How is average position actually calculated?
It is the average of “the highest position your link reached in the results”, not a fixed rank on the results page. Google's help page (checked 2026-11-13) says that when your link appears several times for one query, only the highest one counts, and then all the queries it appeared for are averaged. Google's example: you appear at positions 2, 4 and 6 for one query, which counts as 2; at 3, 5 and 9 for another, which counts as 3; the average position is 2.5.
The help page adds two points: a position is recorded only for links that got an impression, so queries with no impressions are not in the average; and the rank you see when you search yourself may differ from the report, because results are affected by search history and location.
This means an average position of 8 can be “always at 8” or “1st on some days and 15th on others”; one number in the report cannot tell them apart.
Position 8 but no clicks: what is going on?
MochiStamp has an Arabic query “صور كرتون 3d” (3D cartoon images): from 2026-10-17 to 11-12 it had 68 impressions and only 1 click, with a weighted average position of 8.9. Across the 7 days with data, the daily position stayed between 7.9 and 10.0, and the single click came on 11-11. It sat in the lower part of page one and never reached the top few spots on any day.
| Date (2026) | Impressions | Clicks | Position that day |
|---|---|---|---|
| 11-03 | 1 | 0 | 8.0 |
| 11-04 | 1 | 0 | 9.0 |
| 11-08 | 10 | 0 | 10.0 |
| 11-09 | 11 | 0 | 9.4 |
| 11-10 | 16 | 0 | 8.8 |
| 11-11 | 18 | 1 | 8.8 |
| 11-12 | 11 | 0 | 7.9 |
So the reason for a case like “position 8, no clicks” is simple: the page sits in the lower half of page one, where the click-through rate is low anyway (see the table below). We have no screenshot of the results page for this query and no AI Overview data, so we cannot say what sits above it; for what to do next, see Impressions but no clicks: what to do with pages ranking 8 to 20.
An average position of 3.2: was it always in the top 3?
Not necessarily. The Persian query “استیکر در جواب مرسی” (a sticker to reply “thank you”) had 46 impressions and 9 clicks in the same window, with a weighted average position of 3.2. Split it up: from 10-29 to 11-05 it climbed from 10th to 2nd, and from 11-06 it was 1st every day. The 3.2 is a blend of those two stretches and does not mean it was 3rd on any day.
| Period (2026) | Impressions | Clicks | Click-through rate | Weighted position |
|---|---|---|---|---|
| 10-29 to 11-05 | 17 | 1 | 5.9% | 7.1 |
| 11-06 to 11-12 | 29 | 8 | 27.6% | 1.0 |
| Whole window | 46 | 9 | 19.6% | 3.2 |
This page (MochiStamp's Persian guide on reply stickers) was committed on 11-04 and has had impressions in the dashboard's page table since 11-06. The click-through rates of the two stretches differ by nearly 5 times, but each stretch has only a dozen or two dozen impressions and 1 and 8 clicks; the sample is small, so all we can say is that there were clearly more clicks once it reached 1st, and we cannot derive a click-through rate rule.
Conclusion: when you look at a query's position, look at the daily changes first and the window average second; an average that lands in the middle often means it moved position partway through.
At which position band does click-through clearly improve?
On MochiStamp, the click-through rate for positions 1–3 is 10.8% and for positions 4–10 it is between 2.2% and 2.5%. The window is 2026-10-17 to 11-12, counting only the 923 impressions for which Search Console gave the query text (14% of the site's impressions); each query is placed in a band by its own weighted position, rounded down.
| Position band | Queries | Impressions | Clicks | Click-through rate |
|---|---|---|---|---|
| 1–3 | 23 | 166 | 18 | 10.8% |
| 4–7 | 52 | 395 | 10 | 2.5% |
| 8–10 | 57 | 267 | 6 | 2.2% |
| 11–20 | 18 | 33 | 4 | 12.1% |
| Beyond 20 | 30 | 62 | 1 | 1.6% |
The 12.1% in the 11–20 band comes from only 33 impressions and 4 clicks; it does not mean “the lower, the better”, it means the sample is too small, so do not treat it as a rule. An earlier 28-day version is in Impressions but no clicks: what to do with pages ranking 8 to 20; its window is one day shorter and its band cut-offs differ, so the numbers are slightly different.
This table covers only queries with text; anonymized queries have no text and are not in it; for why, see Why do a Search Console query table's impressions add up to less than the whole site? What are “anonymized queries”?.
What should you watch for when judging by average position?
1. Do not rely on a single number; break it down by day
In Search Console, shorten the date range or view the query by date. Whether the position moved partway through is visible at a glance.
2. Do not draw conclusions from queries with few impressions
Position is recorded by impression, so an “average of 1.0” from a few impressions may just be what those few happened to catch. Our 11–20 band has only 33 impressions.
3. Leave the newest days out for now
Search Console's data for the last day or two is not complete yet; see How many days is Search Console data behind? Why are the latest days low?.
4. For queries at positions 8 to 10, first check whether there is a dedicated page
With one page per query, the position can sometimes climb a notch; for when a dedicated page is worth it, see Is one page per query worth it? Data from single-query pages on a small site.
What are the limits of these numbers?
- Only one site, MochiStamp, and 27 days; queries with text cover only 14% of the site's impressions.
- The positions are our own average weighted by daily impressions; how Search Console weights the whole site is not stated on Google's page.
- The query table and the page table are stored separately, so we cannot tell which page a query landed on.
- We have no history of title changes and no AI Overview data, so we cannot separate the effect of titles and AI answers on clicks.
The page-level case behind a Persian query ranking around 3 and an Arabic query ranking around 8 is in Do Arabic and Persian long-tail pages work? Data from MochiStamp's reply-sticker pages.
FAQ
How is average position calculated in Search Console?
Google's help page says that when your link appears several times for a query, only the highest position is taken, and then all the queries it appeared for are averaged. For example, a best position of 2 for one query and 3 for another averages 2.5.
Why does an average position of 8 get no clicks?
8 is in the lower part of page one, where click-through is low anyway. In MochiStamp's data, the 8–10 band has a 2.2% click-through rate against 10.8% for 1–3. For one Arabic query, 68 impressions produced only 1 click.
Can average position be misleading?
Yes. It is an average over a period: one MochiStamp query averaging 3.2 actually climbed from 10th to 2nd over the first 8 days and then held 1st every day. Look at the daily changes.
Why is the rank I see when I search different from Search Console?
Google's help page explains that results are affected by search history and location, so the position you see yourself may differ from the report.
What position counts as good?
In our data, positions 1–3 have a click-through rate of about 10.8% and 4–10 about 2% to 2.5%. The sample is one small site and these are not general numbers.
Sources
- Google · Performance report (Search): average position (checked 2026-11-13)
- MochiStamp · Search Console query data, Oct 17 to Nov 12, 2026 (Local data dashboard, read 2026-11-13)
- kagentx evaluations: the data on the case study pages; each example gives its date.