A page that moves from sixth to second place in Google does not win a little more traffic. It changes order of magnitude. Click-through rate (CTR) studies by position remain the simplest way to work out what a keyword could be worth.
They come with one caveat, and it is a large one. Averages hide very wide gaps depending on how a results page is laid out, and AI Overviews have moved the reference points most teams still quote from memory.
What follows gathers the figures published by the main studies, with the scope of each stated, then sets out a method for estimating traffic from your own Search Console data.
What click-through rate measures on Google
CTR is the ratio of clicks to impressions: the number of times a link to your site was shown, or could have been, on a results page. That is the definition used by the Performance report in Google Search Console, where CTR equals clicks divided by impressions.
The average position in that same report needs care. For each query, Google takes the highest position held by a link to your page, then averages across all queries. A page ranked second for one query and fifteenth for another shows an average position of 8.5, though it never actually appears eighth for anyone. To reason about CTR, work query by query.
CTR by position: what the studies report
Two studies do most of the work here. Backlinko analysed roughly 4 million search results from Search Console accounts supplied by Semrush, in an April 2025 update. Sistrix studied more than 80 million keywords, mobile only, first published in 2020 and revised in July 2025.
| Position | Backlinko (April 2025, ~4M results, all devices) | Sistrix (2025 revision, 80M+ keywords, mobile) |
|---|---|---|
| 1 | 27.6% | 28.5% |
| 2 | not broken out (top three: 54.4%) | 15.7% |
| 3 | not broken out (top three: 54.4%) | 11% |
| 10 | around ten times less than position one | 2.5% |
| Page two | 0.63% of searchers click a page-two result | not measured |
Sources: Backlinko, Google CTR Stats; Sistrix, CTR study.
The two agree on the shape of the curve. Position one takes more than a quarter of clicks, and the top three a little over half. Backlinko adds that moving from second place to first brings 74.5% more clicks, while CTR across positions eight to ten is close to identical. Gaining a place at the foot of page one changes very little; gaining a place inside the top three changes a great deal. That difference should decide where effort goes.
Intent and brand matter as much as rank
Advanced Web Ranking updates a study every month from millions of search terms, split by device and by industry. On January 2021 data, it reported a CTR of 33.92% in position one for commercial-intent queries against 23.35% for informational ones. Its first-quarter 2026 report gives movements rather than absolute rates: on desktop, branded queries rose by between 1.99 and 5.78 points against the previous quarter, while on mobile the increase stayed at around one point. A query carrying your brand name is not comparable to a generic one, and mixing the two distorts any average you build.
The layout of the results page changes everything
An average per position assumes every results page looks alike. They do not. Sistrix measured position-one CTR against the features present on the page: 34.2% when the page holds organic results only, 46.9% when the first result carries sitelinks, 23.3% with a featured snippet, 16.7% with a knowledge panel and 13.7% when Google Shopping results occupy the top.
That spread is wider than the gap between position one and position three. So before estimating traffic, look at the real results page for your target keyword, on mobile and on desktop.
AI Overviews: what has actually been measured
AI-generated overviews have shifted click-through rates more than any other results-page feature. Several independent measurements exist, each with its own scope.
- Seer Interactive, February 2025: across roughly 10,000 informational keywords ranking in the top 20, from January 2024 to January 2025, organic CTR on queries showing an AI Overview fell from 1.41% to 0.64%. Seer notes these queries already drew fewer clicks before overviews appeared.
- Ahrefs, April 2025: across 300,000 keywords, March 2024 against March 2025, an AI Overview is associated with a CTR 34.5% lower for the page ranked first.
- Pew Research Center, July 2025: a panel of 900 American adults and 68,879 Google searches in March 2025. Users clicked a standard result in 8% of visits that included an AI summary, against 15% of visits without one. Links inside the summary were clicked in 1% of cases, and 18% of searches showed one.
- Seer Interactive, April 2026 update: 53 brands, 5.47 million queries and 2.43 billion organic impressions from January 2025 to February 2026. In February 2026, organic CTR stood at 2.36% with an AI Overview against 3.82% without. Seer also records a rebound, from 1.3% in December 2025 to 2.4% in February 2026, and 120% more clicks per impression when a page is cited inside the overview.
