Organic CTR by position: what 15 small sites earned

It is not one site

One site supplied 61% of the top-three impressions. With it taken out, the top-three band is 2.5% (2.0% to 3.1%): higher, and still a small fraction of the classic curves. The concentration is real: two sites supply 86% of the top-three impressions, and 4 of the 7 dots below rest on 70 to 143 impressions, where one click moves a site’s rate by about a point.

Each dot below is one site’s own rate in a band. Only sites with at least 50 impressions in the band are drawn, because on fewer, one or two clicks swing the rate.

Why our top-three figures sit so far below the classic curves

The gap is widest at the top. From position 4 down, our 0.9% sits close to Advanced Web Ranking’s September 2026 curves (0.9% on desktop and 1.6% on mobile at position 5) and well below the 2019 and 2020 ones. Search Console does not say why a search was not clicked, so these are the likely causes of the gap at the top, with what our data and the published research say about each. No one of them explains all of it.

AI answers take the top of the page

Google’s Search Console help says an AI Overview occupies a single position in search results and that all the links in it are given that same position. A site cited in an overview at the top can therefore be reported at position 1 for a result most people read without clicking.

Pew Research Center followed the Google searches of 900 US adults in March 2025. People clicked an ordinary result on 8% of visits to a results page with an AI summary and on 15% of visits without one, and they clicked a link inside the summary on 1%. Two of the figures measured on searches with an AI answer sit close to ours. First Page Sage puts position 1 at 3.6% with an AI answer and 22.6% without. Ahrefs measured 1.6% at position 1 for keywords with an AI Overview and 3.9% for informational keywords without one, on desktop in December 2025. Advanced Web Ranking’s curve for results pages with an AI Overview does not sit close: it puts position 1 at 14.8% on desktop.

Inside our own top three, position 1 was clicked less often than positions 2 and 3. That fits citations in AI Overviews being counted at position 1, though it does not prove it.

Google disputes the direction. Its head of Search wrote in August 2025 that total organic click volume from Search to websites had been relatively stable year on year, and called third-party reports of steep declines inaccurate. Google published no figures with it.

Small, young sites

Ahrefs’ June 2026 benchmark of 422,421 sites put the median whole-site click-through rate at 0.58% for sites with up to 10 indexed pages and 1.74% for sites with more than 5,000. Ahrefs notes this is whole-site CTR, which mixes every position, so it says little about the top three. Across every search, hidden ones included, our 15 sites came to 1.0%. Our data cannot test the age of a site: the seven sites with only five to six weeks of data hold 4.6% of the top-three impressions.

The searches small sites rank for

40% of the top-three impressions came from searches that put a site in the top three fewer than five times in six months. The 45 searches that did so 20 times or more clicked at 1.7% (1.3% to 2.2%), so the low rate is not only one-off searches. Rare searches are often long ones, though our data does not measure length, and long searches are where Google shows AI summaries most: Pew found one on 53% of searches of ten words or more and on 8% of one- and two-word searches.

Search Console’s query table understates the overall rate

Search Console leaves anonymised queries out of its query table. Google describes them as queries not issued by more than a few dozen users over two to three months. Only the query table gives a position we can put in a band. On our sites the hidden searches were 46% of impressions and 64% of clicks, and they were clicked at 1.4%. Every band on this page is built from the visible rows. Across all searches the sites’ rate is 1.0%, against 0.7% in the visible rows, but Search Console gives the hidden searches no position, so whether each band would rise, and by how much, cannot be measured.

Our data can’t separate market from site

Twelve of our 15 sites are outside the US. Advanced Web Ranking’s curves use US results and First Page Sage’s are weighted to US searches. If market explained the gap at the top, the US sites would sit closest to those curves. They sit lowest: the three US sites’ top-three band is 1.0% (0.8% to 1.3%), against 2.5% (2.0% to 3.1%) for the twelve outside the US. Almost all the US top-three impressions come from one site, the largest in the study. So there is no sign that market mix explains the gap, but one site cannot settle it. Lower down the order reverses. At positions 4 to 10 the US sites are clicked at 1.5% and the others at 0.6%, so the mix of sites does shape the lower bands.

Position itself still matters

None of this means position stopped counting. In an eye-tracking study at Cornell (Joachims and colleagues, 2005), people still favoured the first link when the researchers swapped the top two results. A Microsoft Research experiment (Craswell and colleagues, 2008) found that a “cascade” model, in which people read down from the top and leave at the first worthwhile result, best explained position bias in the top ranks. An Expedia field experiment (Ursu, 2018) found the effect of position is smaller than simple comparisons suggest, because better results also sit higher. Our bands fall with position too: from 1.6% at the top to 0.5% on page two.

