Skip to main content
All CollectionsBlueprintsMonitoring and troubleshooting
A/B test performance with Blueprint trials
A/B test performance with Blueprint trials

Test and compare performance changes against a control, helping optimize strategies effectively.

Updated over a week ago

Blueprint trials offer a way to test new strategies and measure performance against a control. Users can experiment with various versions of Blueprints to identify the one that yields the best results and easily implement changes that improve performance.


Setting up a Blueprint trial

  1. Go to Blueprints in the top menu.

  2. Click on a Blueprint in the list.

  3. Select Create Trial at the top of the Blueprint, located beneath the metrics row.

  4. (Optional) Select accounts to link them to the trial Blueprint and trigger a republish. Tip: We recommend selecting accounts after configuring the trial variables.

  5. Click Create Trial. The duplicated Blueprint for the trial will be marked with a beaker icon and the name appended with "TRIAL." This will initiate an A/B test with your selected accounts and the control accounts.

  6. In the trial Blueprint, implement the changes you want to test. Tip: We suggest limiting the variables being tested so you know which changes impact performance. Ideas include shared vs. single budgets, creative variations/ad copy, geo strategies, and bidding strategies.

  7. When finished, click Review/Edit Trial.

  8. Under Move Additional Accounts to Trial, select the accounts to test the changes on. Note: Assigned accounts will launch on the trial Blueprint and pause in the control Blueprint.

πŸ›‘ Caution: Avoid making changes to the original Blueprint once a trial is running, as it can impact control metrics. Any changes made will be reverted once the trial is complete.


Reviewing trial results

  1. Navigate to the trial Blueprint, and click Review/Edit Trial.

  2. Compare the trial Blueprint with the control Blueprint. The metrics reflect the performance of both the control and trial Blueprints over the last 30 days.

    • Control shows the metrics for all accounts using the original Blueprint and does not include the accounts assigned to the trial.

    • Trial shows the metrics for all accounts assigned to the trial.


Concluding a trial

  • Option #1: Apply successful changes. If the trial Blueprint's changes resulted in better performance compared to the control Blueprint, click "Promote This Trial." The trial will conclude, and the successful changes in the trial Blueprint will be applied to the control Blueprint.

  • Option #2: Reject changes. If you choose not to implement the changes, click "Delete Blueprint Trial." This will end the trial, and no changes will be made to the control Blueprint. Accounts assigned to the trial will resume running on the control Blueprint.

Did this answer your question?