Changing a title is easy; knowing whether it helped is harder. Search traffic also changes with position, demand, devices, and the mix of queries. A test log makes the hypothesis and its limitations visible so a small movement does not become an unsupported success claim.
Write the hypothesis before editing
Specify the mismatch you intend to fix. Perhaps the page answers a cost question but the title hides the cost focus. Save the old title and description, the proposed change, the date, and the affected URL. Make sure the visible article actually delivers the new promise. A more clickable claim that misrepresents the content is not an improvement.
Compare like with like where possible
Use comparable date ranges and inspect relevant queries, countries, and devices rather than only the sitewide total. Record impressions and position alongside clicks and CTR. If the page starts appearing for broader queries, CTR can fall while useful traffic grows. Conversely, a higher CTR on very few impressions can be too unstable to interpret.
Make a cautious decision
Keep, revise, or reverse the change based on the combined evidence and the accuracy of the message. Search engines may display text different from your supplied title or description, so the published edit is not proof that every searcher saw it. Avoid repeatedly rewriting the same page before there is enough information to understand what changed.
Decision worksheet
| Item | Record or check | Decision rule |
|---|---|---|
| Hypothesis | Make the answer clearer | State the reader problem |
| Change record | Old and new text with date | Keep a reversible history |
| Context | Queries, impressions, position, device | Explain other possible causes |
| Decision | Keep / revise / reverse / insufficient data | Do not force a winner |
Worked example — illustrative, not measured
Illustrative observation: clicks rise from one to three after an edit, but impressions also triple. That is not convincing evidence that the title improved CTR. Keep the counts visible, examine the query mix, and label the result inconclusive rather than reporting a dramatic percentage increase as proof.
Put this into practice
- Choose one page with a clear message mismatch.
- Save the baseline and exact text change.
- Wait for a comparable observation period.
- Evaluate accuracy, query mix, and counts before attributing causality.
A mistake to avoid
Do not interpret zero clicks on a handful of impressions as proof that a title is bad. Sparse data may not distinguish competing explanations.
Basis and limitations
Google’s Performance report includes clicks, impressions, CTR, and average position. Those metrics describe observed search performance, not a controlled experiment on your title. Google: Search performance report (checked October 22, 2026).
The workflow and illustrative example are editorial recommendations, not measured vendor performance.