What Is the GSC Performance Report?

Google Search Console's Performance report is the diagnostic core of the platform — and its four metrics are the most systematically misread numbers in organic search. Clicks, Impressions, CTR, and Average Position appear side by side, but they are not a flat list. They are a diagnostic chain, and reading any one of them correctly requires understanding its relationship to the others.

The chain runs in one direction. Impressions measure coverage: how often Google served one of your URLs in results. Clicks measure response: how many users acted on those appearances. CTR captures the ratio — Clicks divided by Impressions — which is a snippet quality signal, not a traffic signal. Average Position locates where those impressions appeared, which sets the realistic ceiling for any CTR benchmark comparison. A page averaging position 8 with a 2.1% CTR is not underperforming — that is within range for its position. A page averaging position 1 with a 2.1% CTR is a serious problem.

The practical diagnostic flow runs the same direction. Low impressions on a page you expect to rank point to an upstream problem: indexing, crawlability, or content relevance — see Index Coverage & Indexing in GSC before adjusting any content. High impressions with low clicks point to the snippet. Acceptable CTR with inadequate volume closes the case at Average Position: positions 6–10 collectively earn 1–4% CTR regardless of title quality, and no amount of meta-description optimization overcomes a page-2 position.

The 2025–2026 period introduced three compounding factors that require explicit context before interpreting any Performance data from this window. On September 11, 2025, Google deprecated the &num=100 URL parameter, eliminating bot-driven impressions for deep-result positions that had suppressed Average Position figures for years. AI Overviews — present on a growing share of informational queries — compressed organic CTR at every position: Ahrefs' February 2026 analysis of 300,000 keywords found position 1 CTR falling from 7.3% (December 2023 baseline) to 1.6% (December 2025) on AI Overview-triggered queries. And a GSC logging bug inflated impressions platform-wide from May 13, 2025 through April 27, 2026 — clicks were not affected, but CTR and Average Position derived from those inflated impression counts are unreliable for that window.

For a full orientation to GSC's report inventory and property verification, see What is Google Search Console. For the SEO measurement context — how Performance data relates to Google Analytics 4 session counts, rank tracker position data, and backlink signals — the Performance report is the one data source that cannot be replicated outside Google's own infrastructure.

Key Takeaways

• GSC's Performance report tracks four metrics — Clicks, Impressions, CTR, and Average Position — that form a diagnostic chain, not a flat list. Read them in sequence to locate where the problem is, not just that a problem exists.
• Average Position is an impression-weighted arithmetic mean across all users, geographies, device types, and personalization signals. It reads better than rank tracker output by design — both numbers are correct; they measure different things.
• On clean SERPs, position 1 earns 39.8% of clicks and position 2 earns 18.7% — a 21-percentage-point gap that is the largest single positional drop on the page (FirstPageSage, 2026). The top three positions together capture 68.7% of all clicks.
• Ahrefs' April 2025 analysis of 22 billion clicks across 887,534 GSC properties found anonymized queries account for 46.77% of all GSC traffic. The Queries table always shows a partial view — the chart totals always exceed the table row sums.
• The Search Analytics API returns up to 50,000 rows per day per property per search type (versus 1,000 in the interface) and includes hourly data for the last 10 days via the HOUR dimension.
• GSC retains 16 months of Performance data on a rolling basis. Data older than 16 months is permanently deleted; there is no retroactive recovery once it ages out.
• GSC impressions from May 13, 2025 through April 27, 2026 are permanently inflated due to a platform-wide logging bug. Clicks from that window are reliable; impression-derived metrics (CTR, Average Position) are not benchmarks for that period.

The Four Performance Metrics: Definitions, Counting Rules, and Diagnostic Use

GSC defines each metric precisely, and the counting rules contain edge cases that explain most of the data discrepancies that trip up practitioners. The table below draws directly from Google's authoritative documentation at support.google.com/webmasters/answer/7042828.

