When to use them
Use experiments when a focused test is easier than the full dashboard.
The normal dashboard is broad. It tracks the website as a whole. An experiment is useful when you want a smaller view for one question group, one page cluster, or one planned routine.
Use experiments for isolated changes
Good experiments start with one clear change.
Examples:
| Test | Why an experiment helps |
|---|---|
| New comparison page | Watch only prompts tied to that comparison |
| New product page cluster | Track questions linked to that product area |
| Source-improvement push | Review sources for one topic group |
| Competitor response | Follow prompts where one competitor appears |
Pick one reason
Start from one gap or one content push.
Add matching questions
Use only prompts that can show the result.
Schedule follow-up
Let CLEA review the test after data exists.
Stop at the decision
Keep, adjust, or close the test.
Baseline must exist first
Do not start a test with no previous data.
If there are no prompt results yet, you cannot compare later movement. Wait for a few useful runs, or use the experiment as a planning space until enough data exists.
Main dashboard is better for normal monitoring
Use the main dashboard when you want the full website picture.
Experiments are for smaller work. They should help focus attention, not replace the core dashboard.