7 Dec 2012

Why FBX could force agencies to break the single ad server model

With FBX rumored to be in 'advanced talks' to acquire the Atlas ad server from Microsoft, and AppNexus supposedly out of the running, could operational pressures force agencies into using multiple ad servers?

The first consideration is the power that FBX would wield - 'if you want to play with us, then you have to use our toys'. Possible, but unlikely. Facebook wants revenue, and lots of it, and isolating big parts of the market is not the way to go about it.

Instead we need to look at why ad servers existed at all - 3rd party verification that you are getting the number of impressions you paid for, easy campaign management and easy creative management. Holding on to these benefits would only occur if you ran all your media buys through one platform - and if you want to include FBX in the buy then that platform has to be Atlas.

And if we look into the near future, the pressure increases. I talk to many CMOs who say they are finally getting to grips with attribution modeling, and beginning to understand media in a more holistic fashion. Given display should always be measured by impressions, not clicks, including the impression-level data from their ad server is critical. Currently the FBX View Tag allows for conversions to be tracked with Dart or Atlas tags, but not impressions. If FBX does buy Atlas, and allows full impression tracking, Atlas could become the needed platform for attribution.

Some of these problems can be solved of course by getting impression-level data from two ad servers, making sure it's time-stamped and then blending the data together. For some that might be enough, but if you are contemplating this path, know you are in for a world of pain caused by a cookie loss of up to 40% when trying to bridge all the systems, and just the sheer logistics of trying to make this data actionable and meaningful.

So Atlas is looking quite good right? It is a second place technology, good heritage, has benefitted from being a SaaS product inside a reputable software development shop, is the same price and it can do the job...

Well Dart agencies will disagree; they will tell you about speed and capability, and many love the tight integration into the Google technology / media stack. They will talk about the headache of retraining, as well losing lots of historical data.... there are in fact lots of reasons, I don't need to go on - it's highly unlikely FBX acquiring Atlas could force a rush from DoubleClick on its own.

More likely, we will see agencies using both platforms and dealing with the consequences. And if that happens, it creates two intriguing situation:

1. It puts FBX in a bucket all by itself. Marketers are telling me they don't know what to do with FBX - on the one hand it's just another source of exchange / RTB inventory and some feel it should be treated that way. Many are uncertain though; they get hung up on this being in a social environment and question what team should manage if, how consumers will react and how it should be measured. They aren't wrong in my opinion, there are lots of questions that need monitoring. But for certain, if they force a connection to Atlas, and Dart users start to feel like the ugly sibling, then immediately FBX is forced into being 'something different'.

2. It might force agencies to move away from last-click attribution, something I would be delighted to see!! If everything can't run through one ad server any more then last-click or last-touch attribution has to go away, and a more sophisticated model has to emerge. Woo hoo! Perhaps no more site retargeting credit for everything... :)

So my prediction, if the acquisition occurs, is that it will be a good time to be an Atlas sales rep, they will pick up some agencies who shift completely, but certainly get a lot of new accounts, but all spending on FBX only.

One question remains for me. The reports say that this acquisition would be to help Facebook create an off-site network - why buy an ad server and not a DSP, or better still, a Programmatic Marketing Platform?


1 comment:

Sean Blankenship said...

Great overview Dax. As a marketer, I'm pushing the organization into attribution modeling. Analytics fragmentation is very challenging; multiple sources, multiple agencies, multiple agendas, and of course risks (cookies) abound. At the end of the day, my belief is that we must ultimately pick our core KPI's and force the digital ecosystem to perform accordingly. Attribution will help us define and optimize those core KPI'S