Facebook prepares to amp up its ad war with Google with Atlas

Just as in Greek mythology, Atlas, the ad-serving product reportedly headed soon to Facebook, is tasked with an enormous burden: upholding the social network's advertising empire. Don't shrug off Atlas; it will be the vehicle that Facebook uses as it tries to unseat Google, the current ruler of display advertising.

Rumors began swirling in December that Facebook was preparing to buy Atlas, the advertising platform that Microsoft picked up when it purchased aQuantive in 2007. The deal, according to Ad Age, is now nearing its close, and Facebook will pick up the property for less than $100 million.

The deal will bring to fruition months of speculation around the social network's intent to develop its own ad stack, or a one-stop shop where advertisers and agencies can, in bulk, purchase, sell, optimize, and track ads across the web. The clear objective: to carve out a new space outside the confines of the social network where Facebook can deliver socially powered ads -- and take over more of advertisers' display advertising budgets in the process.

Buying Atlas "is absolutely a direct line to creating an ad network," Rebecca Lieb, digital media analyst at Altimeter Group, told CNET. "It's also a direct run at Google."

Atlas rivals Google's DoubleClick product for serving ads on the web, which means the buy would go "against that very core part of Google's business and revenue," Lieb said.

Put simply, she said, Facebook wants Google's inventory. "Facebook is going after a limited set of valuable ads, and it's going after them with incredibly strong data -- data that rivals what others have on us."

As it stands in display land, Google earned 15.4 percent of the U.S. display ad revenue market in 2012, according to estimates from eMarketer. Facebook was a close second. The social network captured 14.4 percent of the net U.S. display ad market, which eMarketer valued at $15 billion for the year.

"Facebook has grown strong in the display ad space, and its mobile display business is strong," Clark Fredricksen, eMarketer's vice president of communications, told CNET. "The impact of an ad network could be significant."

Facebook already knows how to run display advertising on Facebook, of course, and it uses its nascent Facebook Exchange to help advertisers target ads to you based on your behaviors off of the site. Atlas, which helps companies place ads on Web sites and track their effectiveness, would help Facebook take advantage of its greatest asset -- the mountains of data on who we are, who we know, and what we like -- to place highly targeted units in front of consumers as they browse the web.

Facebook knows, for instance, that you "like" Virgin America, which means it could show you a Virgin America ad, instead of one from Delta, the next time you visit a travel Web site. It's this "like-level" data, coupled with behavioral data, that Williams believes Facebook can use to shift wallet share and budgets away from Google.

The social network may have an advantage with some ad buyers. An array of publishers already use "Log In with Facebook" buttons on their sites and make use of the company's Open Graph platform to exchange member activity information between their sites and Facebook. In theory, Facebook could also design ads that feel blended with the publisher's site -- just not yet.

If you're expecting Facebook to revolutionize the look of the display ad and make it akin to a Sponsored Story, as we've seen in tests on Zynga.com, you're mistaken. For now, Facebook would need to abide by the Interactive Advertising Bureau's standards around the format of banner ads. The units you see littered across the web are one-size-fits-all for a reason: They let agencies build ads that can be plugged into most publisher sites.

There's also the problem of integration. Facebook can run Atlas out of the box as a standalone ad server on day one, but embedding the social network's data in the system won't happen overnight. It could take months or years before Facebook is able to integrate Atlas in a way that would give advertisers what they want: wider distribution and better targeting.

The ad stack theory is a sound one, but Ari Paparo, who previously worked for ad platform AppNexus, a company that was said to be bidding on Atlas before Facebook entered the picture, has a different theory. Facebook, he told CNET, has an attribution problem, meaning the social network has a hard time showing advertisers the results of their spending on the site.

"The industry has used a 'last click' model for the past 10 years wherein all the credit for all ads is given to the click that generated a visit to a Web site" he said. "The primary mechanisms for attribution [DoubleClick for Advertiser and Google Analytics] are owned by their archrival Google."

Atlas, which can track the performance of ad campaigns, would give Facebook a way to fix its attribution problem, and that's a solution that could convince advertisers to spend more on ads that run on the social network.

Whether Facebook wants Atlas to create its own an ad stack or solve its attribution problem, or both, is debatable, but one thing is clear: It's gunning to steal business from Google.