FIELD GUIDE · REVIEWED OCTOBER 15, 2026

Compare Search Console date windows

A fair comparison uses equal-length periods and matching returned dimensions. Review the underlying metrics and coverage before treating a difference as a trend or a cause.

Choose dates that answer the question

Decide what you want to compare before reading the change. A week against the immediately preceding week can reveal a short-term movement; matching calendar periods may be more useful for a seasonal business. Use equal-length windows and check whether they include the same weekdays or known events. Search Console data for recent dates can be incomplete, and totals can move for reasons unrelated to a site change. Write down the exact start and end dates, property, search type, and dimensions so another person can reproduce the comparison.

Read the metrics and missing rows carefully

When comparing query or page rows, interpret clicks, impressions, CTR, and average position together. A change in one metric does not explain the others: fewer clicks can accompany more impressions, and a lower numerical average position generally indicates a better rank. Comparison tools match rows returned in both periods. If a dimension is absent from one result, its value is unknown for that comparison; it is not silently counted as zero or proof that the query disappeared. Google can omit anonymized and low-volume rows, and a bounded result can leave out additional rows. Check the report’s periods, row counts, truncation indication, and coverage notes before interpreting totals.

Worked example and next checks

Suppose a query has 100 clicks in the earlier window and 80 in the later window. The change is (80 − 100) ÷ 100 = −20%, or 20 fewer observed clicks. Suppose average position moves from 8 to 6. Because lower position numbers indicate a higher placement, that is an observed improvement of two positions. The two movements can coexist; neither proves that a particular edit caused the click decline. Check impressions, CTR, the landing page, indexing state, and whether the windows cover comparable days. Look for a matching change in the search results or business context, then compare again using the same dimensions and a suitable later window.

Try the workflow

Explore the interactive sample first. For an approved connected account, ask your authorized MCP client:

For sc-domain:example.com, use compare_search_periods for August 15–28, 2026, grouped by query. Show the previous equal window (August 1–14), matched-row changes and missing-row caveats. Repeat grouped by page if needed. Propose checks without assigning cause.

Relevant tool: compare_search_periods. See product documentation and connection settings.

What the evidence cannot tell you

Search Analytics rows are not guaranteed to include every query or page, especially anonymized and low-volume data. Results may be incomplete for recent dates or constrained by row limits. Matching returned dimensions makes the comparison more interpretable but cannot remove seasonality, changing demand, search-result changes, or other confounders. A before-and-after difference alone does not demonstrate that a site edit caused the change.

Sources and build scope

Original editorial examples, reviewed October 15, 2026. Google defines clicks, impressions and average position; its Search Analytics API reference documents grouping, inclusive Pacific Time dates, fractional CTR and bounded results.

Product behavior checked against Searchlight source 4f160c3, calculation version searchlight-insights-v2 (repository access required). This is implementation provenance, not proof of production availability or live-provider acceptance. Roadmap features are not described as shipped.

← All guides