The published curves disagree with each other

At position 1 the figures in the table below run from 1.6% to 35.7%, a factor of about 20. The spread comes from method: whose Search Console accounts, which searches, which device, which months, and whether results pages with AI answers are counted. Each study below is published by a company that sells SEO software or SEO services, as we do.

Publishers who say AI answers took their clicks

The traffic figures below are the companies’ own claims. None has been tested in court.

1. 24 February 2025: Chegg sues Google

   The education company’s complaint in the US District Court for the District of Columbia claims that as Google answered more of the searches it relied on, it “experienced declines in click-through rates” to its website.

2. 30 June 2025: independent publishers complain to the European Commission

   The Independent Publishers Alliance said AI Overviews cause publishers “traffic, readership and revenue loss”, as Reuters reported on 4 July. Google replied that new AI experiences in Search create “new opportunities for content and businesses to be discovered” and that “sites can gain and lose traffic for a variety of reasons”.

3. 12 September 2025: Penske Media sues Google

   The publisher of Rolling Stone, Billboard and Variety claims its organic affiliate revenue had fallen by more than a third from its peak by the end of 2024, because Google Search sent fewer visitors.

4. 9 December 2025: the European Commission opens an investigation

   It is examining whether Google used publishers’ content for AI Overviews and AI Mode without appropriate compensation and without letting them refuse. The Commission says opening an investigation does not prejudge its outcome.

5. 3 June 2026: the UK regulator sets rules for Google and publishers

   The Competition and Markets Authority requires Google to give publishers a way to keep their content out of AI features in search, and to credit their content with clear links.

6. 30 September 2026: both lawsuits dismissed

   Judge Amit P. Mehta granted Google’s motions to dismiss the Chegg and Penske antitrust claims and declined to hear their unjust enrichment claims. A dismissal on the pleadings does not test the traffic figures.

What we make of this

This section is our reading. Everything above it is what the data and the sources say.

A published CTR curve is an average over somebody else’s searches, and the curves we found differ about twentyfold at position 1. In the top three, our 15 small sites sit at a small fraction of every classic curve, close to First Page Sage’s and Ahrefs’ figures for searches with an AI answer, and well below Advanced Web Ranking’s. On our sites in 2026, a top-three position clicked once or twice in a hundred impressions was ordinary.

So we judge a page’s click-through against its own history: the same page and the same searches, before and after a change to its title or its content. Then we look at the enquiries the page brings, because a click is only worth having when it turns into a call, a booking or a sale.

Reading your own click-through rate

1. 1

   In Search Console’s Performance report, open one page at a time and compare the last three months with the three before.

2. 2

   Filter out searches for your own name. People looking for you by name click far more often than any curve allows for.

3. 3

   Set the page’s total clicks against the sum of its query rows. The difference is the anonymised searches, and on a small site it can be most of the clicks.

4. 4

   If impressions rise and clicks do not, check whether the page is being cited in AI Overviews. Search Console’s generative AI report, available to all sites since August 2026, shows those impressions but not clicks.

5. 5

   For one row at a time, the CTR calculator turns clicks and impressions into a rate and sets it against the position.

Sources

• Google, Search Console Help, “What are impressions, position, and clicks?”, and “Data anomalies in Search Console”. Read 1 November 2026.
• Google Search Central Blog, “A deep dive into Search Console performance data filtering and limits”, 19 October 2022, and “Introducing Search Generative AI performance reports in Search Console”, 3 June 2026.
• Pew Research Center, “Google users are less likely to click on links when an AI summary appears in the results”, 22 July 2025.
• Google, “AI in Search is driving more queries and higher quality clicks”, 6 August 2025.
• Advanced Web Ranking, Google Organic CTR Study, September 2026 data. First Page Sage, “Google Click-Through Rates (CTRs) by Ranking Position in 2026”, updated 22 September 2026. Backlinko, “We Analyzed 4 Million Google Search Results”, 2019. Sistrix, “Why (almost) everything you knew about Google CTR is no longer valid”, 2020. Ahrefs, “Update: AI Overviews Reduce Clicks by 58%”, 4 February 2026, and “What is a Good Organic CTR?”, 22 June 2026.
• Joachims T. et al., SIGIR 2005, doi 10.1145/1076034.1076063. Craswell N. et al., WSDM 2008, doi 10.1145/1341531.1341545. Ursu R. M., Marketing Science 37(4), 2018, doi 10.1287/mksc.2017.1072.
• Chegg, Inc. v. Google LLC, No. 1:25-cv-00543 (D.D.C.); Penske Media Corp. v. Google LLC, No. 1:25-cv-03192 (D.D.C.); memorandum opinion of 30 September 2026. European Commission IP/25/2964. UK Competition and Markets Authority, 3 June 2026. Reuters, 4 July 2025.