CTR Benchmark: find results with too few clicks for their position

CTR Benchmark (the page is titled CTR Analysis) answers one question: given where a result ranks, is it clicked as often as it should be? It builds an expected click-through rate for every position from your own Search Console data, compares each query or page against it, and adds up the clicks the under-performers miss. A large gap at a good position nearly always points at the title, the description or the way the result looks.

How the expected CTR is built

AutoPSEO builds a CTR curve for your site from the query rows in the selected date range:

1. Queries with at least 10 impressions are grouped by their average position, rounded to a whole number from 1 to 20.
2. For each position, the expected CTR is the impression-weighted CTR of that group: total clicks divided by total impressions, so busy queries count for more than rare ones.
3. A position backed by fewer than 200 impressions is too thin to trust; it uses a published market benchmark curve instead. The Site CTR vs expected card tells you how many of the 20 positions came from your own data.
4. A row's expected CTR is interpolated between the two nearest whole positions, so position 3.4 sits between your values for 3 and 4.

Because the curve is your own, a site whose results usually carry rich results, or whose brand searches dominate, is judged against itself rather than an average of other websites.

How each row is rated

Rows need at least Min impressions (10 by default; change it and the report reruns) and an average position of 20 or better.

• vs Expected shows the gap in percentage points; hover it for the relative difference.
• Under-performers are more than 20% below expected in relative terms, Over-performers more than 20% above, and everything else is In line.
• Potential Clicks = impressions × (expected − actual CTR) for under-performers, the clicks you would have had at a normal rate for that position.

Switch between Queries and Pages to rate searches or URLs against the same curve. The table opens sorted by potential clicks, with tabs for each rating, and you can save interesting queries to your keyword library from it.

Reading the charts

• Average CTR by position: bars for your CTR at each position and a line for the expected curve. Tick Show market benchmark to overlay the benchmark curve as a dashed line.
• Position vs CTR: one dot per query (the 300 with the most impressions), red for under-performers and green for over-performers, with the expected curve drawn through them.

What a low CTR usually means

• The title or description does not match the search. Google may rewrite titles and pick its own snippet text when yours describes the page poorly; see its guidance on title links and snippets.
• The results page is crowded. Ads, a featured snippet or other features above you take clicks at any position. Look at the live results before rewriting.
• The intent is different. A query that wants a tool or a video will not click a long article, however good the title.
• Position is an average. Google reports the topmost position your site held each time it appeared, averaged over impressions, so a query that swings between 2 and 12 can look like position 6 with a CTR that fits neither.

Over-performers are worth a look too: they show the titles and formats your audience responds to.

How AutoPSEO helps

The content engine applies the same test to the articles it manages, once a week: an article at positions 8 to 20 with at least 30 impressions and a CTR below 60% of the usual rate for its position gets an optimisation task to sharpen its title, meta description and direct answer. See Keeping articles fresh. The ctr_benchmark tool of the MCP server gives your AI assistant the same table.

Related: Page Poaching, CTR Curve Modeling, what is a good CTR.