Calculate true CTR curves adjusted for AI Overviews, Natural Language, and GSC Deep-Linking.
Recoverable traffic from Page 1 anomalies only.
How your actual Page 1 performance segments into standard versus AIO-impacted environments.
| Target Query & Grouping | Position | Imp. | Actual CTR | Expected Benchmark | Missed Clicks (Deficit) |
|---|---|---|---|---|---|
| Upload a GSC CSV to build your custom curves and command center. | |||||
This tool is free to use
If it saved you some time, consider buying me a coffee to support future updates.
Buy me a coffeeGeneric CTR-by-position benchmarks (the "10% at position 3" style figures quoted everywhere) are averaged across every industry and query type, which makes them a poor yardstick for any specific site. This tool builds your own CTR curve directly from your GSC data — separately for standard queries versus queries likely impacted by AI Overviews or written conversationally — then flags every query getting meaningfully fewer clicks than your own benchmark predicts for its position.
Queries are split into two buckets: "AIO Risk" (matching common informational modifiers like "how," "what," "vs," "best") and "Conversational" (long natural-language queries above a word-count threshold you set), both of which typically see suppressed CTR when an AI Overview or featured answer sits above the organic results. Anything flagged Underperforming against its own bucket's benchmark is a real title/meta rewrite candidate — not just a query that happens to have a low raw CTR.
Yes, completely free with no account or usage limit.
For each position 1 through 10, the tool sums actual clicks and impressions across all your (filtered) queries at that position, separately for the standard bucket and the AIO/conversational bucket, and divides clicks by impressions to get a real average CTR for that position on your site specifically. If an AIO-bucket position has too little data (under 50 impressions) to trust, it falls back to 39% of the standard curve's CTR at that position, a rough estimate of typical AI Overview CTR suppression rather than a precise figure.
AIO Risk matches a customisable list of informational modifiers (how, what, why, guide, vs, best, and similar) that commonly trigger AI Overviews or featured snippets. Conversational applies to any query at or above the word-count threshold you set (8 words by default) — long natural-language queries, which are also more likely to trigger an AI-generated answer above the organic listings.
The Min Impressions filter normally excludes low-volume queries as noise. Turning this on exempts long-tail conversational queries from that filter, since they're individually low-volume by nature but collectively can represent a meaningful and fast-growing share of search traffic worth analysing even at low per-query impressions.
No. The CSV is parsed and all benchmark calculations happen entirely in your browser using PapaParse — nothing is sent to a server.
Underperforming means the query ranks on page 1 (positions 1–10) but gets a lower CTR than your own benchmark predicts for that position — a title/meta problem. Page 2+ means the query isn't ranking well enough yet for CTR to be the bottleneck; no CTR benchmark is calculated for it, since the fix there is rankings, not the snippet.