Findings

Each one states the instrument, the window and the sample size. Every one of these is measured on a property we built. Nothing here is estimated, modelled or taken from a third party.

1. 01

   Three query clusters drew 1.37 million impressions at 0.17 percent click-through

   Three groups, covering 24 web addresses in total, averaged positions between 5.6 and 8.6 on Google. In 28 days they were shown 1,366,114 times and clicked 2,374 times, about 17 clicks per 10,000 times shown.

   Three query clusters averaging positions between 5.6 and 8.6 drew 1,366,114 impressions over 28 days and returned 0.17 percent click-through between them. One window shows the rate, not what caused it.

   0.17% click rate 1.4M impressions

   28 days, 2026-08-13 to 2026-09-09 · 3 clusters over 24 canonical URLs, 1,366,114 impressions producing 2,374 clicks

2. 02

   Our ten most-shown pages drew 81 percent of impressions and a third of clicks

   In one 28-day pages report, the ten most-shown web addresses accounted for 81 percent of the impressions counted across 382 addresses. Together they got about 2 clicks per 1,000 times shown. The other 372 together got about 16.

   Ranked by impressions alone, the ten most-shown pages on this site drew 80.7 percent of its impressions over 28 days, at a combined click-through rate of 0.189 percent. The other 372 pages combined came to 1.615 percent, 8.6 times higher, on under a fifth of the impressions. Cut at five or twenty pages, the gap is smaller and runs the same way.

   81% of impressions 10 of 382 pages

   28 days, 2026-08-13 to 2026-09-09 · 382 canonical URLs carrying 2,079,623 impressions and 9,661 clicks

3. 03

   We recorded a 23x revenue spread, then found most of it was a missing import

   Among 20 pages on one site, each carrying the same two ad slots, the best earned 4.6 times as much ad revenue per 1,000 views as the worst.

   The worst page earned Rs 2.08 per thousand views against the best at Rs 47.29. Reading the source showed the worst page carried no ad slots at all. Holding slot count constant, 4.6x of the gap survived.

   4.6x revenue spread 20 pages

   28 days, 2026-08-13 to 2026-09-09 · 21 pages above the view threshold, 14,603 views between them

4. 04

   US page views averaged under a tenth of India's ads per view

   On one site, about 7 ads were shown per 100 page views from the United States, against about 97 from India and 119 to 242 from Canada, Australia, France and the UK. Why is not known.

   Ad impressions per page view ran at 0.072 for the United States, the property's second largest market by page views, against 0.97 for India and 1.19 to 2.42 for Canada, Australia, France and the UK. We do not know why.

   0.072 ads per page view 11,638 US views

   28 days, 2026-08-13 to 2026-09-09 · 19,485 ad impressions across 30,966 page views, 102 countries

5. 05

   AI assistants sent about one percent of our traffic

   AI assistants such as ChatGPT and Perplexity sent 29 of the blog's 2,728 visits in 26 days, about 1 in 100. That counts visits they sent, not how often they cite the site.

   Twenty nine sessions in twenty six days across fifty six posts. Real, spread across more assistants than expected, and far too small to carry a site.

   1.1% AI share of blog visits 2,728 sessions

   26 days, 2026-08-16 to 2026-09-10 · 2,728 blog sessions, of which 29 from AI assistants

6. 06

   A breakdown we had never requested showed 184,133 impressions from translated results

   In 28 days, 184,133 of this site's impressions were translated results, clicked about 14 times per 10,000. We found this by requesting Search Console's search-appearance breakdown.

   In 28 days, 184,133 of this site's impressions were translated results, clicked 255 times. The count comes from Search Console's search-appearance breakdown, which had never been requested on this property.

   0.14% click rate 184,133 impressions

   28 days, 2026-08-13 to 2026-09-09 · 184,133 impressions producing 255 clicks at average position 8.37

7. 07

   We kept FAQ markup Google had stopped rewarding, aimed at AI assistants

   A lead generation site shipped FAQ markup that Google had stopped showing as rich results, aiming it at AI assistants. In the next eleven weeks, 27 of its 153 visits came from an AI assistant. These figures do not show whether the markup brought any visits. Its contact actions are not reported: two reports of them disagree.

   FAQPage markup stopped producing rich results years ago. A lead generation site shipped it anyway, as a signal to answer engines, and 27 of its 153 sessions in the following eleven weeks came from an AI assistant. Its contact actions are not reported here, because two reports of them do not agree.

   18% AI share of visits 27 of 153 sessions

   2026-06-11 launch to 2026-08-25 · 153 sessions, of which 27 from an AI assistant

8. 08

   Inside one position band, combined click-through differed nearly tenfold

   Among pages averaging positions 5 to 10 on Google, the ten shown most got about 2 clicks per 1,000 times shown, taken together. The other 170 got about 18. The two groups did not average the same position. How much of the gap that explains is not known.

   As groups, pages averaging position 10 to 20 had a higher combined click-through rate than pages averaging 5 to 10. Inside the 5 to 10 band, the combined rate was 0.189 percent for the ten most-shown pages and 1.843 percent for the other 170.

   9.8x click-rate ratio 180 URLs

   28 days, 2026-08-13 to 2026-09-09 · 382 canonical URLs carrying 2,079,623 impressions and 9,661 clicks

9. 09

   A page with fifty five times fewer impressions produced more clicks

   The page Google showed most, 466,151 times in 28 days, was clicked 618 times; a calculator shown 8,521 times was clicked 783 times. The ten addresses shown most and the ten clicked most have only three in common.

   The largest page on the site drew 466,151 impressions and 618 clicks. A calculator with 8,521 impressions returned 783. Ranking the site by reach and by traffic produces two different lists.

   55x fewer impressions 2 pages

   28 days, 2026-08-13 to 2026-09-09 · 382 canonical URLs carrying 2,079,623 impressions and 9,661 clicks

10. 10

    The long tail advantage is 2.27x, and narrower inside a position band

    On one group of related pages, queries with fewer than 100 impressions in 28 days got 2.27 times the click rate of queries with 1,000 or more, taken together. Inside an average-position band, the ratio was 1.57 at 5 to 8 and 1.23 at 8 to 11.

    On one cluster, queries under 100 impressions had 2.27 times the combined click-through of queries with 1,000 or more. Inside a single average-position band the ratio is 1.57x at 5 to 8 and 1.23x at 8 to 11. Both are smaller than the cluster-wide figure, and far smaller than the 125x our own notes recorded.

    1.57x click-rate ratio 802 queries

    28 days, 2026-08-13 to 2026-09-09 · 3,328 queries carrying 636,317 impressions and 820 clicks on one cluster

11. 11

    Four days carried a quarter of 28 days of impressions, at the lowest click rates in the window

    Over 28 days, our largest site was shown 2,055,560 times in Google Search. A quarter of those came on four days, 5 to 8 September 2026. On those days about 36 in every 10,000 showings got a click, against about 51 on the other 24 days.

    Impressions rose 52 percent from the first half of a 28-day window to the second and clicks 40 percent, so the combined click rate fell from 0.493 to 0.452 percent. The half-window comparison hides a concentrated pattern: four high-impression days had the lowest daily click rates.

    25% of impressions 4 of 28 days

    28 days, 2026-08-13 to 2026-09-09 · 28 daily totals for one property, 2,055,560 impressions producing 9,629 clicks