WEBSITE PLANNING

Compare Search Results Across Equivalent Date Windows

Build a repeatable comparison with matching durations, filters and reporting completeness.

Write down the windows before reading the trend

Label both start and end dates, the property, search type and filters. Comparing seven days with thirty days by total impressions tells you mostly that the windows differ. For a recurring review, compare equal-length periods and explain holidays, launch dates or business closures that may affect interpretation. Do not quietly change the window to make a result look better.

Exclude incomplete dates deliberately

Google identifies recent performance data as preliminary when it is still collecting. Use a shared complete cutoff for the comparison, and record unavailable dates separately from observed zero values. If one export stops earlier than another, either align the windows or clearly describe the mismatch before calculating a percentage.

Use a hypothetical review sheet

Imagine reviewing a redesign launched midweek. Create a baseline row for the agreed prelaunch period, a deployment row and a postlaunch row beginning after the change. Record impressions, clicks, CTR and position alongside relevant annotations. Do not attribute the first day’s difference to the redesign merely because it occurred afterward. The example establishes a review method, not proof of causation.

Preserve a reproducible result

Save the aggregate export and filter settings with the review date. Add a short interpretation separating observation from hypothesis: clicks increased; the reason remains uncertain. If samples are small, show counts alongside percentages. Schedule the next comparison after enough complete dates accumulate rather than reacting to each partial daily update.

Apply this to your website

Bring the relevant URL and a description of the decision to request a scoped website review. Do not send passwords or private customer records. The primary reference is the linked technical guidance.