Aside from adding the standard protective elements such as creating a Segment, Metrica, and Direct tokens (for Yandex), and adding Google ID (for Google Ads), we also provide additional and highly useful tools for safeguarding your advertising campaigns.
Behavioral Factors: You can find these in your personal account under “My Sites” / “Edit Site Section” / “Blocking Settings.”
Enable the ‘Analysis of Behavioral Factors’ option to block users (bots) who do not meet the criteria outlined below:
- Apply only to advertising sessions: When activated, other behavioral parameters will only assess visits through advertising and block them by IP address (in Yandex.Direct and Google Ads systems). If this option is disabled, all visits will be analyzed. All Yandex-related blocks will be directed to a segment in Yandex.Metrica.
- Minimum number of actions on the site: Sessions that don’t meet this specified indicator will have future ad displays blocked (if set to 0, the parameter is disregarded). For example, if a user clicked on an ad but only performed one action, the system will block future ad displays for them. Actions on the site include clicks, scrolls, and resizing the page, essentially any user interaction.
- Minimum session length in seconds before blocking: Sessions that fall short of this duration will be blocked (if set to 0, it’s not applied). For example, if a user clicked on an ad but spent less than 10 seconds on the site, the system will block them.
- Maximum number of visits per day: Users who exceed this number of daily visits to your site will be blocked. For instance, if a user clicked on an ad more than 3 times in a day, the system will block them. (if set to 0, the parameter is disregarded)
Note: If a user (or bot) meets any of the specified criteria, the system will block future ad displays for them.
Cluster Analysis: You can configure this parameter in the dashboard under “My Sites” / “Site Editing Section” / “Blocking Settings.”
Our system employs machine learning methods to analyze the traffic on your site. By default, this functionality is enabled but can be turned off or adjusted.
- If you don’t want machine learning cluster algorithms to analyze your site traffic, uncheck the ‘Cluster Analysis’ box.
- If you want only sessions that originate from advertising to undergo intelligent analysis (excluding all general sessions), check the ‘Only sessions through advertising’ box.
- There’s also an option to fine-tune the algorithm’s sensitivity if the current settings don’t suit your needs. The recommended sensitivity value is 6. You can lower or raise it within the range of 1 to 10. Higher values will result in more sessions being blocked.
How does cluster analysis work? It’s a machine learning algorithm that, based on a set of parameters (black box model), determines whether a bot has visited the site. Depending on the analysis data, it assigns the visit to one of six clusters (see the figure below). A value of 6 indicates the visit is placed in a cluster where 100% are bots. However, since bot algorithms vary and constantly evolve, we recommend increasing this value to 7-8 if you suspect bot visits on your site due to click fraud. Values above 6 consider data not just from the bot cluster but also from the average cluster, where ‘suspicious’ visits are identified. There is no 100% certainty in this cluster; it may be a well-masked bot or a real user exhibiting suspicious behavior.