Attribution models

You can use attribution to build a report on traffic sources in different ways. This setting helps you pinpoint the traffic source and find out which of them resulted in a conversion. In other words, it helps find the most effective sources.

When making a report, you can use information about multiple sources for each user:
  • First click.
  • Last click.
  • Last non-direct click.
  • Last click on Yandex.Direct ads.

For the “First click”, “Last non-direct click”, and “Last Yandex.Direct click” models, Yandex.Metrica takes information from the user's session history.

Let's look at an example:

A user clicked through to the site from search results, viewed a few pages on the site, and left. The user returned to the site later, but this time from a Yandex.Direct contextual ad. Then later on from a social network ad. Some time later, the user entered the site's address into the browser and completed an order (this is the moment of conversion). As a result, the user had four separate sessions on the site.



To pinpoint the referral source, we must use an attribution model. Sessions are distributed differently by source depending on the model.

Distribution of sessions by traffic source based on the attribution model
Session / Model Last click Last non-direct click Last Yandex.Direct click First click
1 Search Search Search Search
2 Yandex.Direct Yandex.Direct Yandex.Direct Search
3 Ads Ads Yandex.Direct Search
4 Direct traffic (hit) Ads Yandex.Direct Search
Last click

In this model, Yandex.Metrica detects the current traffic source for each session without referencing the user's session history. In the example, the user had four sessions, each with its own source:

The model can be used for a technical analysis of your site (for example, to detect pages that don't have any code snippets using internal traffic analysis).

Last non-direct click

This model lets you calculate the conversion rate more precisely. All sources can be roughly classified as non-direct or direct in terms of conversion rate.

In this case, sessions from direct sources are attributed to a previous non-direct source, allowing you to better measure their performance.

In our example, the sources of the first three sessions are non-direct (search, Yandex.Direct, and ads). The source of the last session is direct. Therefore, this session is attributed to the third non-direct source (ad traffic):

This attribution model gives you reliable results for sites with fast conversion that occurs during a single session.

Last Yandex.Direct click

This model helps you evaluate the performance of Yandex.Direct ads regardless of when in the session chain a user clicked through an ad to your site.

If click-throughs from other sources were registered after the click-through from the Yandex.Direct ad, all the sessions are attributed to the Yandex.Direct ad traffic.



If the user clicked through Yandex.Direct ads multiple times, the session will be attributed to the last click.



The model is used in Yandex.Direct reports by default. If there's no traffic from Yandex.Direct ads, the Sources, summary report is based on the “Last non-direct click” model. You can change the attribution model.

First click

This model is used for sites with deferred conversion: when it takes a user a long time to make a decision about a purchase (or another goal) and they may go back to the site from different traffic sources several times while thinking it over. You can also use this model if you need to know which source attracts the most new users to a site.

This model uses the user's history: the source of the very first session. All of the user's subsequent sessions are attributed to the first source. In the example, the user's first session started from search results: