A logging error over-reported impressions in Google Search Console from May 13, 2025 onward – clicks, rankings and indexing were never affected. Google confirmed the bug on April 3, 2026 and fixed the logging by late April 2026, but only “going forward.”
- Only the window from May 13, 2025 to April 27, 2026 is affected – this historical impression data stays permanently distorted and, per John Mueller (Bluesky), will not be reconstructed.
- Since late April 2026, GSC records impressions correctly again – use that point as your new baseline.
- Your action plan: set annotations in GSC, cross-reference click data with GA4, and treat the bug window as unusable for year-over-year impression comparisons.
On April 3, 2026, Google published a notice on the official Data Anomalies Page that packs a punch: a logging error has been systematically inflating impressions in the Search Console Performance report since May 13, 2025 – for almost eleven months. Not just a minor blip. Systematically.
In my client projects at SEO Kreativ, I regularly see teams making decisions based on GSC data without questioning whether the numbers are actually accurate. Impressions rising? “Great, the strategy is working!” Impressions suddenly dropping? “Alert, we’ve been penalized!” The reality is often more mundane – and this case demonstrates that perfectly.
The kicker: clicks were never affected at any point. Anyone measuring their SEO performance primarily through clicks and conversions probably didn’t notice anything unusual. However, anyone running CTR analyses based on the inflated impressions has been working with distorted data. In this article, I’ll show you exactly what happened, how to identify the impact in your data, and what concrete steps you should take now. How to separate an impression drop from a real ranking loss, worked through on my own data, is in three tests in Search Console.
What Exactly Happened?
According to Google’s official notice on the Data Anomalies Page, this is a pure logging error. The gist: a logging error has been preventing Search Console from accurately reporting impressions since May 13, 2025. The fix will roll out over the next few weeks, and as a result, you may notice a decrease in impressions in the Performance report.
That’s exactly what happened. Per Google’s own update, logging was back under control by late April 2026. The crucial catch came from John Mueller: on Bluesky, he made clear the fix only applies “going forward” – logging was repaired from the correction point onward, but the faulty historical data was not recalculated retroactively (reported by Barry Schwartz on Search Engine Land, May 4, 2026). The affected window therefore runs from May 13, 2025 to April 27, 2026 – and it remains a permanent data gap.
A Google spokesperson additionally confirmed to Search Engine Land that a reporting error was identified that temporarily led to an over-reporting of impressions. Bug fixes are being implemented to ensure accurate reporting.
The important distinction:
| What WAS Affected | What WAS NOT Affected |
|---|---|
| Impressions in the Performance report | Clicks |
| Calculated CTR (denominator inflated) | Average position |
| Impression-based trend analyses | Actual search performance |
| Data logging (collection) | Rankings and indexing |
This is not an algorithm update. Not a penalty. Not a ranking loss. It’s a counting error in Google’s reporting system – nothing more, but nothing less either.
Timeline: From Bug to Fix
To properly contextualize the current bug, it helps to look at the chronology of GSC data issues over the past year:
| Date | Event |
|---|---|
| May 13, 2025 | Logging error begins – impressions are over-reported from this date onward |
| Sep 14, 2025 | Google disables the num=100 parameter – impressions drop 40–60% as bot impressions vanish |
| Aug 24–25, 2025 | Temporary reporting glitch with ~80% drop in clicks and impressions overnight |
| Dec 2025 | Massive delays in Performance report and Page Indexing report |
| Feb 28 / Mar 1, 2026 | Two days of Bulk Data Export missing for some properties (unrecoverable per Data Anomalies Page) |
| Apr 3, 2026 | Google officially confirms the impressions logging bug and begins correction |
| Apr 27, 2026 | End of the distorted window – logging captures impressions correctly again from here (the fix only works “going forward”) |
| May 4, 2026 | John Mueller clarifies on Bluesky that the historical data will not be reconstructed – the data gap is permanent (Search Engine Land) |
2025 was the most turbulent year for GSC data quality I’ve experienced since 2014. The num=100 shutdown, reporting delays, plus the nearly year-long impressions bug – the sheer pile-up makes it clear that blind trust in a single data source is a risk.
Context: The num=100 Drop vs. the May 2025 Bug
This is where it gets really interesting – and simultaneously complicated. Because 2025 saw not one but two effects that massively impacted your impression data:
Effect 1: The Logging Bug (from May 13, 2025) – An internal counting error at Google that over-reported impressions. Entirely on Google’s side, independent of external tools or bots.
Effect 2: The num=100 Shutdown (September 14, 2025) – Google disabled the &num=100 URL parameter that SEO tools used to fetch up to 100 search results per query. The result: bot-generated impressions for positions beyond the top 10 disappeared. Many websites saw impression drops of 40–60%. Clicks remained stable.
This means: between May and September 2025, both effects were running simultaneously – the bug was inflating impressions upward, while the num=100 shutdown then corrected them downward. In my client projects at SEO Kreativ, we observed that the September drop likely partially masked the bug effect in many cases.
The practical consequence: if you’re analyzing your impression data from the May to September 2025 period, you’re dealing with a combination of genuine growth, bug inflation, and still-present bot impressions. A clean separation is virtually impossible after the fact.
