Cost-Per-Action Advertising As Powerful As Algorithmic Engines That Run It

Performance-driven advertising is all about action. Performing an action is a significant leap forward in a user's conversion path. However, actions that are not translated into sales are, to put it simply, missing the target. After all, you can't go to the bank with clicks and engagements, only real dollars.

Enter pure performance CPA -- cost per action, where the action is as powerful as the machine-learning algorithmic engines that run it. It's the only display model in which advertisers pay only for actual sales. When a direct line can be drawn between money spent and money earned, the ROI is crystal clear.

The goal of a CPA-focused company is to generate conversions rather than clicks and engagements. Its interests are aligned with those of the advertiser and it will certainly run the extra mile and beyond to deliver actual sales because its own business depends on it.

If that's the case, then why is CPA display so scarcely offered? The answer is that the ability to fully support this model, from the core, is anything but trivial. Every pillar of a CPA-focused operation has to be built with a conversion output in mind: from data collection and analysis, optimization engines, real time bidder to campaign management and de-duping.

Optimizing for conversions rather than clicks is a complex undertaking, requiring cutting edge machine-learning predictive algorithms that can accurately forecast which impression will result in a cost-effective conversion. We’ve all heard this before, but it truly all boils down to showing the right banner to the right person at the right time in the right price. There is a science to this that many companies promise but can’t deliver on, and this is especially true when talking about non-CPA driven business models.

Overcoming the challenges of building a great CPA machine with the right infrastructure and properly wired algorithms that support it, CPA providers can sustain a model in which they assume all the risk while driving incremental sales for their customers.

When it comes to frequency capping, it's important to understand that CPA companies treat display with a surgeon's scalpel and not with a sledgehammer. Opting for quality rather than quantity, a "spray and pray" approach goes counter to the very essence of CPA. Since it pays for all the media but only gets paid if a conversion is achieved, it simply cannot afford to waste impressions.

Another factor to consider is de-duping. In CPA, the concept of de-dupe (and the systems that are able to support it) is central. Unwillingness to de-dupe is a complete disregard of today's complex path to conversion.                

And then there is the view-through conversion. I believe brands understand that these campaigns work even if they are not a perfect attribution science. After all, most ad spend is done offline where the view-through model rules.

Many studies have shown views lead to visits, improve search marketing and drive up revenue. Up-and-coming metrics like ad viewability, in addition to larger and richer ad formats, will further increase PV's value, as will a sensible 24 to 48-hour look-back window (rather than the completely illogical 14 to 30-day window that is still common today). Having said that, views should be considered an added value and play a secondary role to what matters most: true conversions.

3 comments about "Cost-Per-Action Advertising As Powerful As Algorithmic Engines That Run It".
Check to receive email when comments are posted.
  1. Donald Faller from double wide media, November 30, 2012 at 4:04 p.m.

    oooooo data speak!

    You mean you want to count people who you convince to purchase?

  2. Marie Achille from Alea Vip, November 30, 2012 at 8:16 p.m.

    This was the best article on how to get the best ROI I've read so far!

    Good overview on how to focus PPC marketing campaigns and get maximum results.

    Thanks Micheal,

  3. Pete Austin from Fresh Relevance, December 1, 2012 at 10:56 a.m.

    It's about ROI, return on investment, but this article is only about the "R". In the real world, you do better by focusing on the "I". Cherry-pick: start with small investments that get real returns now. Big returns require big investment, on big projects, but these are very difficult to predict and manage and may over-run.

Next story loading loading..