CTR by Google position, measured on one site instead of averaged across many
Every CTR-by-position table you have found is an average across categories that behave nothing like each other, and none of them is your site.
Across roughly 1.9 million non-brand impressions from one site, CTR falls from 11.82% at position 3 to 4, to 0.81% at position 8 to 10. That is a 15x collapse across five positions, and it is the reason a page stuck on the second half of page one earns almost nothing however good it is.
The table
Non-brand queries only, over 267 days. Brand queries are excluded because they ran at 31.55% CTR against 1.94% for everything else, and leaving them in inflates the top bands beyond usefulness.
| Position band | Queries | Impressions | Clicks | CTR |
|---|---|---|---|---|
| 1 to 2 | 7 | 522 | 143 | 27.39% |
| 2 to 3 | 36 | 5,208 | 1,606 | 30.84% |
| 3 to 4 | 108 | 67,037 | 7,926 | 11.82% |
| 4 to 5 | 86 | 24,136 | 2,194 | 9.09% |
| 5 to 6 | 180 | 188,902 | 7,025 | 3.72% |
| 6 to 8 | 222 | 565,775 | 7,804 | 1.38% |
| 8 to 10 | 188 | 937,691 | 7,569 | 0.81% |
| 10 to 15 | 93 | 121,952 | 2,391 | 1.96% |
| 15 to 30 | 25 | 22,906 | 741 | 3.23% |
Two rows you should not trust, and I am printing them anyway
Position 1 to 2 rests on 522 impressions across seven queries. At that sample size the 27.39% could be almost anything, and the honest thing is to say so rather than quietly drop the row. The 10 to 15 and 15 to 30 bands rise again, which is not a discovery about Google: it is a small number of unusual queries where the searcher scrolls because the top results do not answer them.
Where the impressions actually live
The 8 to 10 band carries 937,691 impressions, more than any other row, and returns 7,569 clicks. Most of what a young site earns in visibility lands exactly where visibility is worth least. That single fact decides roadmaps: another page at position 8 adds impressions, and moving one page from 8 to 4 adds traffic.
How to build this for your own site
Export your queries with clicks, impressions and average position, remove every query containing your brand or a misspelling of it, bucket by position band, then sum clicks and impressions per band and divide. Do not average the per-query CTRs, which weights a nine-impression query the same as a nine-thousand-impression one. Appendix D of the book walks through it with the export open.
What it is worth to move a page
Multiply the impressions a page already earns by the CTR of the band you are trying to reach, then compare that with what it earns now. On this data, moving a page from the 8 to 10 band into the 3 to 4 band multiplies its clicks by roughly fifteen, with no change in how many people see it. That arithmetic, on your own numbers, is the only honest way to decide whether a position is worth the quarter it takes to reach.
Cite this
Mailarpawar, P. (2026). CTR by Google position, measured on one site instead of averaged across many. RankHonestly. Measured on 267 days of Google Search Console data. https://rankhonestly.com/case-study/ctr-by-position
These figures are free to quote, republish and chart, with a link back.
Use the numbers
- Run this on your own export
- /guides/average-ctr-by-google-position
- /guides/what-is-a-good-click-through-rate-seo
- /guides/impressions-but-no-clicks
More from the same dataset
- Programmatic SEO, measured: 11 of 94 pages earned 88.6% of the clicks
- The first 90 days of a new site, month by month, with the work shipped beside it
The whole record
This page is one table out of it. The book is all of it, for $19 once. What is in the book, or read Chapter 20, free to read.