End-of-the-Line for Last Click Attribution ? End-of-the-Line for Last Click Attribution ?

End-of-the-Line for Last Click Attribution ?

by ,
on 6 March 2018


Announced during the annual Marketing Next conference on May 23,2017, Google Attribution is the latest from Google to put an end to the last click attribution model.

Why the last-click ?

The last-click attribution model has become the standard over the years. Strongly pushed by Google, this model had the advantage of favoring the Google Adwords advertising model. The latter was based on paid advertising in search results and for years, the conversions in this channel were positioned at the end of the purchasing tunnel which created very flattering results for Google Analytics as compared to other acquisition channels notably “classic” displays.

The repeated limits of last-click attribution 

But remarketing has come through. And with it, numerous players who decided to take advantage of a simplistic attribution model that could only gratify the last player in the chain. Criteo and other retargeting solutions, affiliation based in large part on promotional codes, email retargeting, etc., and the sector concentrated on the end of the tunnel as comparative measurement with different actors favored this.

Then mobiles revolutionized use, and over the years, cross-device attribution became a major problem for all the players in advertising and Web analysis, with Google leading the pack.

But Facebook took advantage of the situation by highlighting its ability to reconcile cross-device attribution within its platform as users were obviously connected. Further, post-impression and cross-device conversions made it much easier for advertisers to justify branding budgets.
This demonstrated that the sector was no longer counting on Google only for performance analysis and Facebook’s discourse on Google’s limits bore fruit, with numerous advertising budgets changing networks.

At the same time, Google needed to continue to grow and had to propose other advertising offers, almost always more branding-oriented such as notably Youtube. These also suffered from the comparison imposed by last-click attribution.

With this overview of Google’s situation, it’s easier to understand its desire to put an end to last-click attribution. It has become a major economic challenge: to develop its advertising offer by granting it more credit than before and to end Facebook’s progress.

Google Attribution and the Data-Driven Model

Over time, different attribution models had appeared in Google Analytics and in Google Adwords as well as the possibility of comparing them to last click.

This first step was interesting but even with this information, the same questions remained: which model to choose? For what reasons? How to determine if it was adapted to my sector of activity? Should it be regularly revisited?  All these questions being asked by many advertisers for several years explains the persistence of the last-click model despite its faults.

The Google solution to finally move on to something else is two-fold:

  • To create a dynamic attribution modelthe data-driven model
    This model based on an algorithm that automatically adapts to changes in user behavior and to new activated channels makes it possible to resolve 3 problems: no need to choose a model, the model is the same for each sector (it adapts alone), and no need to regularly revisit the model.
    This first step is already in place with the data-driven attribution model available in Google Analytics 360 (the paid Google Analytics version) and in Google Adwords. Besides, it is currently more power in Google Adwords as it takes cross-devices into account and enables direct optimization activation.


  • To launch a product dedicated to attribution makes it a relevant subject: Google Attribution
    To hasten last-click’s demise is the first objective of Google Attribution as well as forcing advertisers to take a stand on the subject and not to content themselves with the default model. Further, this allows Google to evolve its offer in relation to existing attribution solutions such as Mazeberry. So instead of being an annex parameter in Google Analytics or Google Adwords, attribution will be a proper tool, able to connect to other Google tools for a unified model.

Google Attribution will be available in several months and will allow Google to communicate on the subject of attribution specifically, highlighting it as a separate building block in its strategy. It will then make it possible to push its attribution model to other Google tools: Analytics, Adwords and DoubleClick.

So, for now?

Since this conference, the message relayed by Google to agencies and top advertisers is clear and pushes for the replacement of the last-click model by a data-driven model.

Of course, it’s possible that friction will be too important for this change to occur. The ease of last-click may push many to not want to change a model they know well.

But Google wants you to change models, not because it’s better for you – even if that will certainly be the case – but because its business models are suffering in relation to Facebook.

It’s true that this new model will also have its faults including: lack of transparency and relative complexity compared to the truly simplistic last-click. This could deter numerous advertisers.
One could also be wary of a post-click cross-channel model that’s only post-impression in the Google world (with DBM or Adwords) which will obviously bestow it an advantage over Facebook and other companies in results analysis.

However, despite its faults this new model will have considerable advantages over last-click.
And it’s high time to move on to something which – even if imperfect and pushed by Google for its own economic reasons – will allow post-impression, cross-channel, cross-device, and automatic analysis. And so, to be everything that last-click attribution is not in an era where our professions can no longer do without this type of analysis.



, , , ,

to read next...

Leave a Reply

Your email address will not be published. Required fields are marked *


End-of-the-Line for Last Click Attribution ?