Ship, adjust, or abandon
Every experiment should end with a decision.
The decision can be simple: keep the change, adjust it, or stop it. The important part is that the experiment does not stay open forever with no next step.
Experiment decision
The result should end with one plain decision: keep the change, adjust it, or stop it.
Ship, adjust, or abandon result lens
Use the result to choose one path.
| Decision | Meaning |
|---|---|
| Ship | Keep the change and make it part of the normal workflow |
| Adjust | The signal is partial, so improve the page, prompt, or routine |
| Abandon | The test did not help or cannot be read clearly |
Read the goal
Return to the reason the test exists.
Check the signals
Review prompts, sources, competitors, and answer text.
Choose one path
Ship, adjust, or abandon.
Record the note
Save the decision for future work.
Evidence comparison
The decision should mention evidence.
Do not only say "ship it." Say what moved. For example: "Ship the page update because source reads improved in three experiment questions and answer wording is clearer."
Movement that matters
Movement should match the goal.
If the goal was to improve comparison prompts, do not ship only because an unrelated prompt changed.
Decision
Close the loop.
If you ship, move the lesson into normal work. If you adjust, create the next task. If you abandon, write why so the same weak test is not repeated.