The Three Ws
As standalone entities, each of the key emerging targeting methodologies, demographic (which tells you WHO consumers are), behavioral (which tells you WHAT they like) and contextual (which tells you WHERE they are) each have signal strengths and weaknesses. The challenge, as Tim Vanderhook, CEO of Specific Media, outlines below, is to fit each of these three Ws together, making the targeting whole greater than the sum of its parts.
Behavioral Insider: What do you think are the biggest misconceptions advertisers and networks have regarding behavioral targeting and its role?
Tim Vanderhook: The biggest misconception about behavioral targeting is that so many networks seem to believe it's an end-all or be-all. The thing is that behavioral without a strong focus on demographics is very limited. It is the best indication you can have that someone is in market for something, but it really doesn't tell which ad -- much less which creative -- to serve. If all you have to go on is that someone's browsing auto sites, do you serve them an ad for a Toyota Prius or a Ford F-150?
BI: What role does behavioral data play in your network?
Vanderhook: What we've always focused on is the importance of targeting as whole. The difference with us, I think, is we look at behavioral as a feature rather than as something that stands by itself. At the core of what we do is a demographic predictive system. We know who every one of our users are, based on cookie data. So if a client comes to us and says, I want to reach an audience of 25- to 44-year-old businessmen -- can you help us find them? -- our goal has always been to be able to tell them we can deliver only those specific target customers.
BI: So does that mean you see behavioral as over-hyped?
Vanderhook: By no means do I mean to dismiss the importance of behavioral. Behavioral data adds tremendous value to demographics. What we attempt to do always is answer three questions: Who is the user, what kinds of products are they interested in, and where are they? If you can answer all three you have a 360-degree view.
BI: What benchmarks do you use to judge how well you're targeting?
Vanderhook: As a network that focused particularly on brand advertisers, our goal is to make effectiveness tracking as transparent as possible. The metrics advertisers want to deploy are up to them. But we enlist Comscore to track our user base and the effectiveness of our ad reach. As a network primarily geared to brand advertisers, the most useful test for us in gauging the effectiveness of a display campaign is in looking at whether we can increase -- and by how much -- the clicks on branded search terms.
BI: Are you extending your targeting efforts into video?
Vanderhook: We expect to roll that out shortly. There are unique challenges to creatively serving ads for video so that users will accept them as adding to their experience. But the fundamentals are the same. The profile of who, what and where I mentioned remain just as relevant. If someone's been searching for "Home mortgage rates" or "F-150's," you know the category of the product you want to advertise. If you know their demographics as well, you know a lot more about which particular price points and products to serve. That profile doesn't change when someone downloads videos.
BI: How about mobile?
Vanderhook: I'm a pessimist about mobile in the near-term -- at least in the U.S. I'm not sensing a market for the mobile Web here yet, at least not in scale.
BI: How would you gauge the current state of the learning curve as far as targeting efficiency?
Vanderhook: The learning curve is accelerating. But what you need to remember is what a dramatic shift traditional media buyers need to make in their thinking to adjust to the new possibilities of targeting. For traditional media buyers trying to reach the right buyers, there were three steps. First they'd research which content attracts the type of customers they want to attract. Next they'd buy ads with them -- and finally, they'd try to figure out if the ads performed.
In the new paradigm there are still three steps, but they're very different. Now it goes more like this. You (the advertiser) tell us what customers you're trying to reach. Then we identify them -- and finally, we report on the content.
So it's a big shift to learn to target users first, content later. It takes time to fully adapt to the new world, but the improvements in efficiency are compelling. Say your goal is to target women 25 to 44. The first instinct of a traditionally trained buyer is to say 'OK, buy iVillage.' But if iVillage skews 58% female and 42% male, then you're still wasting four out of 10 impressions. You put that approach side by side with a new targeting model where 100% of your impressions are going to be served to women, and the necessity to transform how you look at planning becomes overwhelming. I see more and more buyers at that point now.