Google Search Console Impression Drops, 2025 & 2026
Around September 10, 2025, many site owners reported large drops in Google Search Console impressions, often concentrated on desktop. Many also saw average position improve while clicks barely moved.
Two separate reporting events overlap in the same period. Sorting out which one you are looking at helps determine whether you have an SEO problem, a reporting artifact, or both.
This post was originally published in September 2025 and explained the decline with a single cause. Google has since confirmed something broader, so the explanation below has been corrected.
What Google Confirmed
Google’s Data Anomalies page carries this entry:
A logging error prevented Search Console from accurately reporting impressions from May 13, 2025 until April 27, 2026. This issue has been resolved.
External guidance
Google confirmed that a logging error prevented Search Console from reporting impressions accurately from May 13, 2025 until April 27, 2026, affecting impressions, click-through rate and average position, but not clicks.
Three details matter more than the headline.
Impressions were not the only casualty. Google says impressions, click-through rate, and average position were all affected. Clicks were not. Metrics calculated from impressions, including click-through rate and average position, inherited the problem.
The affected period lasted just under fifty weeks. It opened in May 2025 and closed in April 2026.
September 2025 sits inside it. The logging error had already begun before that month and continued after it, so it does not by itself account for a sharp change in one month. It does mean the numbers on both sides of that drop were already unreliable.
The Order These Events Happened In
The two events, and how they overlap
Confirmed by Google Observed behavior, not explained by Google
-
The logging error begins (Confirmed by Google)
Google later said that Search Console impression reporting became inaccurate from this date.
-
Google stops honoring num=100 (Observed behavior, not explained by Google)
The unofficial parameter allowed one request to return up to 100 results, and some rank-tracking and bulk-SERP tools had relied on it. Google never announced the change or explained it.
-
Impression declines reported widely (Observed behavior, not explained by Google)
Many site owners reported sharp impression declines, often heavier on desktop, while clicks changed little or held steady. The timing pointed at the parameter change, but Google did not confirm a link.
-
The logging error is fixed going forward (Confirmed by Google)
Impression counts may look lower from this point because Google corrected the reporting issue going forward. Google did not retroactively rewrite the historical data.
What Changed in September 2025
Google stopped honoring the unofficial num=100 URL parameter, which allowed a request to return up to 100 results instead of the usual 10. Some rank-tracking and bulk-SERP tools had relied on it. Without it, retrieving the same number of results could require many more requests.
The timing lined up closely with widespread impression declines. That led many people in the industry, including me, to infer that automated result retrieval may have contributed to the recorded impression pattern.
That remains one plausible reading of the pattern. It is not something Google confirmed. Google never said the parameter change caused the impression drop, never said automated requests had been counted as impressions, and never published an explanation for removing the parameter at all.
External guidance
A third-party analysis of 319 properties found that 87.7% saw Search Console impressions fall after Google stopped honoring the num=100 parameter in September 2025, and that 77.6% lost unique ranking terms.
That analysis identifies a pattern across the properties in the study. It does not show the same cause applied to every site, and the observed period overlaps Google’s broader logging-error window, which is a second explanation the study could not separate out. Review the original analysis before applying its percentage to another portfolio, especially its site selection, comparison method, and query and device breakdown.
What “Resolved” Does Not Mean
This is the part that gets misread, and it changes how you should treat your own history.
Google fixed the logging going forward. It did not go back and repair the data. John Mueller said the old data will not be fixed, which Search Engine Land reported alongside Google’s notice.
Search Console did not retroactively rewrite the affected historical data. Impression, click-through-rate, and average-position data from that period may be inaccurate and should be interpreted with caution. Exports from that period carry the same limitation.
Google said clicks were not affected by this logging issue, which makes clicks the most reliable continuous Search Console series across the affected period.
Why Average Position Can Look Better
Average position is not a fixed rank for one keyword. It is an aggregate across many impressions, spanning queries, devices, locations, and result types. The same page can sit at position 3 for one search and position 63 for another.
If lower-position impressions stop being recorded, the remaining impression mix may be concentrated closer to the top of the results, and the average can improve. No important page or query has to rank better for that to happen.
The example below simplifies Search Console’s aggregation to show why a reported average can change when the mix of recorded impressions changes.
How an aggregate average position improves without a ranking change
| Aspect | Before the reporting change | After low-position impressions stop being recorded |
|---|---|---|
| Higher-visibility queries | Position 8, 100 impressions | Position 8, 100 impressions |
| Low-visibility queries | Position 70, 900 impressions | No longer recorded |
| Total impressions | 1,000 | 100 |
| Reported average position | 63.8 | 8.0 |
Treat average-position changes during the affected period cautiously. It is one of the metrics Google says was misreported for those fifty weeks, so an improvement inside that window is not solid evidence of anything.
What Is Still Unknown
Worth stating plainly, because much of the commentary on this is more certain than the evidence supports.
- Google never confirmed that automated rank tracking caused the September 2025 decline.
- Google published no correction factor, so there is no way to reconstruct what your real impressions were.
- Google did not say every impression count in the affected period is unusable, only that the reporting was inaccurate.
- An impression count on its own has never revealed whether a site lost rankings, demand, clicks, or revenue.
