Google's Impression Bug Ran for 11 Months Before Anyone Noticed
If you pulled Search Console impression data in the last year to build a report, make a case for more budget, or demonstrate SEO progress to someone who signs checks, part of what you showed them was wrong.
Google disclosed today that a logging error in Search Console has been over-reporting impressions since May 13, 2025. The bug inflated impression counts in the Performance report for nearly eleven months before Google identified it. A fix is now rolling out, and the company says the correction will take several weeks to fully propagate.
Google updated its Data Anomalies in Search Console support page with a note that users "may notice a decrease in impressions in the Search Console Performance report." A spokesperson told Search Engine Land: "We identified a reporting error in Search Console that temporarily led to an over-reporting of impressions from May 13, 2025 onward."
Clicks were not affected. Neither were other metrics. Only the impression logging was wrong.
The Second Impression Correction in Seven Months
This is not the first time Search Console impression data has turned out to be inflated in recent memory. In September 2025, Google quietly removed support for the &num=100 parameter, a URL setting that had allowed automated tools and crawlers to pull up to 100 search results per page. When that parameter disappeared, so did the phantom impressions those tools were generating. According to Nytro SEO analysis, 77% of tracked sites lost keyword visibility overnight, and non-human impressions had been inflating counts by 20-50% depending on the niche.
That correction happened in September. This new logging bug started in May 2025, which means there was a roughly four-month overlap where impressions were being inflated by both the crawler noise and the logging error at the same time. Nobody flagged the logging bug because the September parameter change created so much noise that any remaining inflation was impossible to isolate.
If you were comparing your post-September "clean" baseline to the months before it, even your corrected numbers were still slightly wrong. And if you walked a stakeholder through a "here is our real impression data after the correction" presentation last fall, you will probably want to revisit that conversation.
Clean Up Your Reports Before Google Cleans Up Theirs
The good news: clicks remained accurate the entire time. If your reporting leans on click-through rates, actual traffic, or conversion data from organic search, nothing changes. Those numbers were real.
The bad news is that impressions are probably in more of your reporting than you think. If you use any third-party SEO tool that pulls from the Search Console API, those numbers were also inflated. Ahrefs, Semrush, whatever dashboarding tool your agency uses to generate pretty charts for clients. Any tool downstream of the Performance report was working with the same bad data.
Impressions tend to be the first metric in an SEO report. They are the biggest number, the one that trends upward most reliably, and the one most commonly used to tell a growth story. A lot of quarterly business reviews include a Search Console impressions chart with a green arrow pointing up. Some portion of that green arrow, for the last eleven months, was a logging error.
If I had client reports or stakeholder decks that relied on impression data from the last year, here is what I would do:
Pull your May 2025 through March 2026 data now, before the correction finishes rolling out. Screenshot or export it. You will want the "before" numbers available when someone asks why impressions dropped.
Prepare a one-paragraph explanation. Something like: "Google disclosed a logging error that inflated impression counts in Search Console since May 2025. Our click and traffic data was unaffected. The impression correction is now rolling out and we expect our baseline to normalize within a few weeks."
Shift your primary reporting metric. If impressions sit at the top of your monthly or quarterly report, this is a good reason to move clicks, CTR, or indexed-page performance into that spot instead. Impressions have always been the most directional and least actionable number in the Performance report. They tell you roughly how visible you are. They do not tell you if that visibility is converting to anything.
Re-baseline after the correction completes. Google says the fix will take several weeks. Once it stabilizes, pull a fresh 28-day snapshot and use that as your comparison point going forward.
Impressions Were Already the Weakest Metric in Your Reporting Stack
I think this bug is going to create some short-term panic and then quietly improve how SEO teams communicate results. And honestly, that is probably a net positive.
Impressions in Search Console have always been a fuzzy metric. They count any time your URL appeared in a search result, regardless of whether the searcher scrolled far enough to actually see it. A page ranking in position 8 gets an "impression" even if nobody scrolled past position 3. That is how the metric is designed, and it has been that way since Search Console launched. It is not wrong, exactly. It is just not measuring what most people assume it measures.
The more useful number has always been clicks. Clicks mean someone saw your listing and decided it was worth clicking on. That is a real signal of demand. Impressions are an estimate of opportunity. Clicks are a measure of whether you captured any of it.
From what I have seen across different verticals, the teams that put impressions at the top of their SEO reporting tend to be the ones optimizing for visibility rather than revenue. That is a fine goal for brand awareness work, but if your SEO program is supposed to drive pipeline or actual sales, impressions were always the wrong number to lead with. This bug just makes that harder to ignore.
The metrics worth tracking have not changed: clicks from organic search, click-through rate by query cluster, pages generating actual sessions, and where possible, the revenue or pipeline those sessions produce. Those are the numbers that survive a logging bug, a parameter change, and whatever the next Search Console data issue turns out to be.
I would estimate that for most sites, the correction will land somewhere in the range of 5-15% lower impression totals compared to what Search Console was showing. Some verticals with heavy tool-based monitoring probably saw more inflation than that. You will not know exactly how much was the bug versus real impressions until the fix fully rolls out and stabilizes.
Getting Ahead of the Drop
Google said the correction will roll out over several weeks, which means impression counts will decline gradually rather than fall off a cliff. That is actually easier to manage from a reporting perspective, because a sudden drop tends to set off alarms. A gradual normalization gives you time to explain.
The worst scenario is if you are in the middle of a quarterly review cycle and someone pulls the data while the correction is still propagating. The numbers will look unstable. Get ahead of it. Send the note now, not when your CMO Slacks you asking why organic visibility dropped 10%.
If this feels like deja vu, it should. Between the September parameter removal, a crawl stats UI bug in March, missing bulk export data in February, and now this logging error, Search Console reliability as a reporting tool has taken a few hits in quick succession. None of them are catastrophic individually. But stacked together, they are a solid argument for not anchoring your entire SEO narrative to a single Google-provided dashboard.
Search Console is still the best free tool for understanding how Google sees your site. I am not suggesting you stop using it. But cross-reference with your analytics platform. Track actual revenue from organic. Use Search Console as one input among several, not the foundation under everything.
The impressions will settle. Whether your stakeholder trust does too depends on when you send the email.
Notice Me Senpai Editorial
