Average Position in Google: How It Is Calculated
Average position in Google Search Console is the average topmost rank of your result in Google Search, weighted by impressions. For every impression Google records the highest position at which your site (or page) appeared; the average position is the sum of those positions divided by the number of impressions. That is why a report can show 8.4 even though the page was never literally “at eight point four”: for some queries it was 4th, for others 26th.
Below: the average position formula, a worked example, why the number changes when impressions grow, and how average position differs by query, by page and by site.
Average position formula
Search Console does not average the positions of rows in a table. It averages over impressions:
Average position = Σ (topmost position × impressions) ÷ Σ impressionsThree rules follow from how Google records the data:
- Only impressions count. If your page did not appear in results for a query, that query does not contribute to the average at all — it is not treated as position 0 or 100.
- The topmost position wins. If two pages of your site appear in the same results, the site-level report records the higher one.
- Lower is better. Position 1 is the top of the results; a growing number means the result moved down.
Every result block occupies a position on the page, so features such as a featured snippet or a carousel take positions too. Google's definition is in the help article What are impressions, position, and clicks?
How is average position calculated: a worked example
A page received 1,000 impressions over 28 days for three queries:
| Query | Impressions | Position | Impressions × position |
|---|---|---|---|
| buy bicycle kyiv | 800 | 4.0 | 3,200 |
| bicycle shop near me | 150 | 12.0 | 1,800 |
| cheap bicycle | 50 | 45.0 | 2,250 |
| Total | 1,000 | 7,250 |
Average position = 7,250 ÷ 1,000 = 7.25.
A simple average of the three positions would give (4 + 12 + 45) ÷ 3 = 20.3 — almost three times worse, because it gives the query with 50 impressions the same weight as the query with 800. That is the most common mistake when people recalculate position from an exported table. In a spreadsheet use =SUMPRODUCT(impressions, position) / SUM(impressions).
To calculate the average position for a group of pages or queries from an export yourself:
- Export the Performance report and open the Queries or Pages tab.
- Filter the rows you need — for example, all URLs in
/blog/or all non-brand queries. - In a new column multiply impressions by position for every row.
- Divide the sum of that column by the sum of impressions.
The result will be close to, but not exactly equal to, what Search Console shows with the same filter: the export contains at most 1,000 rows and omits anonymized queries.
Inside one query the same logic applies to individual impressions: if the page was 3rd for 600 impressions and 7th for 400, the position for that query is (600 × 3 + 400 × 7) ÷ 1,000 = 4.6.
Why average position changes when impressions grow
Because the formula is weighted by impressions, any change in the mix of queries moves the average — even when nothing changed for your main queries.
| Before | After | |
|---|---|---|
| Main query: 800 impressions at position 4 | 3,200 | 3,200 |
| Second query: 200 impressions at position 10 | 2,000 | 2,000 |
| New query: 600 impressions at position 38 | — | 22,800 |
| Impressions | 1,000 | 1,600 |
| Average position | 5.2 | 17.5 |
The page began to show for a new query on the fourth page of results. Average position “dropped” from 5.2 to 17.5, yet the page lost nothing and gained impressions. The reverse also happens: when a page stops showing for distant queries, its average position improves while traffic stays the same.
Other reasons the number moves:
- Personalization, country and device. Results differ by location and device; the report mixes them all unless you add filters.
- Data delay. The last two or three days are often incomplete.
- Short windows. Over a day the average jumps; over 28 days it shows a trend.
So read average position together with impressions and clicks, never alone. More on those metrics — in the article about impressions, clicks and CTR.
Average position by query, by page and by site
| Level | Where in Search Console | What is averaged |
|---|---|---|
| Query | Queries tab, or a query filter | Topmost position of your site for that one query across its impressions |
| Page | Pages tab, or a page filter | Position of that URL across all queries it was shown for |
| Site | The chart totals with no filters | Topmost position of any of your pages per result, across all queries |
These numbers are not interchangeable. A page can rank 3rd for its main query and still have a page-level average of 15 because of dozens of long-tail queries. The site-level average is not the average of your pages either: when several of your pages appear for the same query, only the highest one counts for the site, while each page keeps its own position in the Pages table. For tracking one keyword in one city, use a rank tracker: it records the actual rank in a fixed location and device instead of an average of real impressions.
Average position in Google Ads (AdWords)
Google Ads had its own “Avg. position” metric for ads: the average rank of your ad among other ads in the auctions where it was shown. Google retired it in September 2019. Instead, Google Ads reports impression rate metrics: Search top impression rate (share of impressions shown above organic results) and Search absolute top impression rate (share shown as the very first ad). If you are reading an old AdWords report, the logic was similar — an average over impressions — but it only concerned paid ads and has nothing to do with the organic average position in Search Console.
Average position vs rank tracker: why the numbers differ
Search Console and a rank tracker answer different questions, so a gap between them is normal rather than an error in either source.
| Average position in Search Console | Rank tracker or one-off check | |
|---|---|---|
| What it measures | The average across every impression in the period | The rank on one results page at the moment of the check |
| Location and device | All of them, unless you filter | The ones set for the check |
| Which queries | Every query that produced impressions | Only the keywords you added |
| Whose site | Only your verified property | Any domain |
A typical case: the tracker shows 3rd place for a keyword on mobile in one city, while Search Console shows 7.8, because it averages impressions from other locations, desktop searches and days when the page ranked lower. For a one-off check of a single keyword with a chosen country, language and device, use the Google rank checker; for regular tracking with history, Rank Tracker.
To work through average position across hundreds of queries at once, export the Performance report from Search Console and drop the file into the Search Console export analyzer. It lists striking-distance queries in positions 8–20 by click potential and queries whose CTR is below the benchmark for their position; the file is processed in your browser and never uploaded to our server. What to do with pages close to the top is covered in pages in positions 4–20.
Where PromoPilot shows average position
PromoPilot takes position from Search Console and averages it by impressions, the same way. Variants of one address — with and without a trailing slash, with parameters — are combined.
- “Pulse” — current position and its chart. The position scale is inverted: a line going up means moving towards the top. A growth badge appears only when the improvement exceeds the noise threshold. See what “Pulse” is.
- Project card — “At the top threshold”: how many pages sit on positions 4–20.
- “Radar” — the pair “current position → realistic goal” for each candidate page.
- Launch report — position at the start and after 30, 60 and 90 days.
- Chrome extension — the “Potential” tab with the position of the open page over 28 days.
What is a good average position
There is no universal norm; it depends on the competition in the niche. A useful guide: positions 1–3 collect most clicks, 4–10 bring noticeable traffic, 11–20 almost none, and beyond that practically nothing. Moving from 26th to 18th looks good in a report but adds few clicks; moving from 8th to 4th adds many. That is why pages on positions 4–20 with real impressions are the best candidates for work — see pages at the top threshold and how to choose a page.
Common reading mistakes
- Averaging positions from an export without weights. Always weight by impressions.
- Comparing page position with the position for a single query. They measure different things.
- Panicking when position worsens after impressions grow. Check clicks first; usually the page simply reached new queries.
- Measuring over one day. Compare equal windows: 28 days with the previous 28.