To understand the impact of bot visits on a Metrica report, first compare the same period with different breakdowns. Do not change data collection just for a prettier conversion percentage: filtering at collection and temporarily viewing a segment have different consequences.
What exactly the system determines
Metrica distinguishes known bot User-Agents and behavioral bot detection. These signals do not mean that every automated visit is necessarily recognized. Metrica's metrics and those of another advertising system also cannot be considered identical based on one similar name.
Record the current counter settings, the goal, and the report period. Then compare the available options for displaying bot traffic while keeping all other conditions the same. If you change the source and the date at the same time, the cause of the difference will be lost.
How to read the result
Suppose that after selecting a breakdown the number of visits changed, while confirmed orders in your system remained the same. The percentage could have changed because of the denominator. This does not mean that sales grew at the moment the report was switched.
Look at where the difference is concentrated: source, landing page, device. For a suspicious area, formulate a testable hypothesis and compare it with the available technical data from the site. Do not declare a specific visitor a bot based only on a short visit.
When it comes to configuring collection
According to the help documentation, collection filters apply to new data; discarded information cannot simply be brought back with a report toggle. Therefore, before such a setting, you need to understand the consequences and record the date of the change. For an initial analysis, it is more reasonable to use the available comparison of breakdowns.
In the conclusion, state which dataset was selected and why. If some sources have not been checked, keep that limitation. Removing inconvenient visits from a report does not confirm the quality of a service, organic interest, or the return on advertising.
Related task: Your own visits and test actions: how not to mistake a site check for demand.
Sources
Metrica: bot visits · Metrica: collection filters. Verified 23 September 2026. Practical scenarios are an editorial methodology; the training examples are not client data.
If a separate public announcement for this work has been prepared on Telegram, you can view the terms for post views. The form below applies only to such an announcement. The service does not change filters and does not confirm the quality of a site's traffic. Purchased activity does not confirm organic interest, leads, or revenue.