In contrast to the online space -- where behavioral data is abundant, but truly meaningful customer data remains a challenge -- mobile operators and publishers have an incredible wealth of highly personalized subscriber data to start with. Their challenge, as Nick Lim, product management director of Enpocket, explains below, is learning how to relate all the rich demographic, usage and other kinds of information at their disposal to transform it into real-time intelligence.
Behavioral Insider: What's different about framing a targeting strategy for mobile versus online?
Nick Lim: There are two things that really distinguish mobile targeting from online. First, the amount of subscriber data that's leveragable is far greater and richer than almost anything you can find online. Second, with portability you have far more history and cumulative data over time, because, unlike cookies, which have a very short shelf life, cell phone numbers often are kept for a long time.
BI: At this stage mobile clearly looks to have an edge in terms of accessibility of rich demographic information. But how can that demographic information be broadened for a wider range of targeting, especially behavioral?
Lim: Our newest mobile ad platform, Enpocket 6.1, is an important extension to what we've done before. In previous releases we've had targeting capability for demographic profile information, basic things like age, gender, and income stored. Over time we've been adding interest level behaviors, but what marks the new release is that advertisers can now drill down into real interest-level classification and do it according to their own needs.
If you look at behavioral targeting as it's normally practiced, you're usually classifying a target customer's interests in terms of a fixed hierarchy. So you'll start with someone who's evidenced interest in sports, and then go to the sub-class football, and then to further sub-classes of interest, the NFL or college, say.
BI: How does this psychographic data relate to behavior?
Lim: In the old days, the goal of targeting was just to match the general profile you had of who your target customers were with sites which might by proxy attract them, based on content. Behavioral targeting for the most part remains based on that kind of one-size-fits-all content-centric view of who your customer is. Once you are able to tie a demographic profile with more detailed psychographic profiles based on more specific interests, there's clearly an improvement in response metrics like click-through. But more important, from an advertiser point of view, you learn a great deal more about who is actually responding. That is, you can learn to track how your ads are performing far better.
Lim: For instance, say you got 150,000 responses running an ad targeting sports site visitors. When you match demographics and behavior against interests, you discover you really got 30% of a female mix in that response. So, by constantly gauging how demographics, interest classifications and behavior relate, you're continually able to learn and improve your assumptions about whom you're targeting, when, where and with what message. You may find, in trying to behaviorally target finance site visitors who go to sports sites, you don't do so well generally, but you do extremely well among Hispanics with a certain interest profile. That's something which would be missed if you were just doing content or behavioral targeting in isolation.
BI: What kinds of benefits will this give advertisers moving into the mobile space?
Lim: The more you're able to really identify what components of your targeting profile are really working best, and which aren't, the more you are able to point advertisers to the specific users most likely to engage with particular messages. So you're learning more over time about not only where you can target them, but what kinds of messages are optimal.
BI: Can you outline how these connections can be modeled?
Lim: We begin with subscriber data as it relates to demographics and psychographic characteristics, and then overlay that information with mobile usage information, what pages they visit and what campaigns they respond to. All the pages a user visits can be tagged not only by content but by very rich psychographic categories. For instance, if a user goes to a sports site we capture patterns about what types of sports they're interested in, and then go deeper into what kinds of areas within that sport. Are they a college or a pro fan? If it's college, do they follow the Ivy League or the Big Ten? What kinds of sports information sources do they use, and how do those relate to who they are more broadly? And, then, finally, how do these profiles relate to behavior over time?
BI: The biggest pick-up in behavioral targeting so far appears to be from direct response advertisers. What benefits do you think this melding of psychographic and behavioral approaches might hold for brands?
Lim: Direct response advertisers will always be the first adopters of new targeting methodologies, if only because they tend to have a more clearly delineated and articulated model of who their specific buyers are. But we are seeing much more interest now in how interests and behavior inform purchase patterns from brands. For instance, CPG brands who sell toothpaste now see numerous sub-markets of different kinds of toothpaste buyers, and are bringing out brands aimed at as many micro-niches as possible. Increasingly brands also want to follow specific market segments over time. That is, they no longer think of just focusing on 18- to 24-year-olds, but look for a specific cohort within that group and look to follow them over time as they age. To do that means to learn to look at how profiles and behavior interact and evolve over time.