Sources: Seer Interactive, 2025; Seer Interactive, 2026 update; Ahrefs, 2025; Pew Research Center, 2025.
Read these as direction and order of magnitude, not as a rate to apply to your own market: they lean heavily on United States data and on informational queries, and overview coverage differs by country, language and topic. The dependable measure is your own: track clicks and impressions on the queries where overviews appear for your pages. The general lesson holds. Historic CTR curves overstate traffic on informational queries, and much less so on transactional or local searches.
Estimating the traffic a page could win
The base formula is simple: estimated traffic equals monthly search volume multiplied by the CTR at the target position. Reliability rests entirely on those two inputs. The method below keeps the error down.
1. Start from Search Console impressions, not tool volume
Keyword tool volumes are smoothed estimates, often bucketed and out of date. If your page already ranks in the top 20, the Performance report gives the real impression count for every query. Filter by page, then by query, and export three months to absorb seasonality.
2. Record your own CTR by position
For each query, note the position and the CTR observed. With enough queries you end up with your own CTR curve, one that already contains your sector, your brand recognition and the real layout of the pages you compete on. Published averages then serve as a fallback. Exclude branded queries first, or you flatter every projection that follows.
3. Project, then discount for page features
A worked example. A query generates 1,500 impressions a month; your page sits in position seven with an observed CTR of 3%, so 45 clicks. If you target position three and your own curve, or failing that Sistrix, puts that position near 11%, the ceiling is 165 clicks a month. Then look at the results page: if an AI Overview or a Shopping block owns the top, revise that figure down using the gaps measured above.
4. Prioritise by click gap, not by position gap
Rank your queries by the difference between current and potential clicks. A query with 5,000 impressions in position four is often worth more effort than one with 200 impressions in position twelve, even if the second looks easier. That table becomes the roadmap for content work, title tags and internal linking, and the clearest input to bring to scoping an SEO audit. A marketing measurement dashboard saves repeating the exercise by hand every quarter.
Limits worth keeping in mind
- Averages flatten very different situations. Sistrix measures position-one CTR from 13.7% to 46.9% depending on the results page. A single value such as "position one equals 28%" is only an order of magnitude.
- Mobile and desktop diverge. Advanced Web Ranking publishes the two series separately, with movements running in opposite directions by device and industry.
- Branded queries inflate the averages. They behave nothing like discovery queries; strip them out.
- CTR can be improved at an unchanged position. Backlinko reports that title tags of 40 to 60 characters achieve a CTR 33.3% higher than those outside that range. Check that your page titles and meta descriptions give a reason to click.
- Studies age quickly. Data from 2020 or 2021 describes results pages without AI Overviews. Date the source you quote, and prefer your own numbers.
Common questions
What is the average click-through rate for the first position on Google?
Around 27% to 29%: 27.6% for Backlinko (April 2025, around 4 million results) and 28.5% for Sistrix (mobile, 80 million-plus keywords). That average moves sharply with layout, from 13.7% with a Shopping block at the top to 46.9% with sitelinks.
How many clicks does a result on page two of Google get?
Very few. Backlinko reports that only 0.63% of searchers click a result on the second page. In practice, a page beyond position ten is close to invisible for that query: either work it into the top ten or stop counting on it.
How do I calculate the potential traffic of a keyword?
Multiply the monthly impressions for the query, from the Search Console Performance report or a keyword tool, by the CTR you expect at the target position. Prefer the CTR your own site records there, then adjust for what sits on the results page: ads, an AI Overview, Shopping, a featured snippet.
Do AI Overviews really reduce traffic?
They do on the queries where they appear, and most clearly on informational ones. The Pew Research Center (March 2025, 900 American adults) recorded a click on a standard result in 8% of visits with an AI summary against 15% without. Seer Interactive measured 2.36% organic CTR with an overview against 3.82% without in February 2026.
Why does average position in Search Console not match what I see in Google?
Google takes the highest position your page held for each query, then averages across every query and impression. A page in position two for one query and fifteen for another reports 8.5 without ever occupying that place, so analyse CTR query by query.
Can click-through rate be improved without changing position?
Yes. The title tag, the meta description, the URL and structured data all influence clicks at an unchanged rank. Backlinko reports a CTR 33.3% higher for titles of 40 to 60 characters. On pages that already rank well, this is usually the quickest change to make.