One implication of these counting rules is consistently missed: the CTR you calculate from the Queries table is never the true property-level CTR. Anonymized queries (those not issued by enough users to clear Google's privacy threshold) appear in the chart totals but are absent from the table rows. For a typical site, Ahrefs' April 2025 analysis found this gap accounts for 46.77% of all traffic — meaning table-level CTR calculations always overstate the ratio against the full impression pool. The chart header is the correct denominator for property-level CTR.

Organic CTR by SERP Position: The 2026 Benchmark Data

The most actionable use of the Performance report is comparing your CTR at each position against external benchmarks. If your position-1 pages are delivering 8% CTR and the clean-SERP baseline for position 1 is 39.8%, the gap is the diagnostic — not the position itself.

The clean-SERP figures above are the diagnostic baseline. In practice, CTR at any position varies by query type, device, SERP features present, and whether an AI Overview is triggered. The 39.8% position-1 figure applies to informational queries on uncluttered SERPs — branded navigational queries regularly exceed 60%, while queries with shopping modules, local packs, or AI Overviews can compress organic CTR well below these medians.

MB Adv Agency has found that clients with average positions between 4 and 6 show the highest CTR improvement potential relative to effort — small, precise revisions to titles and meta descriptions at those positions move the CTR needle more than any other single intervention. Pages at positions 1–3 with below-benchmark CTR require SERP feature diagnosis first; a low CTR at position 1 almost always indicates AI Overview competition or a mismatched query intent, not a weak title. For the paid-search equivalent — CTR, CPC, and impression-share interactions in Google Ads — see Google Ads Metrics & KPIs.

The Page Experience report — which tracks how Core Web Vitals performance correlates with organic position — sits alongside the Performance report in GSC. See Core Web Vitals & Page Experience for how CWV affects the position data you read in this report.

Search Types and Separate Performance Reports

The default Performance report view shows Web search data. GSC tracks five additional search surfaces — each with its own impression-counting rules and, for Discover and Google News, separate report sections entirely.

All click, impression, and position data is stored and reported separately per search type. A URL that appears in both web results and image results has independent data for each — switching the search type filter in the Performance report is not a view preference, it loads a different dataset. The Discover and Google News reports cannot be combined with the Web results report in the interface; export to CSV or Sheets is required to merge them. For product-feed and shopping context where Google Merchant Center image data intersects with Search, see Google Merchant Center.

Performance Report Dimensions

The Performance report segments data across five dimensions. Each dimension changes the aggregation scope — not just the visual grouping — so adding a filter changes which data pool the metrics are calculated against.

Combining dimensions (e.g., Page + Query or Page + Device) narrows the data pool to the intersection, which is where real diagnostic value lives. A page averaging position 3 overall that averages position 11 on mobile is a mobile experience or mobile-intent problem — visible only when the Device dimension is applied against the Page dimension. For Google Tag Manager users building custom GSC data layers, the dimension structure here maps directly to the Search Analytics API dimension parameters.

Filters, Regex, and Comparison Mode

The Performance report's filter layer transforms it from a snapshot into a diagnostic instrument. Regex filtering (launched April 2021), the branded query filter (March 2026), and comparison mode each serve distinct analytical purposes — and each has edge cases that affect how results should be read.

The regex filter is the most underused capability in the Performance report. A regex like ^(what|how|why|when|can) isolates informational-intent queries across the entire property — immediately exposing which pages answer questions (and whether those pages are capturing the traffic their intent warrants). Applying (?-i)brand_name in case-sensitive mode filters branded queries with precision that the AI-assisted branded filter cannot match for complex brand strings. For step-by-step guidance on building these filters, see GSC API & Automation where the regex syntax is documented alongside API-equivalent filter parameters.

The Search Analytics API: Beyond the 1,000-Row Cap

The Search Analytics API and the Performance report interface share the same underlying dataset — but the API removes the 1,000-row ceiling, adds hourly granularity, and returns anonymized queries as a filterable zero-length string. For any site with thousands of pages or queries, the interface is a directional sample; the API is the complete picture.