Impact on Your SEO Data
At first glance, “only impressions affected” sounds harmless. In practice, however, this metric feeds into a surprising number of analyses:
Click-Through Rate (CTR): CTR is calculated as clicks divided by impressions. If the denominator (impressions) was inflated, CTR appeared artificially lower. Anyone optimizing title tags or meta descriptions based on low CTR values over the past months may have been solving a problem that didn’t actually exist.
Visibility Trends: Rising impressions are interpreted by many teams as an indicator of growing visibility. If part of that growth was caused by the bug, reporting statements to clients or stakeholders were overly optimistic.
Year-over-Year Comparisons: This is where the real problem sits – and it’s permanent. Because Google only corrected “going forward,” impressions in the window from May 13, 2025 to April 27, 2026 stay inflated. Any year-over-year impression comparison touching this window pits correct data against distorted data – and shows a decline that isn’t real. There’s no fixing it retroactively. For reliable trends across this period, click data is the only way.
What remained stable: Clicks, conversions, and actual organic traffic in Google Analytics are unaffected by the bug. If your GSC clicks and GA4 sessions are consistent, nothing has changed about your real performance.
I still regularly see monthly SEO reports that highlight impressions as the primary success metric. This bug is a good occasion to reconsider your reporting hierarchy: clicks and conversions should always come before impressions.
Infographic: GSC Impression Anomalies 2025–2026
Action Steps: What You Should Do Now
1. Set Annotations in GSC
Mark May 13, 2025 (bug start) and April 3, 2026 (fix start) with Custom Chart Annotations. If you haven’t yet annotated the September 2025 drop from the num=100 shutdown, do it now.
2. Use Clicks as Your Primary Metric
Clicks were unaffected by the bug. For measuring SEO success, they’re more meaningful than impressions anyway. Cross-reference your GSC click data with organic sessions in GA4. If both are stable, nothing has changed about your actual performance.
3. Discard CTR Analyses from the Bug Window
Every CTR analysis from the period May 13, 2025 to April 27, 2026 is based on an inflated denominator and can’t be smoothed after the fact – the historical data was not corrected. For CTR optimizations, use data from late April 2026 onward, since logging has been running correctly again.
4. Proactively Inform Stakeholders
If you deliver SEO reports to clients or internal stakeholders: communicate the bug actively. A brief paragraph in the next report prevents awkward questions when impressions drop in the coming weeks.
5. Set a New Impression Baseline from Late April 2026
Since April 27, 2026, GSC has been capturing impressions correctly again. Use the values from that point onward as your new baseline for future impression comparisons. For the entire window before it (from May 13, 2025), use click data for trend analyses – the impressions from that period are, and will stay, unusable.
6. Diversify Your Data Sources
GSC remains the most important data source for organic search data – but it shouldn’t be the only one. GA4 for traffic and conversion data, server log files for crawling analyses, and third-party tools for ranking data together paint a more robust picture than any single source alone.
Frequently Asked Questions (FAQ)
Are my rankings affected by this bug?
No. Google has explicitly confirmed that the error only affects data logging. Your actual rankings, page indexing, and clicks on your search results are not affected. This is purely a reporting error.
Since when exactly is my impression data unreliable?
According to Google’s official notice on the Data Anomalies Page, the logging error began on May 13, 2025 and runs through April 27, 2026. All impression data in this window is over-reported and will not be corrected. Click data is unaffected and remains reliable.
Is the correction complete – and will the historical data be repaired?
The correction is complete: since late April 2026, logging has been capturing impressions correctly again. But – and this is the crux – Google only fixed it “going forward.” The distorted historical data in the window from May 13, 2025 to April 27, 2026 will not be reconstructed retroactively. John Mueller confirmed this on Bluesky (reported by Search Engine Land, May 4, 2026). So this data gap is permanent – for impression comparisons across that period, use click data as a substitute.
Is this the same bug as the September 2025 impressions drop?
No, these are two separate events. In September 2025, Google disabled the num=100 URL parameter. This caused bot-generated impressions from SEO tools to disappear, resulting in drops of 40–60%. The current bug is an internal logging issue at Google that has been inflating impressions since May 2025, independent of external bots.
What should I communicate to my clients or stakeholders right now?
Inform proactively about the bug and its impact. The key message: if impressions decrease in the coming weeks, that’s the correction of a counting error – not a performance loss. Clicks and organic traffic in GA4 are the more reliable indicators. A brief paragraph in the next monthly report is typically sufficient.
Conclusion: Clean Data Requires Clean Processes
This bug demonstrates once again: Google Search Console is an indispensable tool – but it isn’t infallible. In my client projects at SEO Kreativ, I’ve documented more GSC data anomalies in 2025 than in the five years before combined. That’s not a reason to turn your back on GSC. It’s a reason to work with it more professionally.
Specifically, that means: clicks and conversions as primary metrics, regular data exports as insurance, GA4 as a cross-check, and annotations as institutional memory. Anyone who masters these basics won’t be caught off guard by the next GSC bug – they’ll just nod and adjust their baselines.
The correction is done; impressions in your reports settled lower in late April 2026 – and the nearly year-long data gap stays as a lasting hole in your history. Now is the right time to sharpen your processes – before the next anomaly hits. And it will.
