• by Steve Smith , Staff Writer @popeyesm, March 2, 2007

Veterans of the BT space will recall Omar Tawakol as a fixture at digital ad conferences in the early 2000s. As the CMO for Revenue Science he was a perennial panelist explaining the technology and virtues of behavioral approaches. Now, as chief advertising officer at mobile search firm Medio, Omar is still pursuing the BT model but, "in reverse," he says. Medio powers various types of on-deck search for T-Mobile, Verizon, and Amp'd. The range and depth of data that mobile can deliver about consumers generally may be more valuable to marketers than the basic act of segmenting and ad serving.

Behavioral Insider: What is the link between the BT disciplines you developed at Revenue Science and mobile search?

Omar Tawakol: The thing you learn when doing behavioral is that some of the best behaviors you find on the Internet are behaviors that have signs of intent. Search had proven [itself]; through all the different types of data we worked with at Revenue Science, search was always the best. It was more granular and it was more lean-forward, since the consumer actually types it. Because the mobile phone is a single point of data access there is an incredible amount of richness a carrier can have. It's as if you had access to an entire Internet experience online.

BI: What data points are available to you now on mobile searchers?

Tawakol: It depends upon what data is allowed for you to have access to. If you have a good relationship with the carrier, you may have the ID of the phone. It depends on whether you are working with WAP or an application, where you can track behaviors across sessions.

BI: What is the state of those relationships. Are there carriers sharing?

Tawakol: I can't comment in specifics. I can say Medio has access to that information but I can't talk about which carrier does and doesn't [share].

BI: What kind of segmentation is possible now?

Tawakol: Let me step back and lay out something. What you saw in classic BT online was typically [that] people did segmentation as if they knew the answer to the question. Ford would say they wanted you to give them people who went to the following SUV pages. So you ID someone like Kelly Blue Book or Edmunds and tell them these are the behaviors we are interested in. Go get me people who did that. It is glorified database look-ups. The buyer says, these are the acts I want to buy.

There is another form of segmentation which happened a little bit in the BT space online.. and we have started to take it up in the mobile space. We call it 'clustermerization' or identifying 'clustermers.' You work almost opposite to segmentation. We let the data speak You get access to search history, WAP browse history, transaction history and demographic information. You put all those attributes together and you have a very rich database. But instead of forcing a conclusion on the database, you let the segmentation and data mining techniques come out and identify discrete types of clusters that you can then do all sorts of things for. You can identify trends in the data that are unusual, like people who love 'Star Wars' content actually like classical music, which is something your wouldn't know in advance.

The data has bubbled up naturally occurring audience segments based on behavior. The interesting thing is that now you have carriers who have access to all that [data] and can group people that are similar and tell us what their behaviors are. How do we sell services to one group of people because we know [those who are similar to them] and what they have done.

BI: What sorts of ads and campaigns come from that knowledge?

Tawakol: This biggest trend I see in advertising is that ads that look like content are much more useful to the customer. They treat it as something that's useful. [In mobile search] it is not seen as advertising. The customer asked a question and you answered it. The goal is to use BT to extend that, so ads outside of the search experience are so relevant to the needs of that consumer that they look at it as an embedded piece of content.

BI: It sounds as if you think mobile has data points that can be used to better inform other forms of marketing.

Tawakol: An interesting opportunity I see on the phone -- and probably also online -- is to do behavioral targeting in reverse much more aggressively. I think about BT in the opposite way it's been done for the past four years. Generally, we figure out a group of people to send a campaign to. And it gets to them some of the time. But if you did it in reverse, you could get them all the time. Which is to say, instead of using BT to plan and deliver a campaign, you can use it in the planning process. I will still buy contextually and by search or geographically, but BT allows me product maps of who are the people who are actually responding to my creative. I could fine-tune my buying to really get the right people. Even if the next iteration of the buy is not behavioral in delivery, you are still using it as a tracking mechanism to figure out who is the right audience and who is actually responding to my campaign. It becomes a way of instrumenting all campaigns.

On the mobile phone that becomes really rich, because you get that one point of information that can look at all the history and say, who is the right recipient of this ad. You can go further and say that ads that take place on other media or print can use BT in reverse. It's a huge potential of the phone to let you understand audiences for ads that aren't even delivered by phone. I would call them behavioral maps. It may be that BT in reverse is bigger on mobile than direct BT. Instead of using BT to figure out who to serve the campaign to, it would be used to make all campaigns smarter.