Once your experiment is running, we recommend testing that it is working as intended. The approach you use will depend on whether the subject of the experiment is a Salesforce user or a Salesforce record. 

Either way, we recommend impersonating at least ten different users for each new experiment. Be sure to impersonate users who could potentially be included in the experiment—for example, users who fall within the assignment of a page builder experiment. 

Your team may have many additional QA/QC steps, but these are the minimum tasks we recommend completing for each new experiment you run:

1. Validate the correctness of variation experience

   • Ensure the variation experiences align with expectations (e.g. layout is as desired)

   • If changes are needed, change the experiment record status to Paused, modify the variation experiences, then reactivate the experiment

2. Validate the variation experience’s bucketing action

   • Ensure different variation experiences are successfully being served

   • The number of times each variation experience occurs may vary greatly. What’s important is to confirm that each active variation experience is served at least once during your validation.

   • If the experiment subject is set to Salesforce User, the variation experience should change periodically as you switch between users. You may need to impersonate several different users before a different variation experience is served.

   • If the experiment subject is set to Salesforce Record, the variation experience should change periodically as you click on different Lead/Contact/Case records. You may need to click several different records before a different variation experience is served.

3. Review your Optimizely App Error Log

   • Identify any errors that may have occurred during your evaluation