Google Inc. isn't the only Internet ad purveyor that recently filed to go public. Advertising.com Inc. and Claria Corp. -- which match ads to Web-surfing histories rather than to search queries -- also registered last month to hold initial public offerings.
Their filings attracted none of the media hype of Google's. Yet these upstarts reflect a sector of online advertising that is growing nearly as fast this year as search-based ad systems were last year.
This sector is called behavioral targeting, at least this time around. Back in the 1990s it was dubbed personal profiling, when first-generation networks like Engage Technologies, Excite@Home and DoubleClick mined people's surfing patterns in an attempt to identify specific audiences, such as middle-aged accountants driving BMWs who travel overseas. Then came the Internet stock crash. Excite went bankrupt and most others quit the profiling business, which had triggered a huge privacy backlash.
Now comes the second wave of Internet profilers, employing a wider variety of methodologies with varying degrees of success. Unlike some past attempts, today's systems rely on anonymous profiles that lack personally identifiable data. Moreover, industry analysts say much of what was theoretical and experimental back then has since been baked into real products that advertisers and publishers are fine-tuning in an attempt to boost the efficiency of Internet ads.
"Behavioral targeting is finally coming of age," said Karen Anderson, media director for Internet ad agency Modem Media Inc. "That's because advertisers and marketers have caught up to where the technology providers wanted them to be five years ago."
Others are less optimistic. "This really is a nascent marketplace," said Nate Elliott, who is writing a report on the subject for Jupiter Research. "The big question is whether it can improve performance enough to justify the higher prices" the new ad placement services command.
Among today's vendors are Tacoda Systems Inc. and Revenue Science Inc., which offer software for letting Web publishers sell their audiences to advertisers in new ways. One goal is to sell ads on pages that haven't been popular with advertisers in the past by giving advertisers more data about who's reading them.
Under more traditional ad systems, ads are either paired with specific types of content appearing on a Web page -- an ad for Dell laptops might run on a channel or group of pages focusing on technology, for example -- or ads run across an entire network. Channel or contextual ads often sell at higher rates than run-of-site ads because supply is limited and demand is theoretically greater.
Tacoda's tools aim to expand the number of Web pages an advertiser might be interested in by identifying sought-after readers on other sections of the same site. The tools track what visitors read and correlate that with other details visitors provided about themselves when they registered at the site. The visitors can then be served ads tailored to their perceived interests. Some early studies suggest these targeted ads are generating decent click-through rates, which often determine how much a Web publisher is paid.
"We have delivered hundreds and hundreds of behaviorally targeted campaigns," said Dave Morgan, Tacoda's chief executive. "The results show that if they are well targeted, these campaigns can deliver results that are 20 to 40 times as good as contextual advertising in some categories."
