Know what the report describes
Google’s Performance report documentation covers clicks, impressions, click-through rate and average position, with grouping by queries, pages, countries, devices and dates. These describe search-result performance. The workflow below is an independent research-based method for investigating a collection page; it does not predict ranking improvements or replace sales reporting.
Write the question narrowly: for example, why a particular collection received fewer search visits during a completed period. Save the exact page address and the date the concern was raised. Broad questions about whether SEO is working tend to mix several unrelated changes.
Choose comparable periods
Compare complete periods of similar length and note trading events, stock changes and major site releases. A partial week should not be casually compared with a full week. If the store has seasonal demand, include a separate view that helps the team consider that context.
Record the selected property and filters beside the result. Another colleague should be able to reproduce the view without guessing. A saved screenshot without dates or scope is weak evidence because a later reviewer cannot tell whether the apparent movement came from the page, the audience or the reporting selection.
Separate visibility from response
Look at the direction of clicks and impressions together. Then inspect the page’s query mix before deciding what an aggregate rate means. A collection appearing for a broader set of searches may attract a different audience from the one it reached previously.
Group queries by shopper intent using a short, explicit classification: product type, use case, brand or informational research. Review ambiguous terms manually. This is an editorial analysis method, not a claim that a tool can perfectly infer what an individual searcher wanted.
Inspect the page as a shopper
Open the collection on a phone and a larger screen. Check the visible title, introductory copy, product selection, availability and route to useful filters. Ask whether the page answers the intent represented by the query group being discussed.
Compare what changed in the catalog with the change log. A collection dominated by unavailable items poses a different problem from a page whose wording no longer matches its range. Record a concrete observation, such as an absent size filter, instead of a vague instruction to improve the content.
Choose one explainable intervention
Prioritize a change that addresses the observed mismatch and can be reviewed later. Write down the hypothesis, affected URLs, responsible person and intended review date. Avoid changing navigation, copy, product membership and page addresses together unless a larger release genuinely requires it.
Keep an unchanged comparison page where it makes practical sense, while recognizing that pages can face different demand. Describe the limits of the comparison. Search performance has multiple influences, so an improvement after editing cannot on its own prove that the edit caused the improvement.
Leave a useful investigation record
Close the review with the evidence, the most plausible explanation and the uncertainty that remains. If the available data does not distinguish two explanations, say so and name the next check. A decision to gather better evidence can be more useful than a hurried rewrite.
At the agreed review date, use the same scope and record any intervening releases or promotions. Keep search observations separate from order outcomes, then connect them using appropriate store reporting. This produces a repeatable operating habit instead of a dashboard discussion driven by whichever line moved most recently.
Official reference: Google Search Console: Performance report overview. Source checked on 29 October 2026. AMZ12 analysis and suggested checks are independent editorial guidance.