Why Your Best-Ranking Pages Get Zero Clicks

[Operator’s read] I run GSC exports across a handful of sites every week, and this one I ran on my own domain this month. A page of mine sits at an average position of 8.9 — genuinely good rank — across 141,000 impressions in 90 days. It converted 103 of those into clicks. That’s a 0.07% click-through rate. Meanwhile a completely different page on the same site, ranking worse on average, converts at over 9%. Same domain, same authority, same crawl budget. The only real difference is what’s sitting in the title tag.

Table of contents

Position and clicks are not the same metric

Most SEO advice optimizes for rank: more backlinks, better internal linking, cleaner technical fundamentals, all in service of moving up the results page. That work is real and it matters. But rank only controls how many people see your result. Click-through rate controls how many of them act on it, and it’s driven by something rank doesn’t touch at all — whether your title and description look like the answer to the exact question the searcher typed.

A page can do everything right technically and still lose on this second axis. I’ve got the receipts from my own site to show it.

Three pages, three outcomes

Here’s what 90 days of Search Console data on alejandrorioja.com actually looks like when you sort by impressions instead of clicks:

The first three rows are ranking fine and getting ignored. The fourth is ranking worse on average and getting clicked at roughly 70x the rate. I didn’t touch backlinks or technical SEO to get that gap — it’s entirely a function of what each result promises versus what the searcher is actually trying to do.

The two failure modes I keep finding

Once I started sorting my own pages by “high impressions, low CTR” instead of by traffic, the same two problems showed up over and over.

1. The query is navigational and the result is editorial. My worst-performing page by CTR ranks for a branded login query. The searcher already knows the destination — they want the login screen, not a 1,500-word explainer about how to get there. Every impression on that query is someone who’s about to bounce back to Google and click the actual login link instead, no matter how well the page is written. No title rewrite fixes this. The page is answering a question nobody asked; it needs a different target query entirely, or a much thinner, faster path to the thing the searcher wants.

2. The title describes the topic, not the answer. My second-worst page is a listicle-style title that names a category (“high CPC keywords”) without stating what’s actually in it — no number, no year, no specific angle. Compare that to the tool page in row four: its title says exactly what the searcher gets and lets them do it in one click. A title that summarizes the subject reads as “an article about this.” A title that states the outcome reads as “click here and this is solved.”

What I’m actually rewriting

For pages that fail check #1 — wrong intent entirely — I’m not touching the title. I’m either retargeting the page at a different query where an editorial result genuinely competes, or accepting the impressions are dead weight and stopping there. No amount of copywriting turns an explainer into a login page.

For pages that fail check #2 — right intent, weak title — the fix is mechanical:

1. State the outcome, not the topic. “High CPC AdSense Keywords” becomes something that says what the reader walks away with — a number, a year, or the specific decision it helps them make.
2. Match the exact phrasing of the query in Search Console, not a paraphrase of it. Pull your own top queries per page (gsc_search_analytics if you’re running this through an agent, or the Performance report by hand) and check whether the words a searcher actually typed appear in your title at all.
3. Write the description as the answer, in one sentence, not a teaser. A description that withholds the answer to get a click trains readers to distrust your snippets over time, and it doesn’t work as well as people assume — Google frequently rewrites vague descriptions anyway, which you can check with a manual site: search for the page.
4. Re-check after 2-3 weeks, not after 2-3 days. Google doesn’t re-render snippets instantly, and CTR moves slower than rank does.

The prompt I run against my own exports

I export the Search Console table (Performance report, filtered to Position and CTR, sorted by impressions descending) as a CSV, then run this:

Here is a CSV export from Google Search Console: page, average position,
impressions, clicks, CTR, for the last 90 days.

Find every row where position is 15 or better (lower number = better) but
CTR is under 1%. For each one, tell me: (1) whether the query intent looks
navigational/branded versus informational, based on the page URL and topic,
and (2) if informational, rewrite the title tag to state the specific
outcome instead of the topic, in under 60 characters. Group the navigational
ones separately — flag those as "wrong target, don't rewrite the title."

That single pass is what surfaced the login-query mismatch on my own site. It’s a five-minute job once the export exists, and it catches the exact failure mode that rank-tracking tools never will, because rank trackers are built to tell you position moved, not that nobody cared once it did.

Where this connects to AI search

The same audit gets more urgent, not less, as AI Overviews and chat answers absorb a growing share of informational queries — zero-click search is compounding for exactly the pages this fixes, because an editorial page with a vague title is the first kind an AI summary swallows whole. A title and description that state a concrete, extractable answer are also more likely to survive as the cited source rather than get paraphrased into an Overview and skipped. If you haven’t run a broader traffic-drop diagnostic recently, start there before assuming CTR is your only problem — this audit is one input into that, not a replacement for it.

FAQ

What CTR should a page ranking in the top 10 actually get?

It depends heavily on query type, but as a rough operator’s benchmark: informational queries at position 1-3 commonly land 15-30% CTR, position 4-10 commonly land 2-8%, and anything under 1% at a top-10 position is a real signal something’s wrong with the title, description, or query-intent match — not just normal variance.

Is this the same thing as the AI Overview zero-click problem?

No, and that’s the point of running this audit separately. AI Overviews suppress clicks that would otherwise happen by answering the question above the result. A CTR mismatch means the page never had a shot at the click in the first place, because the title didn’t match what the searcher was trying to do. Fix the second problem before blaming the first — I’ve found pages where the “AI Overview killed my traffic” story fell apart the moment I checked whether the title matched the query at all.

How often should I run this audit?

Quarterly is enough for most sites. Search behavior and Google’s own snippet rewriting shift slowly enough that a monthly re-check mostly just re-confirms the last one. Run it sooner if you’ve just published a batch of new pages or migrated a site.

Should I use a CTR-tracking tool or just export from Search Console manually?

Start manual — Search Console’s own Performance report has everything in this audit for free, and the export-to-prompt workflow above takes under ten minutes once you’ve done it once. If you’re managing this across a dozen-plus client sites, a rank tracker with built-in CTR trending like Semrush saves the repeated export step; for a single site, it’s not worth paying for something Search Console already gives you.