is that it’s a venture capital and private equity-driven space. So you have to expect lots of business combinations to occur there.
How does that change the way you work with those companies?
It doesn’t change the way we do business with them, but being in a crowded space colors how that industry behaves. It’s definitely a culture of announce everything you’re doing as though it were landing a man on the moon for the first time. And do deals with everyone you can at any cost, everywhere, all the time.
That’s the culture. It doesn’t bother us. We play in a lot of ecosystems and in that one, it’s very frothy, where everyone wants to announce the next deal with the next guy. You haven’t seen us do that because we’ve been to this movie before and we’re pretty workmanlike about how we do our partnerships and make our partner deals.
Experian also provides data for addressable TV advertising. What are you seeing there?
That’s an industry that has been kind of slow to change, but something is definitely happening. First, there’s an exemplar in digital media. People look at the digital media industry and a lot of people wonder why that’s not happening (in television). Maybe the pipes aren’t in place yet, or the buying systems aren’t in place, or programmatic is still a little scary. But people realize these are all different versions of the same marketer goal. The advertisers’ goal, at the end of the day, is the same. They want the result at the best possible cost.
So we see people settling into familiar patterns of executing against addressable audiences that we’ve seen in database marketing, teleservices, email and display. We see those same behaviors starting to take root in television.
Do addressable TV ads only appear on satellite and cable right now? And how much available inventory is addressable?
As a percentage of all available inventory, it’s a very small number. Most of the action is in the two minutes of every hour that’s controlled by the cable company. That’s where you can fully control a buy and an audience. And that alone narrows it down to a small subset.
And the data management pipes that enable the insertion of a targeted ad are being ramped up within each cable company at a different speed and starting point. That also affects where each deploys and who’s farthest along with it.
So it’s a pretty small percentage of all available impressions, but everyone is now on a ramp. Everyone is on a path to get to where they can monetize.
When someone says, “The pipes aren’t ready,” what exactly does that mean?
That’s an infrastructure question, and I’m not an infrastructure guy. But in plain English: Think about that moment when you’re watching your favorite show on linear TV, as it’s being broadcast, and it cuts to an ad. There’s no black screen – unless the black screen was part of the show you were watching. It’s a seamless cut.
That seamless cut is unbelievably hard to accomplish from a technology standpoint. And yet, it’s unbelievably important to consumer engagement with the ad. If you have a technological glitch where the screen goes dark and the ad doesn’t come on, the number of people who leave or change the channel spikes.
So getting that technology right is critical. That technology involves outside companies who’ve invented that technology, coming into an MSO (multiple-system operator) with a very precious technology stack, and integrating that software into the existing stack. That’s a huge process for any company, and requires very careful coordination and planning.
What is Experian’s role in the addressable TV buy?
We sit at the center. They’ve got a subscriber database. Every one of those subscribers has one or more set-top box IDs attributed to it.
A measurement company might ascribe viewership of channels to those set top box IDs. Rather than have those data matches happen one to one, they all funnel through us. That’s where the (data) currency comes into play. So we map that currency to different company IDs.
When an advertiser wants their best customers, who also watch certain programs according to the measurement service, and who also look a certain way, to receive a certain ad, we get an order from the cable company that says, “Serve us up those Experian IDs.”
We send them a file that has the Experian ID matched with their subscriber ID, and they can execute that through whatever software they’re using.
Keep in mind, those IDs are hashed. No one has actually seen those identities married up, which is important from the privacy aspect. All that goes back to the advertiser and their agent, the cable company, is a set of numeric IDs. They don’t mean anything, other than these are set-top box IDs you should fire an ad to.
How do you determine which data set is the best? Epsilon once commissioned a study claiming it has the best data, so how does one actually evaluate that quality?
It’s a very vexing problem for us because it’s very easy to be fooled. And there’s a built-in incentive for these companies to fool clients because it’s expensive to build these data sets. It’s vexing for us because we have not reduced our investment in our data assets. We’ve increased our investment in them, and that makes us unique. You’d have to pore through financial statements to see this.
We have significantly increased our spend. So it’s frustrating to us when somebody makes a claim, but it’s almost like political campaigns where someone can purport something that’s prima facie true, but it’s not really true if you dig underneath and do a fact check.
What sleight of hand is commonly used?
The most common way is by count. Many people are trained to think more is better. My favorite is when somebody says “We have data on more than 375 million Americans.” That’s the equivalent of saying, “We’re charging you twice for people we know don’t exist.”
You’ll see this with smaller data companies: They’re preying off people who want to hear more is better. More isn’t better, it’s just more junk.
Two people get divorced, and now the database has Couple at 123 Main Street. It also has Divorced Husband at 25 Main Street, and Divorced Wife at… So one household became three households in the database.
Why not say “We have data validated as of one week ago” or something?
That could be one way to do it. We’ve chosen to do a costly and time-consuming study where we don’t control the outcome, but we get the truth back from somebody who has no vested interest in making us feel better. We pay them to find our flaws. And we don’t go out to the world and say, “Look at the results of our study.” Those who do that are only touting the good stuff.
Let’s talk about Tuesday’s report from the FTC about the so-called data broker industry. Is this the first time the FTC has called for legislation?
I’m not sure if this is the first time, but it’s not surprising to me. We’ve been working with them and others who’ve looked into the so-called data broker industry. We weren’t named in that report.
We’ve been active supplying them with information because we want the world to know what we do. But in terms of what’s in the report, that’s all stuff we’ve been committed to consumer notice and choice and data accuracy forever and ever. We didn’t find anything in there that was news to us. It’s stuff we’re very familiar with. They also closed the report with a best practices section.
What’d you think about that?
The best practices in the report are things we’ve been doing for as long as I’ve been at Experian. Our best practices go well beyond what was in that report and that was what was so interesting to me. What we would have independently defined as our list of best practices is much longer than that, inclusive of those things.
Made me feel good because that must not be true universally or the FTC wouldn’t be publishing a report with best practices in it.
There seems to be a conflation between people finders and marketing data providers.
That’s a challenge with this topic. Our economy is a data-driven economy. When we talk about a data industry, it’s a little artificial because the complexity of the whole world is based on that. (People finders) aren’t marketers. Those are companies providing information to a fee to consumers and companies who look them up online.
What do you think of the label “data broker?”
It’s not descriptive. Any company that I know of depends on data — uses, buys and sometimes sells data in the process of doing business. “Data broker” is confusing to people, especially those who’ve never turned their attention to this industry at all.
And in many other industries, the term “broker” is reserved for people who don’t make a good, simply get it from somebody else, and it maybe passes through their hands. They are affiliates to a transaction.
That doesn’t describe the companies we work with, including ourselves. We originate data. And we manufacture it to an extremely high level of accuracy. We’re custodians of something that we provide to the advertisers who consume it. There are people who traffic in data but that’s not us.
Edith Ramirez described it as a business whose primary business is the traffic and sale of data, without having a direct relationship with the consumer.
That wouldn’t apply to us. We have ten million of our own consumers that we work with. We have our Experian.com credit report business, where consumers interact directly with us. But I don’t know. I can think of a lot of other definitions that seem to fit better than the word “broker.” Tell me what a real estate broker does with real estate, or a stockbroker does with stock. Then come back to data broker, and it doesn’t describe what we do. We’re an originator and a provider, so data provider seems better to me.