Facebook has shifted the focus of its ad tech platform Atlas and moved it into its measurement division
The change, which took effect around two weeks ago, means Atlas now sits in Facebook's "marketing sciences" group, which is run by Brad Smallwood, the company's vice president of measurement and insights.
The ad tech division, which contains the lucrative Facebook Audience Network, will continue to be run by Brian Boland, Facebook's VP of ad tech. David Jakubowski, who was running Atlas, will also remain in Boland's team. The rest of Atlas will now work under Smallwood.
There will be no layoffs as a result of the change and the Atlas brand will stay for now, Facebook confirmed.
The restructure is a recognition of how Facebook's use of Atlas has dramatically changed since it was acquired all those years ago.
It's all about "people based marketing" - not ad buying
Back then, Facebook's acquisition of Atlas - an ad server and measurement platform - was seen as a way Facebook could make a huge dent to Google DoubleClick's monopoly of the display ad business. Atlas was a way advertisers could use Facebook's deep targeting data to buy ads on other websites not owned by Facebook.
Facebook had been working on building a demand-side platform (DSP) - software that lets advertisers use automated processes to buy online ads - but in March this year it pulled the project. The company said it called off its DSP plans because there were too many bots and bad-quality ads on the open web.
So Atlas became, predominantly, a measurement tag advertisers can use to match up users as they cross from one device to another. The ad server part of the platform still exists, but it's used mostly for "data fidelity" to ensure accurate measurement rather than acting as a true competitor to DoubleClick.
To some people, that will be seen as a shame - like buying a Ferrari only to use the steering wheel. When AdExchanger reported on Atlas' progress in February, the trade title was fairly derisory that most clients were simply using the platform for measurement, which it said was "not the promise of Atlas."
One of Facebook's big plays to advertisers is that it offers "people based marketing." The idea of Atlas is to take that approach to measuring ad performance to the entire web.
Facebook has expertise here because many of its users log into the social network on their mobiles, desktops, and tablets, which gives it first-party data. That means Facebook's Atlas can let advertisers know if someone who saw an ad (from across their entire digital campaign - not just ads on Facebook or ads bought through Facebook) for a pair of sneakers on their mobile phones went on to buy the product on their laptops. It also lets advertisers know if they reached the audience they paid to reach and not just the same people over and over again.
Boland said: "Our effort in measurement was to approach it not with a lens to do everything to make Facebook look good. It's to do everything to make people know and and believe this is truthful. We will stand by that truth: Either Facebook ads work [better than ads on other publisher sites] or they don't. And there's no hiding."
That comment is timely as it comes a week after Facebook admitted it had been over-inflating one of its video view metrics. Facebook has corrected the error and VP Carolyn Everson has since said the company plans to be more public straight away if it makes mistakes again in future.
Smallwood said the aim isn't to encourage advertisers only to use Atlas, but to use a range of measurement partners to analyze their digital ads. However, Atlas charges advertisers on a cost-per-mille (CPM/cost per thousand impressions basis) so it's likely Facebook will be encouraging advertisers to use Atlas as their go-to digital ads measurement partner.
Ultimately, the move represents what was right to do for the business, according to Boland.
"I like that a work at a company where it is culturally acceptable for a bunch of people to move across the organization to a different place," he said "We have a culturally-installed set of values about doing what's right for the future of the business. I've shifted responsibilities a couple of times ... we're a very customer-centric organization and we look at ways to remove things that hinder you from executing well."
It has been an eventful year for Facebook's ad tech group
Atlas has never revealed the total number of advertiser partners it has (although it did disclose the names of 10 customers last year), nor how much revenue it generates. The acquisition was nowhere near Facebook's biggest: TechCrunch reported the transaction price was between the $50 million to $100 million mark (but closer to $50 million). But nevertheless, it's certainly a change in direction product-wise from when Atlas was first bought in 2013 - even if the "people-based marketing" goal remains the same.
As for the ad tech group, that might change too.
Boland said: "I'm moving away from calling it the ad tech group. 'Ad tech' has lots of pre-conceived notions, very anchored to the LUMAscape of 2009 - cookies and desktop. I'm very focused on the monetization for publisher side of the business ... and how we solve real problems for advertisers, how it grows the ecosystem ... not in the abstract: People-based marketing."
(The LUMAscape is the infographic created by investment bank LUMA Partners, which details where all the players in the ad tech ecosystem sit.)
There has been a lot of change in the ad tech division already over the past few months. Facebook shut two big projects this year: FBX, its desktop exchange, and LiveRail, the video ad exchange it bought for around half a billion dollars in 2014.Facebook Audience Network (FAN) is Facebook's real ad tech strong point outside of Facebook's own ad platform. It's an old-school ad network model, but across mobile and it's a money-making machine. The company said FAN had a $1 billion annual revenue run-rate last year, although sources have told us the number now is closer to $2 billion. Boland declined to comment on the number.
Facebook is "very good at learning"
So, FAN aside, is the Atlas re-org a final nail in the coffin for Facebook and its attempts at building an ad tech stack? Has Facebook finally resigned itself to leaving that fight to Google and the rest of the LUMAscape?
Perhaps that's the wrong way to look at it.
Michael Collins, CEO of ad tech company Adelphic, told Business Insider: "Gaining access to Facebook's data outside their walled garden will improve measurement and attribution. Also, making this data available across real-time bidding and walled garden campaigns shows that Facebook is finally recognizing the value of the larger programmatic ecosystem."
Rob Norman, chief digital officer at WPP's media investment management division GroupM, told Business Insider he thinks Facebook has given up on building an alternative to DoubleClick.
He said: I think Facebook is interested in being big in the Facebook tech business. They are in love with their platforms and the potential number of use cases across Facebook, Messenger, Instagram, and elsewhere. The focus is to do more and better with these platforms."
Every now and then Facebook has made some bets that haven't worked, such its Facebook Paper curated article reader. The app flopped and was shut down in July this year but Norman suspects many elements from Paper were transferred across to its Instant Articles product. Norman predicts the same has and will happen with Facebook and its ad tech projects.
"They are very good at learning and you have to admire that about them," Norman said.