Search marketing was founded on the rock of cost-per-click, but lately large chunks of that rock seem to be eroding or turning slippery with moss, thanks to click fraud fears. As a result, both search marketing power Google and other smaller online ad players are trying out new models that earn revenue not from clicks but through other performance milestones.
Cost per action, the models are called, with advertisers defining the actions they’re willing to pay for. And while most of these platforms are still in the larval stage, a few already have enough heft to attract the attention of both big-name venture funds and real live marketers.
One of these start-up ad networks, Turn.com, grabbed headlines at last November’s Web 2.0 conference both for what it promises to do—charge advertisers for getting consumers to perform a specific action—and for the way it expects to do that—by automating ad distribution with a blend of behavioral and demographic targeting tools, taking the decisions for placement off the advertiser’s task list.
That’s something many online marketers would be glad of, says Turn.com CEO Jim Barnett.
“These debates and discussions about contextual advertising versus categories, or demographics versus behavioral targeting—they’re really nonsensical,” he says. The answer is that they all make sense. But they have to be blended together. Successful ad networks of the future are going to apply very high-end technology to synthesize these targeted approaches, weight them for every single ad call, and then select the right ad for that particular user on that page at that time of day in that geography.”
What that means in practice is that ads will be served up on the fly to visitors to the sites in the Turn.com network, which counted about 30 publishers and 200 domains when it unveiled its beta last fall. Rather than rendering an ad based on pre-set demographic or behavioral factors, a visit to a Turn.com page will start an automated calculation of more than 60 variables, including page content, ad performance, user profile information the publisher might have gathered and optional keywords the advertiser has specified.
Advertisers may come to appreciate the automated nature of the Turn.com platform because it will free them of continually having to optimize lists of thousands of keywords. They can simply upload all their creative into the system with a self-service interface and specify the actions they want—whether a lead, a site visit or a transaction– and the price they’re willing to pay for them.
Turn.com deploys tracking beacons on both ads served and on network publisher pages in order to monitor and report on user actions, impressions or clicks, reporting that data back to both publishers and marketers.
The model seems particularly built for direct-response advertisers. Barnett says it’s still early to evaluate the type of marketers Turn.com will appeal to, but that as more marketers come to appreciate the benefits of both ad automation and the cost-per-action model, he expects Turn’s ad client roster to grow substantially. At beta launch, the company had enrolled about 1,000 advertising clients.
Of course, managing to an action other than a click will almost invariably be more costly than pay-per-click marketing. Barnett says it’s also too early to talk about the levels at which Turn’s prices are settling, but he does point to some broad CPA trends on online networks. “Visits on the Internet broadly average less than $1, while leads can be priced anywhere from $1.50 to in some cases $40,” he says. “A lot of leads today in the CPA space tend to be in the several-dollar range. And transactions tend to range from the high teens to the $20, $30, $40 or $100-dollar-plus range.”
“Right now we’re focused on optimizing our targeting technology,” says Barnett. “We’ve already opened up the network to advertisers, and next we’ll concentrate on opening up for publishers. Right now we’re still approving them on a selected basis, but in Q1 2007 we’ll open the beta up to publishers as well.”
One Web publishing area in which Barnett thinks Turn.com’s automated approach will be particularly applicable will be social content and community-generated pages.
“Almost by definition, you can’t manually target social content, because of the massive volume and the high diversity of the pages,” he says. “Sometimes you don’t even know what’s on those pages because they get created so quickly, and they can change by the time you’re serving an ad against them. So having an automated system to analyze those pages quickly should really outperform any kind of manual targeting.”
Network
