Finding 02
Our ten most-shown pages drew 81 percent of impressions and a third of clicks
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.
In plain terms
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.
- Published
- Window
- 28 days, 2026-08-13 to 2026-09-09
- Sample
- 382 canonical URLs carrying 2,079,623 impressions and 9,661 clicks
- Instrument
- Google Search Console pages report, one 28-day window: 382 web addresses ranked by times shown, and the top 5, 10 and 20 each compared with all the others.
Impressions are easy to read as demand waiting for a click. On the property below, most of them went to a handful of pages, and together those pages had a far lower click-through rate than the rest. Why is not measured here.
The size of it
Over twenty eight days the property returned 2,079,623 impressions and 9,661 clicks across 382 canonical URLs, a rate of 0.465 percent. Ranked by impressions alone, the ten pages shown most drew 80.7 percent of the impressions and a third of the clicks.
| Top pages by impressions | Impressions | Share | Combined click-through | All other pages, combined | Gap |
|---|---|---|---|---|---|
| Top 5 | 1,348,901 | 64.9% | 0.180% | 0.989% | 5.5x |
| Top 10 | 1,677,291 | 80.7% | 0.189% | 1.615% | 8.6x |
| Top 20 | 1,833,240 | 88.2% | 0.284% | 1.806% | 6.4x |
One pull, one window, page level throughout, so each cut and the pages outside it are disjoint and the shares add up. URLs are canonicalised first: fragment rows removed, then www and trailing-slash variants merged. The denominator is the sum of these page rows, not the property day chart, which totals 1.2 percent less and would put a numerator and denominator from two different aggregations in the same fraction. The search-appearance breakdown is deliberately excluded: it may overlap these page rows, so it cannot be added to them.
What the ten pages are
| Kind | Pages | Share of impressions | Combined click-through | Average position |
|---|---|---|---|---|
| Bank branch codes | 3 | 38.2% | 0.222% | 7.92 |
| Festival and observance dates | 2 | 17.7% | 0.068% | 8.61 |
| Fuel cost | 1 | 11.0% | 0.195% | 6.89 |
| Calorie counts | 2 | 9.0% | 0.176% | 5.62 |
| Alphabet chart | 1 | 2.9% | 0.189% | 8.80 |
| Astrology calculator | 1 | 1.7% | 0.706% | 5.43 |
| All ten | 10 | 80.7% | 0.189% | 7.65 |
A likely reason for the group's lower combined click-through rate is that most of these pages answer with a code, a date or a number, which a results page can show by itself. That is not measured here, and it played no part in choosing them. Page quality, speed and titles were not measured either, so none of them is ruled out. Average position is not ruled out either; finding 08 compares the same ten pages only with the other pages averaging position 5 to 10.
The ratio that matters is not 81 percent
It is 0.189 against 1.615. The other 372 pages had a combined click-through rate 8.6 times higher, on under a fifth of the impressions. The site-wide rate blends the two groups, weighted about four to one toward the ten pages with the lower combined rate.
Where the number is sensitive, stated plainly
The cut changes the size of the gap, not its direction: 5.5 times at five pages, 8.6 at ten, 6.4 at twenty. At every cut, the most-shown pages' combined click-through rate was under a fifth of the other pages' combined rate.
Where the denominator comes from
The share divides by the sum of the same page rows, never by a total from another report. Page rows sum to 1.2 percent more than the property day chart, so a page-level numerator over the day-chart total would mix two aggregations, and a denominator from a different pull would mix two exports.
A mismatch of that size would not change the argument, which is exactly why it is easy to miss. The rule that catches it is duller than the finding: a numerator and a denominator have to come from the same count of the same thing.
Why the site-wide rate is a poor target
A brief to lift 0.465 percent points at the most-shown pages first, because that is where the impressions are. On this site those pages have the lower combined click-through rate, and nothing here shows that work on them would raise it.
How to check your own
Export Search Console by page for a full month. Strip fragment rows, merge www and trailing-slash variants, then rank the pages by impressions without looking at clicks. Compare the combined click-through rate of the top pages, their clicks summed over their impressions summed, with the same figure for all the other pages, at more than one cut, dividing only by the rows in the file. The export stops at 1,000 rows, so on a large site the comparison covers the pages in the file, not every page.
Full method
Search Console pages for one window. URLs are canonicalised first: fragment rows removed, then www and trailing-slash variants merged. Pages are ranked by impressions alone, and the top 5, 10 and 20 are each compared with all other pages on combined click-through rate: a group's summed clicks over its summed impressions, not an average of page rates. Clicks play no part in the ranking. All page-level shares on this site are computed against the sum of the page rows in the same export (382 canonical URLs, 2,079,623 impressions, 9,661 clicks), not against the property day-chart total of 2,055,560. Page sums exceed the property total by 1.2%, which is a Search Console aggregation artefact; mixing the two would put a numerator and denominator from different aggregations in the same share.