Metrica Update Default Last Attribution Model, Not Everyone Is Happy

Yandex has updated the default attribution model used within the Metrica analytics platform, and not everyone is happy about the change.

Before, Metrica used the Last Transition model by default, but this has now been replaced by the Last Significant Transition model. More information on the difference between the models can be found via this help page, but a top-level explanation of the two can be found below.

Last Transition Model (Old Default)

 

In this model, for each visit, Metric will determine the source of the transition at the moment, without taking into account the history of the visitor’s visits. In the example, the visitor has four visits, each of them has its own source.

Last Significant Transition Model (New Default)

 

This model allows you to more accurately calculate the conversion. All sources can be conditionally divided into significant and insignificant in terms of conversion.

 

As Yandex put it, this means that “insignificant” sources are attributed to a more significant source. For example, before the change [direct] would have been attributed with the conversion:

Organic > Paid > Direct > Direct

But under the new default attribution model, the Last Significant Transition would be [paid]

Organic > Paid > Direct > Direct

Yandex has made the change to make data “easier” for attributing website promotion efforts, versus technical analysis (the previous method could be used to identify pages missing the Metrica Counter Code. You can, however, do this using third-party scraping tools, such as Screaming Frog.

It’s likely that you will now see fewer conversions coming from internal & direct as a source, especially if the user has previous sessions from other sources in the past.

Whilst you can choose to change the attribution model back to Last Transition, not everyone is happy about the change – with comments on the Yandex blog giving candid feedback that Yandex has made the change to benefit webmasters and analysts for large websites, who may invest more resources in things like SEO or paid traffic (via Yandex.Direct).

This is not convenient for everyone. You have done well for commercial sites that have a lot of attracted or purchased traffic. The model “Last significant transition” and the truth will bring more understanding. But this is not true for all sites. It is better if we were able to set in the settings which attribution model to use by default. Now we have added extra action.

&

It’s like that. Emphasis is shifted to those who pay to attract an audience.

 

Dan Taylor
Dan Taylor is an experienced SEO consultant and has worked with brands and companies on optimizing for Russia (and Yandex) for a number of years. Winner of the inaugural 2018 TechSEO Boost competition, webmaster at HreflangChecker.com and Sloth.Cloud, and founder of RussianSearchNews.com.