CTR drops are frequently misdiagnosed. The default assumption is that something changed in how the site is presenting itself — title tags, meta descriptions, brand perception. But in 2025, the majority of unexplained CTR drops have nothing to do with any of those things. They're caused by changes to the SERP structure that are entirely outside your control.
Before spending time on snippet optimization, it's worth ruling out the three structural causes first. They're responsible for most of the CTR drops I've seen in the past two years.
Cause 1: AI Overviews Are Absorbing Your Clicks
This is the single largest driver of unexplained CTR decline in 2025. When an AI Overview appears above the organic results for a query, organic CTR drops dramatically. Seer Interactive found that queries with an AI Overview saw organic CTR fall from 1.76% to 0.61% — a 61% reduction. Separately, Ahrefs found a 58% click reduction in their updated 2025 analysis.
The prevalence of AI Overviews grew from 18.55% of queries in Q3 2024 to 49.92% in Q4 2025. If your CTR started declining around mid-2024 without any ranking change, AI Overviews entering your keyword space is the most likely explanation.
To check: search your target keywords manually in an incognito window. If AI Overviews are present, your CTR data is now operating in a structurally different SERP environment than it was 18 months ago.
Cause 2: The GSC Impression Bug (May 2025 – April 2026)
This one is specific and consequential. From May 13, 2025 through April 27, 2026 — 50 consecutive weeks — Google systematically over-counted impressions in Search Console. Clicks were not affected. Impressions were inflated.
Since CTR = clicks ÷ impressions, inflated impressions with unchanged clicks produces artificially low CTR. If your GSC is showing a CTR decline that started in May 2025, you may be looking at a measurement artifact rather than a real change in user behavior.
Google confirmed the bug and did not correct historical data. For any analysis covering that period, use clicks as your primary metric. CTR figures from May 2025 through April 2026 are unreliable.
Cause 3: Zero-Click Search Growth
Even setting aside AI Overviews, the baseline zero-click rate has increased. In May 2024, 56% of Google searches ended without a click. By May 2025, that had grown to 69% — a 13 percentage point increase in 12 months, according to Similarweb.
The US-specific number from March 2025 is that only 40.3% of Google searches resulted in a click to any website — down from 44.2% the prior year. This is a structural shift, not a site-level problem. If your content answers questions that Google now answers directly on the SERP, clicks will continue declining regardless of how well you optimize your snippet.
Cause 4: Ranking Erosion Without Full Ranking Loss
Average position in GSC is an average across all impressions. A page can drift from position 2 to position 4 for its primary keyword — not losing the first page, but losing significant CTR — without the overall average position moving dramatically because the long-tail queries are still anchoring the average.
To catch this: filter GSC by the specific high-value keyword and look at average position over time, not the aggregate average position across all queries. Small ranking drifts on high-volume keywords can produce large CTR drops that are invisible in the aggregate view.
Cause 5: Snippet Quality Degradation
After ruling out the structural causes, snippet optimization is worth looking at. The most common culprits:
- Title tag rewritten by Google to something less compelling than your original
- Meta description pulled from body copy rather than your specified meta — often happens when the page content changed significantly
- Competitors improved their snippets and your result now looks comparatively weaker on the same SERP
- A new competitor entered the SERP with a rich result (star rating, FAQ accordion) that makes your plain listing less prominent
Diagnosing Snippet Issues in GSC
Open Performance → Pages. Select the affected page. Look at the query list with CTR data. If CTR dropped across most queries simultaneously, it's likely a SERP-level change (positions, features). If CTR dropped on specific high-volume queries while others held, you may have a snippet competition problem on those specific SERPs.
Then search those queries manually. Compare your snippet to the others on the page. The diagnosis is usually visible — your title is longer, less specific, or less action-oriented than the surrounding results.
The Diagnostic Sequence
When CTR drops and rankings haven't obviously changed, work through this order:
- Check the date range. Does the drop start around May 2025? If yes, factor in the GSC impression bug.
- Search the affected queries manually. Are AI Overviews present? Zero-click features like Knowledge Panels or featured snippets above your result?
- Check the specific keyword position over time, not the aggregate average position. Small position drifts have outsized CTR impact in positions 1-5.
- Compare your CTR to the position benchmark. If it's below benchmark for the position, the snippet is the problem. If it's at benchmark for a lower position, it's a ranking problem.
What to Do About Structural CTR Decline
For queries where AI Overviews have fundamentally changed the traffic economics, there are three rational responses. First, optimize for AI Overview citation — brands cited in AI Overviews earn 35% more organic clicks despite the overall CTR depression. Second, de-prioritize pure-informational queries in favor of commercial or transactional content where AI Overviews are less prevalent. Third, accept lower click volumes while measuring brand impression value as a separate metric.
None of these are as satisfying as a clean fix. But a structural SERP change is not the same problem as a snippet optimization problem, and treating it as one wastes time on intervention that can't address the actual cause.
The underlying question worth asking: which of your high-impression keywords are now structurally zero-click, and which retain genuine traffic potential? Separating these two categories changes which optimization work actually matters.