How to Tell a Reporting Change From a Real Decline
Start with clicks, because clicks are the one metric Google says was not affected.
The diagnostic process, five checks in order
-
Open Performance, then Search results
Compare the affected period against the previous equivalent period.
-
Read clicks before average position
Position is an aggregate and was itself misreported. Clicks were not.
-
Split by device
A desktop-heavy drop with relatively steady mobile traffic is consistent with the September 2025 pattern.
-
Switch from Queries to Pages
Find whether the loss is spread thin across low-position pages or concentrated on pages that matter.
-
Check what happened off the chart
Check leads, calls, bookings, and revenue over the same dates. This is the step that tells you whether the change affected the business.
| What you see | What it more likely means | What to check next |
|---|---|---|
| Impressions down, clicks steady | A reporting or impression-mix change | Device split, landing pages, conversions |
| Desktop down hard, mobile steady | Consistent with the September 2025 pattern | Query and position distribution |
| Average position improved | Low-position impressions left the average | Page-level clicks for priority queries |
| One important page lost clicks | A real page-level problem | Indexing, rankings, competitors for that query set |
| Clicks, leads and revenue all fell | A real visibility or demand problem | Indexing, rankings, demand, site changes, competition |
| The change starts in late April 2026 | Possible effect of Google’s reporting correction | Check Google’s data-anomaly notice before diagnosing an SEO decline |
Why Year-Over-Year Comparisons Mislead Right Now
Comparing May 2026 with May 2025 places a post-fix month beside one Google later said was affected by inaccurate reporting. A decline there may be the correction rather than a loss of visibility.
The same applies to any rolling twelve-month chart that spans the period. Until your comparison window contains only post-April-2026 data, treat impression trends as directional and annotate the chart.
Why I Tell Clients Not to Manage Impressions
I treat impressions as a coarse signal rather than a number to manage. They answer one narrow question well: is Google showing this page often enough to generate a meaningful level of visibility? That is genuinely useful when deciding whether something needs improving or replacing.
Past that, impressions reward the wrong things. A page can gain thousands of impressions by appearing low in results for queries that do not lead to useful traffic, and the chart reads as progress.
The last fifty weeks are a reminder not to build reporting around impressions alone. They are useful as a directional sign of search visibility, especially when you compare pages, queries, and devices. They are not a business outcome, and they become misleading when they are treated as one.
The numbers worth watching instead:
- Organic clicks. Are people still arriving from Google?
- Organic conversion rate. Of the people arriving from Google, are more or fewer taking a valuable action?
- Qualified leads. Are those visitors calling, submitting forms, or booking?
- Revenue from organic search. Is any of it turning into business?
- Landing page performance. Which pages lost clicks, not just impressions?
- Query and device patterns. Did branded, non-branded, desktop, or mobile move differently?
How to Report This to a Client or Leadership Team
If you produce monthly reporting, the useful move is to annotate the period rather than explain the dip every month.
Reporting note
Search Console impressions in this report cover a period affected by two known reporting changes: Google stopped honoring the num=100 parameter in September 2025, and Google has confirmed a logging error that affected impression reporting from May 13, 2025 until April 27, 2026. Google did not retroactively rewrite that historical data.
We are treating impressions as directional for this period and evaluating organic performance through clicks, qualified leads, conversions, revenue, and the performance of priority landing pages. If those measures decline, we will investigate rankings, indexing, search demand, site changes, and competition.
Common Questions
Did Search Console impressions drop in September 2025?
Many site owners reported sharp declines around that month, usually heavier on desktop than mobile. The timing aligned with Google no longer honoring the unofficial num=100 parameter, which some rank-tracking and bulk-SERP tools had used. The connection is a plausible industry explanation, not a Google-confirmed cause.
Google later confirmed a separate logging issue affecting impression reporting from May 2025 through April 2026, which means data from this whole period needs extra caution.
Does an impression drop mean my rankings fell?
Not on its own. Impressions move with reporting changes, query mix, search demand, devices, and result features, as well as with rankings.
Check clicks, priority landing pages, conversions, calls, and revenue before treating it as a traffic problem. If those held steady while impressions fell, the impression change may have limited business impact.
Why did average position improve while impressions dropped?
Average position is an aggregate, not a fixed rank. When impressions from lower-ranking queries stop being recorded, the remaining impressions sit closer to the top of the results and the average improves.
That can happen without a single important keyword or page moving up.
Should I try to recover the lost impressions?
Do not rewrite pages to chase a historical impression count. Work out first whether clicks, leads, revenue, or priority landing pages declined.
If they held steady, treat the impression change as reporting context. If they fell too, do not assume the reporting issue explains the decline. Investigate indexing, rankings, search demand, site changes, and competition.
How This Was Sourced
Google’s public Data Anomalies notice is the only primary source here, and it covers the logging error alone. The September 2025 explanation rests on industry reporting and observed timing, which is why it is labeled as observed rather than confirmed throughout.
The supporting observations include a third-party analysis of 319 websites and one anonymized client property we track. The client example is illustrative and is not presented as a representative sample. Both identify patterns. Neither establishes the cause of an impression decline for any individual site.