Tacoda Forms Behavioral Ad Network, Challenges Paid Search Players

Behavioral targeting technology player Tacoda Systems is developing a pay-for-performance network of content sites that will run text ads and leverage its audience-profiling technology, ClickZ News has learned. The program, slated to launch in late spring or early summer, is a clear challenge to the contextual ad programs run by paid search players like Google and Overture.

“There is so much demand from advertisers for cost-effective, scaleable advertising that’s easy to buy, that’s targeted,” Dave Morgan, CEO of Tacoda Systems, told ClickZ News.

Tacoda has so far convinced at least four of its existing publisher clients, though it won’t say which, to join its new AudienceMatch Network. Tacoda’s clients include USAToday.com, Tribune Interactive, Condénet, Advance Internet and Weather.com. Some sites have already begun testing the program, with more to begin over the next few weeks. Further expansion is expected over the summer. Non-Tacoda clients will also be invited to join the network.

The new network will compete most closely with contextual ad programs run by Google and Overture, jockeying for position on publishers’ pages and in advertisers’ wallets. Tacoda plans to use pay-per-click pricing and a self-serve auction marketplace — conventions popularized by the search players. The biggest difference will be that while the search companies target ads based on what their technology determines to be the content of the page, Tacoda’s network will target its ads based upon member publishers’ knowledge of the site visitor. That knowledge will come from registration data and from information about the person’s behavior across the network: what type of content the person visits, what he searches for, etc.

“It’s much more intuitive for a marketer to buy a person than to buy a keyword,” argues Morgan. “Over time, we think we’re going to see these types of advertising [contextual and behavioral] complementing each other.”

The scheme may sound reminiscent of late-’90s ad networks run by Engage and DoubleClick, which sought to build vast databases of user profiles, and which raised the hackles of privacy advocates. In Tacoda’s case, all audience data will be owned and stored by site owners. The company will simply serve as a go-between, Morgan said.

“I do not think it’s possible to federate a network and own data,” said Morgan, who got plenty of network experience in his previous job as CEO of Real Media. That company was later acquired by, and folded into, 24/7 Media. Morgan cited the need to avoid treading on users’ privacy and to avoid competing with publishers.

A hypothetical example of how AudienceMatch would work: an advertiser would place an order on the network, saying he wanted to reach 18- to 34-year-old men with a demonstrated interest in baseball. A text-based ad would be developed. When an individual landed on a particular page on the network, the Tacoda technology would check with all its networked sites’ databases to determine whether that person met the appropriate criteria. One site might, because of registration data, know the visitor was male and fell within the designated age bracket. Another site may have noted the person spent a lot of time perusing baseball-related content. When a match occurs, the appropriate text ad is delivered to that page, though data are never shared.

“For a number of advertisers, we know that it can dramatically outperform contextual,” said Morgan.

The advertiser only pays if the site visitor actually clicks on the ad. Tacoda doles out revenue generated when an ad is clicked based on where data originated, where the ad appeared, and who sold the ad. Tacoda will sell ads directly. Publishers in the network will also be able to package network inventory with their own.

So far, the move to develop a network, which will exist alongside Tacoda’s existing business, is hailed by industry analysts.

“Creating a network is absolutely the right move for them,” said Nate Elliott, an analyst at JupiterResearch, which shares a parent company with this publication. “There are significant limitations if you don’t network the data. If you are trying to collect and use behavioral data on a single site, there are problems with scale.”

(The folks at Morgan’s former company, now 24/7 Real Media, are also using behavioral data across a network, but it’s not text-ad-focused, nor is it pay-for-performance.)

One of Tacoda’s first challenges will be to convince enough publishers to participate in order to achieve the necessary scale. A factor to the company’s advantage is the weakness of contextual targeting in circumstances where awkward juxtapositions sometimes occur.

“There are some pages where you don’t want to run contextual ads,” said Charlene Li, an analyst with Forrester Research. “Like hard news stories, because of all of the unfortunate situations that can occur.”

Adds Jupiter’s Elliott: “The publishers who are taking a look at these products, if they’re smart, will put all three of them up, or all 10 of them up, and they’re going to let them run for a few weeks. It’s going to vary from site to site. They should run some A/B tests on their inventory, or some A/B/C/D/E/F/G tests on their inventory.”

Related reading