Impressions but no clicks is the most common Search Console complaint I get, and this week it was my turn. Nine of my own guides were showing up in results and getting nothing. So this is a case note on the fix I ran on 20 August: what I looked at, the rule I used to decide which pages were worth touching, and the exact titles I rewrote.
First the short answer, because that is what the query deserves. Impressions with no clicks means Google shows your page and searchers pick something else. Whether you can fix it depends almost entirely on average position. Past position 20, the problem is not your title, because almost nobody scrolls far enough to see it.
What the data showed
I pulled 90 days of Performance data by page. A batch of my guides had impressions in the tens to hundreds with zero or near-zero clicks. Average position told two different stories inside that batch.
Five guides sat between position 12 and 19: page two, close enough that Google clearly thinks the content is relevant, far enough that the click never happens. Those are the fixable ones. The rest sat past position 40. No title rewrite moves a page from 47 to 8. Those pages need better content or better internal links, and I left their titles for later so I would not fool myself into thinking I had fixed them.
That split is the whole method:
Why do my pages get impressions but no clicks?
Because an impression only means your result was on a page the searcher loaded, anywhere on it. Google’s Performance report documentation counts an impression even when the user never scrolls to your result. A page averaging position 15 collects impressions all day from people who chose between positions 1 and 5 and left. Zero clicks at position 15 is not searchers rejecting your title. Mostly they never read it.
But some of them do reach page two, and those are the clicks the title fight is actually for. Which is why the rewrite is worth doing at 12 to 19 and pointless at 40.
What I changed, exactly
The rule for every rewrite was the same: make the title the query as people type it, not my summary of the page. I checked the real wording in the Performance report’s query dimension and against what already ranks. Three examples from the nine, before and after:
- “Crawled, currently not indexed, explained” became “Crawled, currently not indexed: how to fix it”. People do not search for explanations, they search for fixes. The guide is at crawled, currently not indexed.
- “Why your pages are not showing up on Google” became “Why is my site not showing up on Google? 4 reasons”. The question form is the query form.
- “URL is unknown to Google: what it means and how to fix it” became URL is unknown to Google: meaning and fix. Same promise, eleven characters shorter, so it stops getting truncated mid-sentence in results.
The pattern across all nine: query wording at the front, the page’s promise after the colon, under 60 characters so Google shows what I wrote instead of rewriting it. Google’s title link documentation is explicit that the title element is the main source of the title link, and its first piece of advice is to describe the page the way its audience would. That is all a retitle is.
What retitling cannot fix
Worth saying plainly, because title tweaks are the most oversold fix in SEO. A retitle changes which searchers click when they see you. It does not change how often Google shows you, at least not directly, and it does nothing for a page whose content loses to what ranks above it. My four guides past position 40 kept their problems and will get content work, not wordsmithing. And a page with impressions but no clicks at position 3 is a different animal again: there the title really is being rejected, next to visible competitors, and rewriting it is urgent rather than optional.
When to check the result
Not this week. Google has to recrawl each page, then decide to use the new title, and my own data says indexing and re-evaluation run on Google’s calendar, not mine. I logged the date, and the comparison will be 28 days of clicks and CTR against the 28 days before the change, per page, same report. If the five page-two guides pick up clicks, the method held. If they do not, the honest reading is that position 12 to 19 was too far back even for a matched title, and the next lever is content.
One caveat I would add now that I did not have then: impressions from AI features are folded into this same report with no way to separate them, which I go through in do I need to optimize for AI Overviews. There is now one more cause behind that pattern than there was, since AI Overviews began auto-expanding on some queries in late August 2026.
Before anyone suggests measuring that split in analytics, it does not work. AI Overview traffic cannot be isolated in GA4, so this pattern stays a hypothesis with evidence attached rather than a number.
Sources
- This property’s Google Search Console Performance report, 90 days to 20 August 2026
- Google Search Console Help, Performance report: impressions, position, and clicks
- Google Search Central, influencing your title links in search results