Why your average position can improve while your clicks fall

What average position actually measures

Google’s own definition is narrower than most people assume. For each search, Search Console records the topmost position your link held, then averages those positions across the searches in your date range (Google Search Console Help).

The detail that explains most of this article is one line in that same document: a link must get an impression for its position to be recorded. An impression counts when your result is on the page of results the searcher loaded, whether or not they scrolled down to it. If your result sits on page three and the searcher only looks at page one, no impression is counted and no position is recorded for that search at all.

So average position is not “where you rank.” It is “where you rank, averaged only over the searches where your result was on a page somebody opened.”

An illustrative example — the numbers are invented to show the arithmetic:

• A page earns 100 impressions at position 4 and 100 impressions at position 30.
• Its average position is 17.
• Now fewer searchers go on to page three, where the position-30 result sits, and those 100 impressions disappear.
• The average position becomes 4.

No ranking changed. Nothing got better. And almost no clicks were lost, because a result on page three rarely earns any.

Five reasons position rises while clicks fall

1. Fewer searchers reach your deep positions

This is the example above, happening for real. When a site’s low-ranking impressions drop out of the data, the average climbs, because the positions left behind are the better ones. Clicks do not follow, because those deep impressions were almost never producing clicks in the first place.

2. Automated impressions left the data

In September 2025, Google stopped supporting the &num=100 parameter, which had let rank-tracking tools and crawlers retrieve up to 100 results in a single request. Search Engine Land’s analysis describes it as a reporting change rather than a ranking update: impressions fell and average positions shifted even though search results themselves did not change (Search Engine Land, October 2025).

Two practical consequences:

• Any comparison that crosses September 2025 is not like for like for impressions and average position. Clicks are far less affected, because the automated requests were not clicking anything.
• The effect became more visible over time, not less. In December 2025, Search Engine Roundtable reported that the jump in average position had become much more noticeable a couple of months on, and one practitioner quoted there found it strongest on sites ranking for terms that are popular with SEO-tool users (Search Engine Roundtable, December 2025). Queries that SEO tools track heavily are the likeliest to have carried the old noise.

3. An AI summary is answering the question

Pew Research Center analysed the browsing data of 900 US adults across 68,879 Google searches in March 2025. When an AI-generated summary appeared, people clicked a traditional search result in 8% of visits, against 15% when no summary appeared. They clicked a link inside the summary itself in about 1% of visits (Pew Research Center, July 2025).

Pew’s data shows an association rather than isolating the summary as the only cause. But it fits the pattern: your position can hold or improve while fewer people need to click anything.

4. Fewer people are searching

Demand moves. Seasonal queries, a news cycle ending, or a product category cooling will cut impressions and clicks together, while your position stays where it was — or improves, if competitors lose visibility faster than you do.

This is the easiest cause to rule in or out. If impressions and clicks fell by roughly the same proportion, look at demand before you look at anything on the page.

5. Your result stopped earning the click

Sometimes the ranking holds and the result simply becomes less attractive. Google may rewrite the title it shows. Ads, a map pack, videos or a “People also ask” block may push your listing further down the screen. A competitor may add ratings or pricing to their snippet and draw the eye away from yours.

None of this moves your recorded position much. All of it can move your click-through rate.

How to check which one applies to you

Work through these in order. Most sites find their answer in the first three steps.

1. Mark September 2025 as a break in your data. Compare periods that sit entirely after it. If you have to compare across it, compare clicks, not impressions or average position.
2. Look at impressions before position. If impressions fell alongside clicks, the likely causes are demand or reduced visibility (reasons 1, 2 and 4). If impressions held or rose while clicks fell, the problem is click-through rate (reasons 3 and 5).
3. Open the Queries tab and sort by impressions. Compare the click-through rate of each of your top queries between the two periods. A falling total is usually a handful of queries, not all of them.
4. Search your top queries yourself. Look at the live results page. Is there an AI summary? More ads? A map pack? Your result lower on the screen than its position suggests?
5. Separate brand from non-brand. Use a regular-expression filter on your brand name. A fall in branded clicks is a different problem from a fall in discovery clicks.
6. Split by device and country. Patterns often live in one segment and get averaged away in the total.
7. Ignore average position for low-impression queries. Because position is only recorded when a result gets an impression, a query with a few impressions can swing wildly on a handful of searches.

What to fix first

Queries where you already rank near the top, with a low click-through rate. This is the cheapest win available. Rewrite the title and description so they say specifically what the page gives the searcher — a price, a timeframe, a direct answer — rather than describing the business.

Queries where an AI summary appears. You cannot stop the summary appearing, but you can make your page the kind of source it draws on. Answer the question in the first two sentences, state specific facts in complete sentences, and make sure AI crawlers can reach the page. This is the work behind AI search optimization.

Your reporting. Stop reporting average position on its own. Report it next to impressions and clicks, per query rather than sitewide, and annotate September 2025 on every chart that crosses it.

If your numbers don’t fit any of these five

That is worth a closer look, and it is exactly what our free SEO audit does: we run these checks on your own Search Console and send you the working. See our SEO services.