Affiliation, the side victim of the third-party cookie end?


The discrepancies between affiliate platforms and Google Analytics raise industry questions. If abandoning the last click by Google is the ideal reason, then the problem appears elsewhere.

Is Google’s abandonment of the last-click referral model in its ad products since last fall negatively impacting the affiliate marketing industry? “This phenomenon is still very limited, but the fact is that since the end of last year I have noticed a significant decrease in conversions, which is happening with some of our top merchants,” explains Guillaume Bornesin, Director of Audience Marketing at Le Parisien, at Origin Shopping Guide Le Parisien Le Guide. They will be no more than five advertisers to lower their engagement, among the 200 ad programs of this directory which generates about 25 million visits per year and whose revenue increased 30% in 2021 compared to 2020. But that is enough to raise questions.

“I have noticed a significant drop in conversions since the end of last year”

“These advertisers compared the performance tracked by the affiliate platforms we work with, which mostly operate according to the last-click referral model, with estimates provided by Google Analytics, which takes into account the contribution of other levers,” Guillaume Bornezin defines. The result: Considering that their sales may have resulted from other tools, these merchants tended to invalidate the conversions recorded in the dependency.

Guilhem Bodin, associate director at Converteo, a consultancy that specializes in measuring marketing effectiveness, stresses that the results often do not agree: “When we compare the results of the affiliate platform with those provided by Google Analytics, we often see significant discrepancies that are not always in favor of affiliation. “. But for this specialist, this difference can only be indirectly affected by Google ads last fall, since the latter is associated with the Google Ads program, and not with Google Analytics. In fact, Google has replaced the default last click with data-driven modeling of all touch points that occur in the Google Ads environment (search, shopping, viewing, and YouTube). “The end of the last click that Google announced in the fall was not directly related to Google Analytics, as it has migrated to a data-only attribution model since the launch of version 4. In the previous version it is still active, with the last click model remaining by default, however More mature advertisers can configure analysis taking into account many touchpoints.And those advertisers out there, the more mature, who know how to activate multi-touch referral functions on Google Analytics, are the ones who are able to notice larger contributing differences between the information the analytics tool references and the data from basic systems.

Differences that don’t surprise the affiliate platforms themselves: “The difference is important because we don’t count things the same way: affiliate platform trackers show conversions after a click on a tracked link, while Google Analytics starts counting when the page doesn’t start from the same configuration, upon access, Differences in results can be significant in terms of measurement and attribution,” says Paul Mazur, a former trader recently hired as Affiliate Sales Director at Rakuten Advertising.

For this affiliation specialist, while it is essential for advertisers to have a comprehensive and accurate view of the contribution of each marketing lever (affiliation, view, search, etc.), only third-party, independent referral tools can provide fully impartial answers. “Google’s judgment, party, and criteria for its registration are not objective: there are necessarily some judgments about what Google considers most important,” said Paul Mazur. For his part, Guillaume Bournizien advocates for clearer weighting criteria: “It is very normal for advertisers to have a model that takes into account all the leverages that may have contributed to the conversion, but it is necessary to find a fair solution to that balance between all these contributions and the clear criteria for this weighting. : It is important to attribute the right selling to the right channels, but we must not forget about performance, because that is what starts selling.

“The market no longer has the same deterministic tracking capabilities as it once did.”

But the problem may not be there: For Guilhem Bodin, the real reason behind these scaling differences lies elsewhere, in the loss of speed of third-party cookies. “Consent requirement, restrictions on third-party cookies by different browsers and the potential for them to end up in Chrome mean the market no longer has the same deterministic tracking capabilities as it once did.” The result: The dearth of really high-quality cookies will result in a blurry reading of the Internet user’s journey across all attribution tools and on Google Analytics. Inaccurate evaluation that would affect the decision-making process of advertisers, including in relation to affiliation.

This specialist pushes this analysis further to point out that it is the very affiliation model, as well as the logic at the heart of digital referral tools, which is severely challenged by the incremental end of third-party cookies. “Tracking still relies heavily on third-party cookies and alternatives under consideration for their replacement, such as server-to-server linking, referral via identifiers, etc., are not yet sufficient to meet all needs.”

What do you do next? “There is no miracle solution, but one way to analyze affiliation at its fair value in a world without third-party cookies is to do it via AB tests which we can then extrapolate but considering their three different big journeys separately – higher and he concludes That the funnel affiliates (influencers), the middle channel (content sites) and the bottom of the funnel (coupons, coupons, etc.). We mix it all up.”

Leave a Comment