The Search Analytics API is free, subject to standard API quotas. The 50,000-row daily limit is per property per search type — a site pulling web + image + video data has up to 150,000 rows available per day. For any automated SEO reporting stack — whether Looker Studio pulling directly via the connector, a BigQuery pipeline via Looker Studio's BigQuery export, or a custom Python script — the API is not a developer-only tool. It is the minimum viable data source for complete reporting at meaningful scale. For full API integration patterns, BigQuery setup, and the hourly data workflow, see GSC API & Automation.

MB Adv Agency has found that the Search Analytics API, when piped into a BigQuery dataset with rolling daily snapshots, produces the most complete picture of organic performance available — more reliable than rank tracker data for portfolio-level trend analysis because it reflects real user impression distribution rather than single-point crawls. For conversion-side data that complements GSC's click data, see Conversion Tracking & Attribution in Google Ads.

Diagnosing Performance Report Anomalies

Most Performance report anomalies have a pattern. The table below maps the most common symptoms to their likely causes and the first dimension to check — before drawing conclusions from the top-line numbers alone.

One pattern that cuts across all of these scenarios: when multiple anomalies appear simultaneously (position stable, impressions flat, clicks falling, CTR falling), the cause is almost always a SERP feature change — not a content problem. AI Overviews, Featured Snippets, and Knowledge Panels all absorb clicks from results that are still technically ranking. The Search Appearance dimension isolates which feature is involved. For security and manual action signals that affect ranking independently of content quality, see Links, Manual Actions & Security in GSC.

Data Freshness, the 16-Month Window, and the 2025–2026 Logging Bug

GSC Performance data arrives with a 2–6 hour delay under normal conditions and is retained for exactly 16 months on a rolling basis. Both constraints have operational implications that matter for any reporting architecture built on GSC as a primary data source.

The 16-month hard limit means year-over-year comparison — which requires at least 24 months — is impossible from the GSC interface alone unless data was previously exported. Any site that did not run regular API exports or Looker Studio BigQuery connections before June 2025 has permanently lost the data needed to compare June 2026 against June 2025. The correct architecture for long-term GSC reporting: daily API pulls into an external store from the first day a property is verified.

The May 2025–April 2026 logging bug — what it affects and what it does not. A GSC logging error inflated impressions platform-wide from May 13, 2025 through April 27, 2026. Key facts per Search Engine Land's original report: clicks were not affected; impressions were over-reported; CTR (which is derived from impressions) was artificially suppressed; Average Position (computed per impression) was skewed. Historical data was not retroactively corrected — those impression figures are permanently inflated in GSC. Any site that saw an impressions drop in late April or early May 2026 saw the fix take effect, not a ranking loss. Benchmark any impression-derived metric from that window with explicit caveats. For current data anomaly status, check Google's data anomalies documentation.

Methodology

All data in this article is current as of June 2026. GSC metric definitions, counting rules, impression counting behavior, API parameters, row limits, and feature launch dates are sourced from official Google documentation: support.google.com/webmasters and developers.google.com/search. CTR-by-position benchmarks are sourced from First Page Sage, "Google Click-Through Rates by Ranking Position in 2026" (firstpagesage.com), and Advanced Web Ranking, Q4 2025 CTR Changes Report (advancedwebranking.com). AI Overview CTR impact data is sourced from Ahrefs, "AI Overviews Reduce Clicks by 58%" (February 2026, 300,000 keywords, aggregated GSC data). The 46.77% anonymized-query figure is sourced from Ahrefs, "Anonymized Queries Make Up Nearly Half of GSC Traffic" (April 2025, 22 billion clicks, 887,534 properties). The &num=100 deprecation date (September 11, 2025) is sourced from Intero Digital and Ice Nine Online. The GSC impressions logging bug (May 13, 2025 – April 27, 2026) is sourced from Search Engine Land. Reviewed by MB Adv Agency, June 2026.