31,305 Impressions, 81 Clicks: What My Search Console Shows
On this page, 10 sections
My Search Console export for the 90 days ending 11 October 2026 shows 31,305 impressions and 81 clicks across 359 pages of this blog. That is a click-through rate of 0.26%, or one click for every 386 impressions.
I am going to read that number as honestly as the data allows, which means I will say what the export shows, what it cannot show, and what I plan to pull next. I will not hand you a cause I have not tested.
If you run a small blog and your own report looks like this, the useful part is the method for splitting the problem into pieces, not my particular numbers.
Where these numbers come from
I exported the Search Console pages report for the 90 days ending 11 October 2026 and summed the rows for the main host, adityaarsharma.com. The other subdomains I run are excluded.
The export holds the top 500 rows across all hosts. Of those, 359 belong to the main host. That cap matters, because pages with very few impressions can fall outside it. So the honest wording is “the top 500 pages Search Console returned”, never “every page on the site”.
I also did not cross-check the totals against the headline figure in the Performance report itself. Google’s API guide says impressions, clicks and position are calculated differently when you group results by page than when you group by property, so a sum of page rows may not match the property total.
The 81 clicks and 31,305 impressions are the sums of the 359 rows I have. Treat them as that and nothing larger.
An earlier pull I did with a different filter gave 316 pages, 73 clicks and 30,663 impressions. I could not reproduce those figures from the export file, so I dropped them and recomputed everything from the file in front of me. The title of this post carries the recomputed numbers.
Fifteen pages hold most of the impressions
Sorted by impressions, the top 15 pages account for 17,679 of the 31,305 impressions, which is 56.5%. They account for 22 of the 81 clicks, which is 27.2%. Their combined click-through rate is 0.12%.
The other 344 pages hold the remaining 13,626 impressions and 59 clicks, a click-through rate of 0.43%. So the other 344 pages click at about 3.5 times the rate of the 15 pages that show up most.
| Group | Pages | Impressions | Clicks | CTR |
|---|---|---|---|---|
| Top 15 by impressions | 15 | 17,679 | 22 | 0.12% |
| Other pages in the export | 344 | 13,626 | 59 | 0.43% |
| All main-host rows | 359 | 31,305 | 81 | 0.26% |
Eleven of those 15 slugs start with “how-to”. They are small fixes and spreadsheet tricks: importing YouTube details into Google Sheets, an Elementor SVG upload problem, a Contact Form 7 refresh, a page refresh button. The other four are a comparison of AI memory tools, two posts about YubiKeys and one about robots.txt and AI crawlers.
Counting slugs tells me what kind of page earns the impressions. It does not tell me why they earn them or why people skip them.
Five pages with impressions and zero clicks
Five of the top 15 had no clicks at all in the window. Together they hold 4,604 impressions, or 14.7% of the whole export.
- Extracting links from a sitemap in Sheets: 1,177 impressions, average position 34.8.
- Custom link colors in Elementor: 956 impressions, average position 28.8.
- Removing top and bottom white space in WordPress: 956 impressions, average position 38.4.
- YubiKey Security Key versus YubiKey 5: 869 impressions, average position 32.3.
- robots.txt and AI crawler user agents: 646 impressions, average position 15.3.
Four of the five sit at an average position between 28.8 and 38.4. The fifth, the robots.txt post, averages 15.3, which is closer to the first page.
What average position can and cannot tell me
Google’s help page says the position shown in the Performance report is the topmost position a link to your property or page held in the results, averaged across all queries in which it appeared. Its own example is two queries with topmost positions 2 and 3, giving an average of 2.5.
That makes the number an average of queries, not a rank. A page showing 34.8 could rank around 35 for everything. It could also rank near the top for one query and far down for ten others. The page-level export cannot tell those two cases apart.
The same help page says a link must get an impression for its position to be recorded. Its example is a result on page 3 that nobody scrolled to, which records no position for that query. So the positions I see are for results that were at least shown.
Eleven pages sit deep in the results
Eleven of the top 15 have an average position of 28 or worse, ranging from 28.8 to 53.1. Between them they hold 11,891 impressions and 13 clicks, a click-through rate of 0.11%.
That is where much of my impression volume lives: pages that Google shows for many searches, but at positions where I would expect few people to click. I say “expect” on purpose. I have no click-by-position curve for this blog, and I am not going to borrow one from a study of other sites.
Two pages in that group got 3 clicks each: the Elementor SVG post from 1,502 impressions at 36.2, and the Contact Form 7 post from 1,060 impressions at 53.1. A page averaging 53.1 still collected clicks. That fits the idea that an average hides a spread of queries, but I cannot confirm it from this export.
Four pages with the best average positions
Four of the top 15 average a position of 16.2 or better. That is the first page or just past it. They are the AI memory tools comparison at 10.0, the YubiKey multiple-accounts post at 12.8, the Cloudflare Turnstile post at 16.2 and the robots.txt post at 15.3.
Together they hold 5,788 impressions and 9 clicks, a click-through rate of 0.16%. The AI memory tools comparison alone has 3,092 impressions, 9.9% of the whole export, and 4 clicks. That is 0.13% at an average position of 10.0.
The YubiKey multiple-accounts post does better, with 4 clicks from 1,367 impressions, 0.29%, at an average position of 12.8. That is above the blog’s overall 0.26%, but only just.
These are the pages where I would most like an answer. If the average is a clean 10, why so few clicks? I have candidate explanations, but I have not tested any of them, and the export contains nothing that separates them.
What I cannot conclude
- That the titles are weak. I have not compared my titles and snippets against the results around them for any query.
- That AI answers are taking the clicks. Google’s AI features page says traffic from those features is counted in the Web search type in the Performance report. The export I have does not split them out, so I cannot attribute anything to them.
- That the pages are bad. Low clicks from a query a page ranks poorly for says little about the page itself.
- That 0.26% is low for a blog like mine. I have no benchmark for this blog’s topics, and I am not quoting someone else’s.
What I can say is narrower. Most of the impressions come from a few pages, most of those pages average deep positions, and the pages that average well do not convert those impressions into many clicks. Each of those three observations points at a different next step.
What I am pulling next
The missing piece is the query dimension. For each of the top 15 pages I want the list of queries with impressions, clicks and position, so I can sort them into three piles. The labels below are my working assumptions, not findings.
- Queries at positions 1 to 10 with low clicks. These are title and snippet problems, or SERP layout problems, and they are worth opening a live results page for.
- Queries at positions 11 to 30. These are content and link problems. The page already matches the topic loosely, and it needs to be a better answer or be linked from more of my own pages.
- Queries beyond position 30. My assumption is that these are mostly noise. Impressions here can look large while telling me nothing I can act on.
I will write up what the query export shows in a follow-up and say plainly which of my guesses it kills. If it shows that the average was hiding a few good queries and many bad ones, then the headline click-through rate stops being a useful number for these pages at all.
This post comes from the same habit as the post on what actually makes a WordPress site slow: check the number everyone quotes, then see what is underneath it.
Do this today
Open Search Console, go to Performance, and set the date range to the last 3 months. Switch to the Pages tab and sort by impressions. Write down the top 10 pages with their clicks and average position.
Then click the first page and open the Queries tab. Split its queries into the three piles above: positions 1 to 10, 11 to 30, and beyond 30. Do that for your top 3 pages before you touch a title or rewrite anything.
If your export is capped, as mine was at 500 rows, write “top rows” in your notes rather than “all pages”. The API guide says you can page through every row by raising the start row, with a maximum response size of 25,000 rows.