Data Atribution Distortion

Distortion of last click atribution

It is said that having some data is better than not having any data. In some cases this may be true, in others it can be disastrous when making business decisions based on this data. All data analysis is good as long as you have the context in which the data was collected.


We live in a world in which we are multi-device, we are impacted by communications in different media. The last click model, which most companies do their data analysis, does not tell you the full story of what happened on the “costumer journey”. It is as if we measured a soccer team only by that player who scored the goal. But we do not take into account who gave the pass, or who started the play so that that player could make the conversion … sorry for the Goal!

The atribution models

In the Indirect Last Click attribution model, all direct traffic is ignored, and all credit for the sale goes to the last channel the customer clicked on before conversion.

The First Interaction attribution model, the first point of contact, in this case the Paid Search channel, will receive 100% of the credit for the conversion.

The Linear attribution model, each point of contact of the conversion path, in this case the channels will share the same credit for the conversion.

In the attribution model Time deterioration, the points of contact closest in time at the time of conversion get the maximum credit.

The Attribution Model According to position, 40% of the credit is assigned to both the first and last interactions, and the remaining 20% ​​of credit is evenly distributed among the intermediate interactions.

In addition to the aforementioned models, it is possible to create your own personalized attribution model, according to what the analyst considers reflects the reality of the business in the digital medium.

The solution:

Let’s stop using the last click model and start experimenting with attribution models that give us “more real” data and allow us to delve into the consumer and our business.