How Thresholds affect Single-adgroup Testing
In a single-adgroup test, a threshold defines the minimum performance each Ad must have, or a statistical confidence a test must achieve before a valid test result is produced.
- Daily automated single-adgroup tests: the thresholds are used for all test runs. Account thresholds can be overridden at the campaign level.
- Manual single-adgroup tests: you can override the defined thresholds for every manual test you run.
How Thresholds affect Multi-adgroup Testing
In a multi-adgroup test, a threshold is the minimum performance each set of Ads (defined by patterns or labels) must have, or a statistical confidence a test must achieve, before a valid test result is produced.
- Daily automated multi-adgroup tests: the thresholds are used for all test runs. Account thresholds can be overridden at the campaign level.
- Manual multi-adgroup tests: you can override the defined thresholds for every manual test you run (as shown above).
How Thresholds affect the Metrics being Tested
A threshold is used only when it is relevant to the metric being calculated:
- Clicks threshold: used when calculating the CTR and Conv. Rate metrics only.
- Impressions threshold: used when calculating the CTR, Conv. /Impr. and Conv. Value / Impr. metrics only.
- Conversions threshold: used when calculating all metrics, except CTR.
- Confidence and Date Range thresholds: used when calculating all metrics.
What Threshold values should I use?
The Minimum Data You Should Define by Testing Metric
How Much Data Should You Have Before Examining an Ad Test Result?
Where are the Thresholds Defined?
4. Related Info
- How Much Data Should You Have Before Examining an Ad Test Result?
- Working with Statistical Significance: How confident should you be in your test results?
- Do You Have Insignificant Test Results? How to use Maximum Data in Ad Testing