Behavioral Insider: Could you describe your approach to doing TV behavioral targeting and how it relates to and/or is different from the types of BT tools becoming common online, which media buyers are used to?
Engler: Since the 1940s the big premise for all market researchers was that we could tell advertisers and other businesses what their customers looked like. That remained the case till very recently, but over, say, the last seven or eight years that relationship changed. Suddenly businesses didn't need us to get information about who their customers were. In fact the challenge became [that] they had more information than they knew how to use effectively. So the equation changed a bit. Our business became more about making companies smarter about the data they were amassing all the time.
Also, as far back as 10 years ago, we began to realize that the future of marketing would be built around those databases companies were already building around their customers and transactions. We saw the cost of storage going down, and lots of continual innovation in how to analyze and mine data. So the role we set out for ourselves was to evolve new ways of reading marketing meaning into that data.
BI: When was TV Behavioral Graphics introduced and what kinds of targeting does it enable?
Engler: What became Behavioral Graphics began with work we did for Mastercard. They had a record of millions of transactions every day, but didn't know how to extract a deeper sense of their customers from it. What we did was glue our own data profiles, based on the most extensive surveys and interviews with American consumers, to Mastercard's own data and see what we could come up with.
The next logical transition was to do the same thing we did with Mastercard's financial transactions, but to use other databases of transactions, and we partnered with Nielsen, which has the richest data on TV viewing habits, relevant of course to all broadcasters and those who want to advertise on them. Our consumer surveys also included TV viewing information along with a wealth of other data, so it was possible to integrate the two databases to find significant relationships, links and overlaps between viewing data and lots of other things such as brand usage, consumption patterns and other relevant behavior. We developed the product in 2000 and 2001 but spent about three years beyond that testing it.
The concept was to help take TV beyond demographics for both broadcasters and advertisers. For instance, advertisers have long focused on 18- to 34-year-old women as a prime demographic group. Well if you really dig further this demographic really consists of 15 to 20 different sub-groups with very different viewing and consumption patterns.
BI: Haven't most advertisers moved their models beyond demographics already?
Engler: Most advertisers know this explicitly nowadays. They start off with pretty sophisticated models of who their customers are. But the way it works is, once they actually start working on a TV campaign,they say what we really want to do here is focus on cluster number 7 in our model of 20 clusters within the 18-34 female demo. But targeting in such a tight manner is just not something most agencies are used to doing yet. So what media buyers will end up doing in practice is still just going out and finding programming with "good numbers" in the broad 18 to 34 demographic. Which is to say that once they start translating a marketing strategy into actual advertising, they end up cutting off the arms and legs of that great customer segmentation model they started with.
BI: How extensively is the system used right now in broadcasting, and how does it challenge or promise to change the way broadcasters approach their programming?
Engler: Early adopters include TV programmers who are trying to figure out why and how a show's ratings fell among a certain demographic. They can find out whether the defection was within the core cluster inside that demographic or from a secondary group.
Some TV networks are using the data to work with advertisers to apply the cluster model to their campaigns but so far it's happening directly between the sales staff and advertisers more than agencies. Some networks are already starting to use the system in upfront pitches to advertisers.
The next big step beyond that is to tie everything we're doing with TV behavior to movie viewing data, online behavior, in fact every form of media transactional information available.
BI: There's been much speculation about how new platforms for TV, including VOD, interactive TV and DVR, will change the paradigm of how TV is consumed. Do you envision TV Behavioral Graphics as being or becoming more applicable to these models?
Engler: We think from a media standpoint that we're dealing with the proliferation of a myriad of new platforms to distribute content. Things like VOD and DVT and Slingboxes will radically change the way many consumers get their entertainment and information. As it does it will open up the repository of behaviors that can be tracked, and continue to extend the boundaries of what targeting can mean--and take all advertisers, in whatever medium they're working in, way beyond traditional demographics.