How to review a change in search performance.

Confirm the observation

Check the date range, report settings, comparison period, and whether the underlying measurement changed. Separate a broad traffic change from a change limited to a country, device, page group, or query.

Write down the observation in neutral terms. This keeps the investigation from becoming a search for evidence that supports the first explanation someone suggested.

Review the affected pages

Look for a pattern in the pages that changed. Are they products, collections, editorial content, or recently edited sections? Check whether the pages still exist, can be reached, and contain the intended content.

Review customer-facing quality alongside technical details. A migration, product removal, navigation update, or content change may affect different parts of the site in different ways.

Connect the review to the change log

Gather relevant website, catalog, content, and tracking changes. Review known search-system announcements using primary sources if they are relevant to the period, while avoiding an automatic assumption that an external update explains everything.

A timeline helps distinguish simultaneous events from demonstrated causes. Mark any explanation that remains a hypothesis and identify what evidence would strengthen or weaken it.

Choose proportionate next actions

Resolve clear defects first. Broader content or architecture changes should have a reason and a review plan. Changing many unrelated things at once can make the original issue harder to understand.

Document the findings and communicate uncertainty plainly. A useful review provides a practical next step even when the available evidence does not support a single definitive cause.