Attribution Drama

Which cookie wins? Should we pick a first cookie model, last cookie model or give different weights when attributing a sales towards different channels? They used the tool Action Allocator which is a tracking tool at which you can follow the customer journey. They also looked at how many touch points would need to get a conversion and what they could do with this information.
At the case with Tipp24 (gaming website) there has been an average of 2 touch points before an actual sale. These touch points has been divided in 6 different channels:

SEM / non branded search
SEM / branded search
Affiliate post view
Affiliate click
Display view
Display click

At that way they found out which combination models happen very often and which not. But what attribution model should you use? At Tipp24 they tested 3 models: first cookie counts, last cookie counts and a weighted model. With this research methodology they are able to see what happens in terms of revenue and commissions compared to the different channels. What happens is that with a weighted model and a first cookie counts model (mostly the largest) transactional affiliates will lose revenue. And you don’t want that to happen right? So, if you want to create a good system you should create a different model for each type of affiliate. Is it a content informational site? Is it a transactional oriented site? Or is it a get-paid-to site or maybe cash back concept? The business model of the affiliate should be the decisive factor for commissioning. That’s going to be quite hard. Our industry would be better off with one or just a few contribution models. With this solution it would become one big puzzle.

Shouldn’t we better of stop discussing this topic? Stick to last cookie counts for transactional oriented affiliates and just do something with those affiliates which are in the oriental and interested phase of the buying process?