• by Steve Smith , Staff Writer @popeyesm, June 20, 2008

In the latest of his regular reports on behavioral marketing, eMarketer senior analyst Dave Hallerman projects a significant spike in behavioral targeting ad spending in 2012. When we asked Dave to walk us through his numbers, video emerged as a key driver for revenue growth. So we took a small detour through Dave's reasoning. BI: How has your view of spending growth in BT changed since your last report on the sector?

Hallerman: Part of it has changed because I am hearing from more sources that what growth there has been for behavioral targeting is pretty much tracking the ad formats. Most of BT is in display in the broadest sense, and that has been a little bit less at this point. Resistance still is there, certainly on the branding side. The largest growth still lies ahead, as more brand advertisers are figuring out how to play the online advertising game.

BI: So where will spending be this year and track over to 2012?

Hallerman: In the U.S. only, online advertising with some behavioral targeted component was $525 million in 2007, will be $775 million in 2008. In 2009, we project $1.1 billion, 2010 is $1.7 billion, 2011 is $2.7 billion, and 2012 is $4.4 billion.

Several factors contribute to that accelerating growth. A bottom-line one will be video advertising becoming mainstream. Video advertising is the most expensive form of online advertising. Marketers need greater effectiveness because of its great expense and expect greater effectiveness. As there is far more video ad inventory, the share will go to BT might even be more than for rich media today.

BI: Do you foresee that the technologies will be in place for BT to be applied against video? Hallerman: That is another factor holding back the growth. It is very complex technologically speaking. There is a lot of data. How do you extrapolate the meaning from it? You have to deal with all that data and find how that is meaningful in terms of which ads to serve. Also, in terms of the spend, are the advertisers getting used to the fact that they need more creative if they are going to target by people and not pages? You need more creative because you can't just keep throwing the same ad at the same person in different places and get the same effectiveness. So the amount of variables involved technologically still are being developed and not there already by any means.

BI: As a percentage of overall spend how will BT tracking over time?

Hallerman: Of total online advertising, it is 3% this year. If we define display in the broadest sense as static banners, rich media and video advertising, then the percentage of display is 9.5%. By 2012, BT will be 8.6% of total online ad spending, but it will be 23.4% of display ad spending.

BI: You are projecting it will be a fairly standard part of the buy.

Hallerman: A quarter is a strong minority share. It is strongly connected with video growth and not just trying to emulate TV's mass blast sort of advertising.

BI: You mentioned earlier that brand markers are still trying to figure out how to use BT. Is it just that many are still trying to figure out BT specifically or their online strategy generally?

Hallerman: They are figuring out their best online strategy. Things like BT are actually part of the attraction. The whole idea with a lot of traditional marketers is we will get more information about the results of our ads and be able to target them better. So that is an attraction in terms of figuring out the broader tactics for brand marketers. Some of it will be advertising but more and more will be various forms of sponsorship and product placement. If done dynamically, it can also be targeted. Even product placement can be done on the fly in certain contexts.

BI: In researching this, you must run into the problems of definition. A lot of buyers complain about defining BT. How do you see that definition changing?

Hallerman: One of the areas of technological growth will be the predictive modeling. That is part of the definition, and there the behavior is indirect rather than direct. It is going to be harder to monetize, but as the social Internet grows all those behaviors are to be mined. Some of them will not be easy or even useful. Not all behavior is going to be useful for targeting ads. But it is going to be looked at. This is one of the reasons why the technology gets complex. As people spend more time online, more behaviors emerge. Yet they are not all useful for targeting advertising. And the whole issue of how it is sourced through ISPs is a pivotal point.

BI: This seems to be a topic that is heating up. The older traditional BT networks are getting hostile about the potential threat of ISP-based BT.

Hallerman: I wonder if ISP data collection will be the sacrificial lamb that will make cookie-driven data collection look more benign. A key thing when it comes to privacy online is simply the mainstream media discussing it. The general public doesn't read my reports or your articles. But when the mainstream media pays attention, it means that the public will. And something like ISP data collection has more potential for blowback from the public than cookies do.

It will also force more transparency on the publishers and on the ad networks. That may be a something with an unintended benefit. My favorite data of all shows that the favorite forms of marketing are emails from companies people have relationships with and they say they want to hear from. And the most hated are emails from companies they don't want to hear from. Same format. What's the pivot point? People were asked, and they said yes. That dynamic operates in all forms of marketing. It's one of the many reasons why paid search does well. People are given a choice.