Letting The Data Do The Talking

The mantra of behavioral targeting has been getting the right message to the right person at the right time. Though much remains to be done, clearly some progress has been made on matching the right messaging to the right people. The next frontier, however, as Ali Mirian, product manager of publisher solutions for 24/7 Real Media explains below, is making it happen at the right time.

Behavioral Insider: What, in your experience, are the most misunderstood things about behavioral targeting?

Ali Mirian: Very often what media planners intuit and assume about what they want to be targeting behaviorally turns out to be wrong. The best behavioral technology in the world is not going to help you much if you're not constantly updating and challenging the assumptions you start off with about which consumers and behaviors are actually the ones that are the most relevant to your targeting efforts.

To do that. you need to be constantly letting the data do the talking--and learn how to better listen to the data you're getting.



Too many marketers still take the approach of starting with a fixed profile and segmentation taxonomy of their customers and their behavior and just fitting the data into that preconceived scheme. That means they're not really cultivating intelligence. That's probably the biggest misconception clients come in with, and it's a gap of strategic understanding as much as technology.

BI: What are some of the unique aspects of the ways you are deploying behavioral methods versus competitors with Open AdStream?

Mirian: Two big gaps in the deployment of BT are, or have been, technological in nature. Those are a lack of integration and a lack of currency of data. That's the area we've been homing in on over the past six to 12 months.

In the conventional world of behavioral technology as most networks and publishers still practice it, analytics and ad serving are siloed into two separate components. You essentially come up with your analysis of where your target consumers have been going and what they've been doing online, and then, based on that information, you decide, OK, this is what we'll tell out ad server to do. What you're saying is, we know Consumer X and Consumer Y who have bought our products or visited our site in the past also have a tendency of doing x, y and z online. So we'll target Consumers X and Y. That's fine--but by the time you're ready to target consumers, days may have gone by. The behavior you're basing your targeting on is not really current.

BI: How will this approach change the way behavioral campaigns operate?

Mirian: As consumers increasingly compress their buying cycles, it's imperative that media buyers and planners learn to compress their media-planning time frames. Online started this process several years ago by demanding that planners move away from the time horizons they were used to from TV and print--where planning and execution were months apart--to much tighter loops. To be done right, behavioral targeting pushes that process dramatically forward.

To do behavioral targeting optimally, you need to narrow that gap from days to minutes-ideally, seconds. You need, in other words, to be incorporating a time dimension, a model of recency and frequency of interactions, into the way you identify and segment customers. So, say you've identified auto site browsers, whom you can segment as auto purchase intenders. The next level--and one we're ready to take--is, figure out how often and when they've shopped, searched or browsed for specific autos, Was it yesterday, today or last week?

This is the dimension we call FIT, frequency of interaction over time.

BI: What practical advantages does the emphasis on recency bring?

Mirian: We'll have case studies soon, but I can say that our testing and research have shown a very strong correlation between how quickly you can move from identifying a behavior to delivering the right message and how effective the ad will be. The closer to same-session, real-time, actionable data you can get, the more successful you'll be.

BI: What are the most important innovations you see on the horizon for 2007?

Mirian: Another level of integration that has remained undeveloped is connecting data collected from various modes of targeting, behavioral, demographic, day parting, geo, etc. All have a place in the mix. But so far they've been used mostly in isolation. So that's another silo we're working this year to break down. Consumers are multidimensional, so it's time to move targeting strategies out of the one-dimensional boxes we've had them in.

Next story loading loading..