Why are impressions rising while clicks fall?
Short answer
You're being shown for a wider set of queries at positions that don't earn clicks, and the average position hides it. Two mechanisms drive most cases: Google fanning queries out across subtopics so you appear for terms you never targeted, and AI surfaces logging impressions where the answer resolves without a click. Check the distribution of impressions by position band rather than the average, because a growing tail at position 30 lifts impressions and returns nothing.
Average position is the wrong instrument here
Work the arithmetic once and you'll never trust the average again.
Say you hold 1,000 impressions at position 5, earning roughly 5% CTR, so 50 clicks. Then query fan-out adds 4,000 impressions at position 38, earning almost nothing. Your impressions rose 400%, your clicks are unchanged at 50, and your average position moved from 5 to about 31.
Nothing got worse. A report sorted on average position will show a catastrophic decline, and a team acting on it will start rewriting a page that is performing exactly as it did last month.
The correct view is a histogram, not an average. Bucket impressions into position bands and compare the shape between periods:
| Position band | What growth here means |
|---|---|
| 1 to 3 | Real visibility gain. Clicks should follow. |
| 4 to 10 | Modest gain, CTR is much lower but non-zero. |
| 11 to 20 | Almost no clicks. Inflates impressions, flatters nothing. |
| 21+ | Pure impression inflation. Ignore for click analysis. |
If the growth is concentrated in the bottom two bands, your impressions rose and your visibility didn't.
Mechanism one: query fan-out
Google's documentation notes that AI Overviews and AI Mode may issue multiple related searches across subtopics to assemble a response. Your page can be surfaced against subtopic queries the user never typed.
The signature is a wave of new low-volume, long-tail queries in Search Console that you never targeted, each with a handful of impressions and no clicks. Individually trivial, collectively enough to move your totals.
This is not a problem to fix. It's a measurement artefact to exclude before you draw conclusions.
Mechanism two: AI surface impressions
Google's generative AI performance report counts an impression whenever a URL from your site appears within an AI Overview, AI Mode, or an AI feature in Discover. Those impressions were already included in your overall performance totals before the breakout existed.
So appearances inside AI answers have been inflating your impression count all along, and those appearances convert to clicks at a much lower rate. Pew Research found that when an AI summary appears, users click a source link in only about 1% of visits.
If the generative AI report is available on your property, check whether its impression trend explains the gap. If it isn't, treat the divergence between total impressions and click-earning impressions as your proxy.
Mechanism three: the one that is a real problem
Sometimes impressions rise and clicks fall because your best queries lost position while your long tail expanded. The two movements cancel in the average and look like the harmless case above.
To rule it in or out, isolate your top 50 click-earning queries from last year and check their position and CTR today, ignoring everything else. If those specific queries are stable, you have fan-out. If they slipped while the tail grew, you have a genuine decline wearing a disguise.
This check takes ten minutes and it's the one most teams skip.
What to do about each
Fan-out and AI impressions. Nothing, except stop reporting raw impressions as a visibility metric. Report impressions in positions 1 to 10 only, and your trend line becomes meaningful again.
Real position loss on money queries. Treat it as a normal ranking investigation. See Which pages and queries explain most of my organic decline?.
Appearing in AI answers without clicks. This is exposure rather than traffic, and it has value you should measure separately rather than counting as a loss. Our own state of AI search research, covering more than 22,000 answers, found citation presence reaches 97.4% on Google AI Mode and 93.2% on Perplexity, with presence persisting once earned.
Separate the two impression types in Analyze AI
The analysis needs impressions split by position band and cross-checked against citation data, which is two exports and a join if you do it by hand.
GSC Page Performance Summary returns the daily series per page and Top Keywords for Site returns the query-level detail, so the histogram is one run rather than a manual pivot.
Citation Watched Pages and Citation Share then tell you which of your rising impressions correspond to AI citations, which is the part Search Console won't attribute for you.
A monthly agent:
Start (schedule, monthly) → GSC Top Keywords for Site (this period and same period last year) → Code (bucket impressions by position band, compute the shape change, and isolate last year's top 50 click-earning queries to check them separately) → Citation Watched Pages → Citation Share → Prompt LLM (report whether the impression growth sits in click-earning bands or the tail, and flag whether the top-50 set held) → Excel export → Send Email to the SEO lead.
The isolated top-50 check inside that Code node is the part worth copying. It's the difference between a report that says "impressions up, clicks down, investigating" and one that says "your fifty best queries are stable and the rest is fan-out, no action needed."
FAQ
Related answers
- Why is my organic traffic dropping while my rankings stay the same?
- What can Google Search Console's new generative-AI performance report actually show me?
- Which pages and queries explain most of my organic decline?
- How do I measure the impact of AI Overviews on my site's clicks?
Want impressions split by position band and cross-checked against citations? Start a free Analyze AI trial and connect Search Console.