Why Facebook Would Buy Atlas

6 Dec

Today Venturebeat, AllThingsD, Business InsiderReadWriteWeb and others are covering the story that Facebook might buy the Atlas ad-serving infrastructure from Microsoft’s aQuantive assets.

I may be a simple caveman, thawed from the ice by your scientists, but as somebody who has built ad-based web-sites with millions of daily impressions, talked with advertisers and agencies, and built a mobile ad-targeting/-network company which sold to Facebook in early 2011, I’ll tell you that this deal is not about a build-versus-buy decision around the “complexity” of web-wide-scale ad-serving. This is a really great move for Facebook, probably the smartest move I’ve seen from them in a while, but not at all for the reasons these early stories are saying.

Nope. I’d say it’s all about (a) the ad-entry, campaign-management, and reporting tools, (b) the relationships, and (c) the cookies & data.

First, the ad-entry and campaign-management/reporting-tools which Atlas has are not only familiar to ad agencies, brands, and advertisers around the world, they are in fact often required. When you create ads, resell ads, or are a publisher making space for ads and you go after big accounts, you are required to use either Microsoft/Atlas or Google/DoubleClick. They just don’t want to mess around with any custom reporting formats and they don’t want to have to train account managers, agencies, or anybody in their business about how to use your custom or weird tool. Even if you have a custom or weird tool, or a custom ad-format, or a custom ad network, you are asked to integrate Atlas or DoubleClick tracking into your ad. The account will glance at your custom snazzy reports, but they are verifying their ad buy with you based on the Atlas or Doubleclick dashboard and reports.

Secondly, the Atlas team has a huge number of relationships with agencies and brands. Facebook would get those relationships and contacts.

Finally, and perhaps more important of all, Atlas’ backend database contain browser cookies and a vast quantity of inferred demographic data which Facebook would be buying to blend in with its own demographic data and cookies. This would round out its cookies and demographic data for ad-targeting across the entire internet.

Serving billions of small bits of HTML, tracking their placement, click-through rates, and collecting additional information about who has viewed and interacted with those bits of HTML — that is not a build vs. buy decision for Facebook. They do this better, cheaper, and faster than anybody else on the web — you’ve heard of the “Like” button, right? And the Facebook commenting API? Right, that’s something that Facebook does just fine. (In fact I will argue here at some point that they probably do it at 1/5th the cost of the second-best player, Google).

Facebook’s ad-campaign, campaign-management, demographic-targeting and reporting infrastructure has only recently gotten beyond self-serve quality — only recently at the level of complexity buyers consider baseline coming from Atlas and Doubleclick. Buying this tool-set and team to jigger the back-end ad-serving and tracking to run on Facebook infrastructure  is such a good idea, as is using the key people on that team to enhance Facebook’s tools and agency relationships.

Hopefully by tomorrow there will be more in-depth reporting about what’s actually interesting about this deal, because it is a big deal, and it is a very smart deal. But the reporting you’re reading today is, I think, stupid.

About these ads

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 70 other followers

%d bloggers like this: