Finding 01
Three query clusters drew 1.37 million impressions at 0.17 percent click-through
Three query clusters averaging positions between 5.6 and 8.6 drew 1,366,114 impressions over 28 days and returned 0.17 percent click-through between them. One window shows the rate, not what caused it.
- Published
- Window
- 28 days, 2026-08-13 to 2026-09-09
- Sample
- 3 clusters over 24 canonical URLs, 1,366,114 impressions producing 2,374 clicks
- Method
- Google Search Console pages export, one pull. Each cluster is a group of canonical URLs from the same export and window: anchor rows removed, then www and trailing-slash variants merged, the counting rule of findings 02 and 09. Average position is impression-weighted across each cluster's URLs.
Impressions are easy to read as demand waiting to be clicked. Three clusters on one property show how little of that reach became clicks over one month.
What we saw
Over one twenty eight day window, three clusters on one property averaged positions between 5.6 and 8.6. Between them they drew 1,366,114 impressions and returned 2,374 clicks, which is 0.17 percent.
| Cluster | URLs | Impressions | CTR | Position |
|---|---|---|---|---|
| Bank branch codes | 17 | 810,421 | 0.221% | 7.90 |
| Festival and observance dates | 5 | 367,983 | 0.068% | 8.61 |
| Calorie lookups | 2 | 187,710 | 0.176% | 5.62 |
One pull, one window, one dimension, counted on canonical URLs: anchor rows removed, then www and trailing-slash variants merged. Positions are weighted by impressions across each cluster's URLs rather than averaged flat, so a large page cannot be outvoted by a small one.
Why these three can be added together
Every row is a group of canonical URLs, from one export, over one window, so the total is a real total. Two things would break that. Clusters measured in different windows cannot be summed, and a search-appearance row cannot sit among them: it describes the same impressions from another angle, so adding it would count impressions already inside the other rows.
That is why the auto-translated surface is not in this table. It stands on its own as a separate finding, which is where it belongs.
A likely mechanism, not measured here
A query that resolves to one string or one number can be answered in the result itself. A bank branch code is a single value, a festival date a single date, a calorie count a single number. That would explain a low click-through rate, and finding 02 makes that case across the whole site. This finding does not measure it: nothing here records what the searcher saw on the results page.
What this does not show
It does not show that ranking higher lost these clusters clicks, or that the usual remedies cannot help them. Either would need the same clusters measured before and after a change, and this is one window.
It does show why the site-wide rate misleads. Across all 382 canonical URLs over this window the property returned 0.465 percent on 2,079,623 impressions, which reads as a modest click-through rate. These three clusters carry 1,366,114 of those impressions at 0.17 percent, so that average blends families that click at very different rates.
How to check your own
Export Search Console by page for one window. Remove anchor rows, merge www and trailing-slash variants, and group the URLs into query families. Compare click-through by family rather than in aggregate, beside each family's impression-weighted average position. A family with large reach and a click-through rate far below the rest of the site is the one to look at more closely, starting with what the results page already shows for its queries.