Impressions but No Clicks? Read the Position Data First

What does an impression actually mean?

Less than most people assume. Google's Search Console metrics documentation defines an impression as a user having "seen (or potentially seen) a link to your site." That parenthetical carries the weight. An item generally counts as an impression whenever it appears on the results page the user loaded, whether or not they scrolled to it.

There is a narrow exception: inside interactive elements such as carousels or expanding sections, the item must actually be scrolled or expanded into view before it counts.

So a result sitting at position 68 on page seven counts as an impression the moment that page is rendered — and the number of people who load page seven of a Google search and read to the bottom is close enough to zero to treat as zero.

Two further definitions matter when you are trying to reconcile the numbers:

• Clicks are deduplicated per query. Clicking your result, going back, and clicking it again "counts as only one click."
• Impressions are deduplicated too. "Scrolling away and back, or paging away and back, during a single query or session does not count as multiple impressions."

Both metrics are more conservative than people expect, which is worth knowing before you conclude the data is broken.

How much does click-through fall with position?

Steeply, and then it keeps falling. The exact figures vary by study — distrust anyone quoting them to two decimal places as though they were constants — but the shape is consistent across every dataset.

Rough position band	Published CTR range	What 2,500 impressions yields
1	~19–40%	475–1,000 clicks
2–3	~7–20%	175–500 clicks
4–10	~2–6%	50–150 clicks
11–20 (page 2)	~0.5–2%	12–50 clicks
21–50 (pages 3–5)	~0.1–0.5%	2–12 clicks
50+	under ~0.1%	0–2 clicks

The ranges are wide on purpose. A study of 200,000 keywords reported position 1 falling from 28% to 19% between 2024 and 2025 — a 32% decline in a single year — while other datasets put position 1 nearer 40% for navigational queries where the searcher already knows which brand they want. Device, intent, industry and SERP features all move the number.

What does not move is the bottom of the table. Below roughly position 30, no realistic impression volume produces meaningful clicks. If that is where you are, you do not have a click-through problem to solve. You have a ranking problem, and click-through is the symptom you happen to be looking at.

How is average position really calculated?

This is the most misread number in Search Console, and the documentation is unambiguous once you read it carefully.

Google takes the topmost position your property occupied for a given query, then averages that across all queries in which you appeared.

Google's own worked example: suppose your site appears at positions 2, 4 and 6 for one query — that query counts as position 2. For a second query you appear at positions 3, 5 and 9 — that counts as position 3. Your average position is (2 + 3) ÷ 2 = 2.5.

Three consequences follow, and each changes how you should read the number:

1. It is an average over queries, not over impressions. A query that fired once at position 4 counts exactly as much as a query that fired ten thousand times at position 80. A few obscure long-tail queries where you happen to rank well can drag a genuinely poor average upward.
2. Only your best placement per query is counted. If a query returns three of your pages at positions 12, 40 and 61, that query is recorded as position 12. Site-wide average position is therefore systematically optimistic about your real depth.
3. Google itself warns the value "can be misleading", because position varies by result type and device and reflects only impressions users actually saw.

The practical upshot: never act on the site-wide average alone. Segment it.

The diagnostic procedure

Run these four checks, in order, before changing any content. Each is a filter in the Search Console Performance report.

1. Filter to Position < 20. This is the only cohort where title and description work converts into clicks. If the filtered view returns almost nothing, stop — no amount of copywriting will help, and you have just saved yourself a week.

2. Read the Queries tab, not just Pages. Impressions tell you what Google thinks you are about. If they come from broad head terms adjacent to your topic rather than the specific questions you answer, you are being served as a long-shot filler result. That is a targeting problem, fixed by writing to narrower intent — not by rewriting headlines.

3. Compare the Pages tab against your sitemap. Count how many URLs actually draw impressions versus how many you submitted. A large gap usually means pages are still sitting in the crawl backlog as "Discovered — currently not indexed," which is ordinary on a young site.

4. Check for host duplication. If the same article appears under both example.com and www.example.com, Google is splitting signals between two URLs. Confirm one redirects to the other with a 301 or 308, that your canonical tags point at the surviving host, and that your sitemap uses it consistently.

Where do AI Overviews fit in?

They matter, but they are frequently over-blamed for what is really a depth problem.

The strongest public measurement remains the Pew Research Center study, which tracked real browsing for 900 US adults across 68,879 queries rather than modelling it: users clicked a traditional result on 8% of searches showing an AI summary, versus 15% of those without.

Ahrefs approached the same question from Search Console data, comparing 150,000 keywords with AI Overviews against 150,000 informational keywords without, benchmarked to a pre-launch December 2023 baseline. Position-one click-through on ordinary informational keywords fell from 0.076 to 0.039; on AI Overview keywords it fell from 0.073 to 0.016. Isolating the difference produced a headline figure of a 58% reduction, up from the 34.5% the same team reported in April 2025.

Take the methodology seriously before taking the number: that is a modelled counterfactual built on correlated aggregates, not a controlled experiment, and it applies to keywords that trigger an Overview in the first place.

The honest reading is that on affected queries you should expect roughly half — or worse — of the clicks a given position used to deliver. That is a real tax on informational content, and if your catalogue is mostly single-answer explainers you should plan for it. But it does not take a healthy page to zero. If you are at zero across every page, position is the dominant cause and AI Overviews are a secondary one.

What to do instead

If the position filter came back nearly empty, the work is not copywriting.

• Target narrower intent. Competing for a head term against documentation published by the platform that owns the topic is not a winnable fight for a young site. The specific, awkwardly-phrased, lower-volume version of the same question is winnable — and it converts better when you win it.
• Fix orphaned pages first. A page with no internal links pointing at it is both harder to crawl and starved of signal. This is the cheapest fix available and entirely within your control — see internal linking.
• Go deep in one area before going wide. Coverage of a connected question space is what lets query fan-out surface you across related searches, and it is the mechanism actually doing the work when people say "topical authority".
• Give it time. Indexing, evaluation and re-crawling on a new domain run on a scale of weeks to months. An impression trend rising while clicks sit at zero is a site being discovered — which is the correct order for things to happen in.

What not to do

Do not rewrite titles at position 60. It is the most commonly given advice for this symptom and, at that depth, it is inert. Save the effort for pages already in the top 20, where it genuinely moves the number.

Do not publish more pages to compensate. Volume without substance runs directly into Google's spam policies on scaled content abuse, and a thin page ranking nowhere adds nothing but crawl budget spent on your behalf. Five pages that deserve to rank beat fifty that do not.

Do not chase the site-wide average. Segment by page and by query. The aggregate summarises many different situations at once, and improving it is not something you can act on directly.

Sources: Google Search Console — Performance report metrics · Pew Research Center — Google users are less likely to click links when an AI summary appears · Ahrefs — AI Overviews reduce clicks · Google Search Central — Spam policies