• by Ross Fadner , November 15, 2004

Behavioral targeting provider Tacoda Systems Monday unveiled a behavior-based pay-per-click ad network that it will market to direct response and search advertisers. By combining the audiences of its publishers into defined segments, Tacoda believes it will be able to offer advertisers the reach they need for direct response campaigns.

Tacoda's network, called AudienceMatch, will operate on a bid model similar to Google's AdSense or Overture's ContentMatch: advertisers will bid against each other to place text ads within AudienceMatch units, which will appear on certain publishers' pages within Tacoda's network.

The network will be rolled out slowly over the next several months, starting with 60 publishers and 100 advertisers at the beginning of its two-month "launch-phase," and expanding to nearly a thousand publishers and a thousand advertisers. By mid-January, Tacoda expects the network to reach 85 million individuals--about 60 percent of the monthly Internet audience.

The AudienceMatch launch publishers include USAToday.com; Media General's sites, including Tampa Bay Online and the Richmond Times Dispatch; and the Virginian Pilot Online.

"By and large, our publishers were excited," said Morgan. "We had no trouble finding willing participants." But, he said, some publishers were "more conservative than others" in their reception to the network.

The New York Times Digital, for one, will not be a part of Tacoda AudienceMatch. NYTimes.com Vice President of Sales Jason Krebs said Tacoda didn't approach the company with the network idea.

Tacoda first announced its intentions to start combining audience segments from different publishers in April. Since then, many industry observers have wondered whether Tacoda's plan makes sense, or whether obstacles--including privacy concerns and questions about whether individual publishers' sites would be diluted--would prove overwhelming.

Dave Morgan, Tacoda founder and chief executive, said that neither hurdle was insurmountable. "Privacy is something I worry about a lot," said Morgan. He said Tacoda dealt with the issue by placing AudienceMatch listings in an ad unit called the "Offer Inbox," which alerts consumers that the ads are targeted by behavior, and gives them the opportunity to opt-out of receiving those ads.

He emphasized that the network is called AudienceMatch because the system simply matches publisher data without sharing it, which he said is "essential to both data ownership and privacy."

Alan Chapell, a privacy specialist from the law firm Chapell & Associates, remarked that "the only real complication is in perception" as far as privacy is concerned. He said the network practices, as described in the press, sound kosher--but added that the threat of HR 2929, a Senate bill that could render the provisioning of cookies illegal, remains a future concern.

Morgan also said that network aggregation makes sense for direct response units, such as pay-per-click text ads--but not for publishers trying to sell image ads. "We're creating Google AdSense for behavior targeting," said Morgan, who added that, on an individual publisher-to-publisher basis, Tacoda publishers can sell targeted image ads.

GartnerG2 Analyst Denise Garcia said that larger sites will most likely not want to participate in the network because it might create tensions between their internal sales forces and that of Tacoda, which would handle sales on behalf of the network. She said that the ad sales process becomes "commoditized" through a behavioral targeting network, which focuses less on a Web site's content and more on the shared characteristics of various publisher audiences.

But Garcia emphasized that behavioral targeting in and of itself doesn't commoditize an audience; rather, being part of a network of segments does. She said that for advertisers simply looking to hit a target, buying on the network saves them time--and more crucially, greatly expands reach.

Niki Scevak, a JupiterResearch analyst, said Tacoda's pay-per-click ad network will most likely compete with contextual ad networks for remnant inventory on publishers' sites. He said the behavioral network will have to hit a certain scale in order to perform effectively for the various parties--something he called "a huge challenge." The short-term test, he said, will be whether publishers earn more with Tacoda's network than with contextual providers like Quigo and Kanoodle.

Omar Tawakol--senior vice president of product marketing for Revenue Science, a rival behavior targeting firm--called AudienceMatch a "different beast" from what Revenue Science offers its publishers and advertisers.

Tawakol said Revenue Science's target advertiser base is national brand advertisers looking to buy high-value publisher cost-per-thousand inventory with branding dollars. He said advertisers can expect something "very different" from Revenue Science, which is currently developing its own expected behavioral targeting advertiser network.

Morgan said Tacoda hired between 15 and 20 new people to concentrate solely on the ad network. The biggest challenge, he said, was to implement the new technology into the company's existing behavioral targeting technology without having to build a separate, centralized database.