Search Console Query and Page Evidence: Read the Scoped Record

Record access and report settings before interpreting numbers.

Begin with the question and the available evidence. Record the property, Search results report, search type, period, filters, grouping, collection date and export method. If you cannot access a property or export, mark that input unavailable; it is not a record of zero performance.

Google’s Performance report overview explains the basic metrics and dimension choices. A query describes a search term, a page describes the reported URL, country groups the search origin, and device separates the device category. A change of dimension changes the question being answered.

Keep the date convention as well as the dates. The dimensions documentation distinguishes Pacific Time for ordinary date views from local time in the 24-hour view. Note partial or preliminary periods rather than comparing them with completed periods as though collection were equivalent.

Use filters and groupings deliberately.

To inspect a query-to-page relationship, preserve the query criterion and review the associated page records under the same relevant country, device, date and search-type scope. A general query list does not itself show which page supplied every result.

Use the current filtering and comparison documentation for the method available in the report. Record the filters actually applied, including retained filters, rather than relying on a description of what you intended to select.

A country or device slice is not a whole-site interpretation. A query appearing for two pages does not by itself prove harmful overlap: investigate the task, page content, canonical relationships and period before deciding if the owners conflict. The same text in a title is even less evidence than a scoped query/page record.

Keep counts, ratios and averages attached to their units.

CTR is clicks divided by impressions for the same record. Average position summarizes reported positions; it is not a promise that every searcher saw one fixed live rank. Do not average row percentages or positions without a method that preserves the relevant weighting and scope.

Google’s data-counting documentation explains property and page aggregation. Different URLs from one property can be counted differently from one property-level observation. Keep the aggregation label on chart totals, page records and query extracts; do not force unlike totals to reconcile.

Most Search Console performance data is assigned to the canonical URL under the documented rules. A reported page address can differ from a duplicate address visited by a user. Use the appropriate current inspection evidence before treating that difference as a missing page or an ownership failure.

Worked sample: page totals and incomplete query rows.

Every query, URL and number below is invented for teaching. No Semantec or client Search Console property was accessed. The fictional scope is Web text-based Search, Ireland, mobile, 1–30 September 2026 in the ordinary daily report convention. Both example pages use the same stated scope.

The page-level teaching records are A at example.com/lessons/source-record/ with 100 impressions and 10 clicks, and B at example.com/lessons/source-dates/ with 50 impressions and 2 clicks. Their scoped CTRs are 10% and 4%. These are page-level records, not an available property-level chart total.

The supplied query/page extract contains only the four rows in the table. For A, disclosed rows total 80 impressions and 7 clicks; for B they total 40 impressions and 1 click. Relative to the stated page records, unresolved detail remains 20 impressions/3 clicks for A and 10 impressions/1 click for B. Those residuals do not reveal the missing search terms or the reason every row is absent.

For A, the two visible row CTRs are 10% and 5%. Their simple mean, 7.5%, is not the combined visible-row CTR: 7 ÷ 80 = 8.75%. The full supplied page record is 10 ÷ 100 = 10%. These three ratios have different calculation bases; none can stand in for the missing query detail.

The query “source record” appears for both pages in the extract. That observation supports further review of their tasks and records; it does not establish that either page should be merged, redirected or rewritten. No property-level total or causal explanation has been supplied.

Name missing detail without reconstructing private queries.

Google omits some queries for privacy and also limits the rows retained or shown. Its query limitations explain that anonymized queries can be included in totals but excluded when a query filter is applied. A residual cannot be labelled as one known hidden term or a verified list of anonymized searches.

In the example, label the unexplained difference “unavailable query detail in the supplied sample”. Do not invent a query name, distribute the residual across visible rows or convert an unavailable observation to a zero. Keep an omitted row distinct from a disclosed row showing zero clicks.

Check exported placeholders too. The data documentation notes that some unavailable/not-a-number display states can become zeros in downloaded data. Preserve the original state and source context before interpreting an exported zero as an observed count.

Compare like periods and keep interpretation separate.

Check the same property, search type, grouping and filters before comparing dates. Record complete versus partial collection, query mix, canonical changes and the method used. A change in average position or CTR can come with a changed set of observations; it is not automatically a change for one fixed audience and rank.

Keep the source record separate from the interpretation. A high-impression/low-click query might justify inspecting task fit and visible presentation, but it does not prove that a new title will improve outcomes. A position average cannot tell you every individual result or replace a current page/content check.

For the sample, request the permitted complete export or explain its limits, verify page/canonical identity and compare the two reader tasks. Do not make a claim about missing data recovery or personal query access that the available tools cannot support.

Preserve a reproducible, bounded review record.

Keep property and access state, dates/time convention, filters, dimensions, aggregation, source file, disclosed rows, missing states, formulas, interpretation and next action together. Note the responsible owner and the point at which changed evidence should reopen the review.

Return to the measurement and diagnosis overview or learning overview. Use coverage and distinct measurement denominators when a search statistic is being used as a substitute for answer completeness.

This lesson supplies an inspectable fictional sample and method. It does not verify a property, measure a real search result, recover anonymized queries or establish a conversion outcome.