A page on a site I run collected 4,347 impressions over twelve months and produced zero clicks. Average position 8.1 — first page, most of the time.
Zero clicks at position 8 is a strange number. Position 40 with no clicks is just losing. Position 8 with no clicks means Google repeatedly decided you belonged in front of people, and every one of them chose something else.
The obvious read is a CTR problem: bad title, weak description, rewrite it and move on. That read is wrong often enough that it's worth knowing what to check first.
Start with the queries, not the page
Search Console will give you the queries per URL if you filter by page and pull the query dimension. That's the step people skip, because the page-level table is right there on the first screen and the query breakdown is two clicks deeper.
Every query hitting my country-level page looked like this:
amsterdam coffeeshop rules tourists 2026 19 impressions pos 9.6
amsterdam coffeeshops tourists allowed 2026 12 impressions pos 8.2
amsterdam cannabis laws tourists 2026 12 impressions pos 8.9
can tourists buy cannabis in amsterdam ... 14 impressions pos 6.7
amsterdam coffeeshop rules tourists 2026 offic 6 impressions pos 9.8
Not one of them was about the country. All of them were about one city inside it.
I already had a dedicated page for that city. So I checked which of my URLs ranked for the single biggest query:
--- pages ranking for "amsterdam coffeeshop rules tourists 2026"
/blog/amsterdam-coffeeshop-guide-2026 0 cl 35 impr pos 7.6
/world/europe/netherlands 0 cl 19 impr pos 9.6
www./world/europe/netherlands 0 cl 4 impr pos 9.2
Three of my own URLs, one query, zero clicks between them. The third is a www duplicate from a redirect that returns 307 instead of 301 — a separate bug, and a story for another day.
The cannibalisation was real, but it wasn't the disease
The country page had a hand-written title override. Someone — me — had noticed the city got more search volume than the country and retitled the country page to chase it. So a page about a nation's legal framework went to market wearing a city guide's title, against my own actual city guide.
That's textbook keyword cannibalisation and it's worth fixing. The country page now carries country intent: legal status, the national tolerance policy, possession limits, how different cities diverge. The city guide keeps the city questions.
But splitting them doesn't explain zero clicks. Two pages splitting a query should produce some clicks, just fewer than one strong page. Zero is a different signal.
Read the queries again
Look at what those searches actually contain. Not the nouns — the modifiers.
2026. official. allowed. tourists 2026 official.
Those aren't people asking how coffeeshops work. Those are people checking whether a specific thing happened. There had been a long-running proposal to bar non-residents from buying cannabis in Amsterdam — it was back in the news in February 2026, and it was dropped from the city's coalition agreement on 3 June 2026.
So I searched my own two pages for the word "ban."
Zero occurrences. On either one.
I had two pages ranking on the first page of Google for a question neither of them answered. Google put me in the room because my pages were about the right subject. Every searcher then picked one of the seven results that actually said whether the thing happened.
Impressions measure adjacency. Clicks measure answers.
That's the reframe I walked away with.
An impression is Google saying "this page is topically near the query." A click is a human saying "this page looks like it contains my answer." They are not two points on the same scale, and treating a click-through rate as a headline-writing score conceals the case where the headline is honest and the page simply doesn't have the answer in it.
Zero clicks at position 8 isn't "your snippet is unappealing." It's right neighbourhood, wrong door.
Which means the fix isn't in the <title>. If I had done the obvious thing — rewritten the title to be punchier — I'd have made the page more attractive at describing something the searcher didn't ask about, and the number wouldn't have moved. Then I'd have concluded the query was just low-intent and stopped looking.
What actually shipped
Two changes, in this order:
- Separate the intents. The country page answers country questions; the city page answers city questions. They stop bidding against each other.
- Answer the question that's being asked. The city guide now leads with the outcome — the residents-only rule was debated for years, was back on the agenda in February 2026, and was left out of the June coalition agreement; the shops stay open; the tourist tax went up instead.
While I was in there I found a second version of the same failure. The page said "no consumption on the street." The actual rule is a €100 on-the-spot fine in four named districts, in force since May 2023 and actively enforced. I had turned a specific, checkable, useful fact into a vague one — which is its own kind of not-answering.
The check worth stealing
For any URL with lots of impressions and few clicks:
- Pull the queries for that URL, not the page-level average.
- Ask whether those queries match what the page is about. If they don't, you have an intent mismatch, not a CTR problem.
- Query the same terms against all your URLs. If more than one comes back, you're splitting your own vote.
- Then do the dumb thing: search your page's text for the key noun in the top queries. If it isn't there, no title rewrite is going to save you.
Step 4 took ten seconds and explained a year of zero clicks.
The part I'd want a reviewer to push back on
I can't prove the fix worked yet, and I'm suspicious of posts that skip that. Retitling resets how a page is understood, and it usually gets worse before it gets better. The honest position is that I've corrected a page that was demonstrably failing to answer the question it ranked for, and the measurement comes later.
What I'm confident about is the diagnosis, because it's not a theory about ranking — it's a text search. The word wasn't on the page.
