On April 3, Google updated its data anomalies page. A logging error had been over-reporting Search Console impressions since May 13, 2025. Not a week. Not a month. Nearly a full year. The fix is rolling out now. Your impression numbers are going to drop.
Google's statement: "Clicks and other metrics were not affected." Only impressions. A data logging issue, nothing more.
That's true for clicks and rankings. CTR is where it gets complicated. CTR is clicks divided by impressions. Search Console computes that value from the same inflated impression count. If the denominator was wrong, so was the ratio it produced.
TL;DR
- Google over-reported impressions in Search Console from May 13, 2025 through early April 2026
- Clicks and rankings were accurate. CTR as shown in GSC was derived from inflated impressions, so it read lower than reality
- Any analysis using impressions to calculate CTR benchmarks, keyword opportunities, or share of voice should be revisited
- Disclosure came through a changelog entry. No notification, no email, no Search Liaison post
- Impressions dropping in your dashboard is the fix, not a ranking change
Google Search Console impressions bug: what happened and when
On May 13, 2025, a logging error began causing Google Search Console to over-count impressions. Google disclosed the issue on April 3, 2026, nearly 11 months later. The fix is currently rolling out.
According to Google, only impressions were affected. Clicks, average position, and CTR are listed as unaffected. CTR is derived from impressions though, which complicates that claim (more on that below). The corrected data will roll out gradually, so expect impression counts in your Performance report to decrease over the coming weeks.
Google has not said how large the inflation was, whether it was consistent across site types, or whether it compounded over time. They disclosed that it happened. The magnitude is still unknown, which is the actual problem when trying to figure out what to trust.
When impressions drop in your dashboard over the next few weeks, that is the correction. Not a ranking change. Brief your clients and stakeholders now before they see the number and assume something broke.
CTR is clicks divided by impressions. So the CTR shown in GSC was wrong too.
Google said clicks and other metrics were not affected. Click counts were clean. But CTR in Search Console is computed from impressions. If Search Console stored an inflated impression count, the CTR it showed you was suppressed.
100 clicks against 2,000 impressions = 5% CTR. The real number might have been 1,500 impressions, making the actual CTR 6.7%. That gap matters when your conclusion is "we're underperforming at position 2 and need better title tags."
A year of "our CTR is low, we need better title tags" may have been chasing a statistical artifact.
Same problem for keyword opportunity scoring. The standard filter is: high impressions, low clicks. That filter ran on bad data for 11 months. Impression share estimates from third-party tools pulling GSC data were wrong too. The scale of the error is unknown, so I can't tell you which specific conclusions broke. That uncertainty is the point.
How Google disclosed this
A changelog entry on the data anomalies help page. Not a Search Console notification. Not an email. No Search Status Dashboard alert, no Search Liaison post. The kind of page you visit when something already looks wrong.
It's possible they notified large property owners through channels I can't see. But based on what's publicly visible: 11 months of bad data in a widely used reporting tool, disclosed via a page update. Nobody in the industry raised a flag while it was happening. Either the inflation blended into normal variance, or we've gotten too comfortable trusting platform data by default.
Build your own sanity checks. Impressions diverging sharply from clicks, session data, or third-party rank trackers should trigger a question. Not just a Google disclosure to tell you something was off.
What to revisit
Rankings did not change. Clicks are accurate. If your work is grounded in those, the damage is limited. If impression trends drove your reporting, revisit these:
Impressions were always the wrong metric to lead with
This bug made visible something that was already true. Impressions count how many times a URL appeared in a SERP, including results below the fold, zero-intent queries, and keywords you'd never want. Easy to grow. Hard to connect to anything real.
They still end up as the headline number. "Organic visibility increased 34% this quarter." That's impressions. For the past 11 months it was also, to an unknown degree, a logging error.
Clicks, rankings, and conversions were clean throughout. If your work is grounded in those, you're fine. If impression trends drove your strategy or your reporting, this is the moment to fix that. Check that data anomalies page occasionally too. Apparently that's where the important updates live.
Yuval Halevi
Helping SaaS companies and developer tools get cited in AI answers since before it was called "GEO." 10+ years in B2B SEO, 50+ cybersecurity and SaaS tools clients. I run an SEO agency, so I have a direct interest in content that makes platform-dependent strategies look risky. Worth knowing.
Related Articles
ChatGPT's Favicon Is Missing from Google Search Results
The world's most-used AI tool looks like a suspicious phishing site in search results.
ResearchTech Publications Lost 58% of Their Google Traffic
We tracked CNET, Wired, The Verge, TechRadar, and six others. Combined, they lost 63 million monthly visits.
OpinionThe Keywords That Convert Don't Show Up in Any SEO Tool
We built 90 content pages driving 100K visits/month. The real conversions came from zero-volume keywords no tool surfaces.