Prime net real estate

As online ads spew out more data, their value rises.

According to Avenue A/Razorfish, a banner on a leading portal, such as Yahoo or MSN, now costs about $500,000 for a day, about the same as a 30-second spot on a hit TV series such as CBS's CSI. What?!?

Some 20 million to 25 million unique visitors stop by while the ad is up. These spots are so hot that the portals, like TV networks, sell them long in advance. And as a condition for prime real estate, portals demand that advertisers buy inventory on their less popular pages.

Joanne Bradford, MSN's chief media revenue officer, believes the flight to quality is already under way. The niche sites are going to have a harder time competing; the demand outside the elite sites will start cooling off in 18 to 24 months.

At the same time, leading niche sites, blogs, and social networking pages should take advantage of their relatively low prices and targeted audiences to attract advertisers. To join the elite, however, they will have to provide advertisers with reliable numbers.

For this, many are already piling into associations that can vouch for a level of quality and provide uniform metrics. Burst Media one of the early adopters for blog ads, offers advertisers access to nearly 2,000 Web publishers in 407 different categories. It's reaping benefits now!

Small sites provide the audience segmentation that advertisers want but let's face it, what advertisers really want is premium targeting at the price of cut-rate sites.

That's the premise behind behavioral advertising, whose long-cherished goal is measuring how people react to ads.

Companies such as DoubleClick were created to fulfill that promise, but it was impossible with 1990s-era technology, not to mention the 19990s mentality with respect to the web.

But not anymore.

Behavioral agencies such as TACODA and Revenue Science track the online sessions of Web surfers. By tagging them with a cookie when they visit one of the agency's thousands of affiliate sites, the system can follow a single surfer, say, from Yahoo to a popular health site to an obscure NGO site that has a narrow audience.

While the agency's computers don't know the surfer's identity, they can deduce from the stop at the general health site that they're dealing with a specific profile by following the navigation decisions. But that site is noisy, and lots of other ads jostle for the viewer's attention. So the behavioral system hits the viewer with a specific health ad when he's on the much narrower NGO site.

But hey, not just any health ad but an ad that reflects the past pages that were visited. It's a form of advertising arbitrage, making money on the spread between premium relevant placements and cheap sites.

This approach accounts for 8.3% of Internet spending volume today, according to eMarketer Inc. For behavioral targeting to continue growing, the agencies must provide advertisers with more numbers.

That's where the eye-and-brain scan comes in. In tests late last year, TACODA's researchers recruited 30 human guinea pigs at malls in New Jersey and Southern California. They hooked them to an eye-scanning camera and recorded every darting movement as the subjects were shown 50 identical Web pages.

The result: The ads placed on pages unrelated to the advertisements' message actually attracted 17% more looks. To see if the messages sink in, TACODA is planning more brain scans.

It has always been the goal of advertisers to work inside our minds - mainly to get their message in. Increasingly, the key is to measure the responses and gauge what happens next.

But how many of the readers actually focus on the ad?

Studies show that they take in only an average of one of every 12 Internet ads. What's more, in display advertising, even the more concrete metric of clicks is questionable.

But we all know that click measurement has been abused, there's no relationship between clicks and brand awareness. Or is there?