September 2025 Google Search Console impressions drop; what’s really happening? 

Key takeaways:

• If your Google Search Console impressions have dropped since around 10 September 2025, this is most likely due to recent changes in how Google reports and counts impressions, not necessarily an actual decline in SEO visibility.
• Industry experts including Brodie Clark suggest this change is linked to the removal of &num=100 and adjustments in how impressions (especially bot and scraper traffic) are logged. At Found, we are monitoring these developments closely and guiding clients on how to interpret and respond to the data.

The situation: What we’re seeing

Over the past few days (since around 10-14 September 2025), many sites (ours included) have seen a notable decline in Google Search Console impressions, particularly on desktop. Alongside this:

• Average positions in GSC have generally decreased (i.e. showing better average rank). Search Engine Land+1
  
• Clicks have also dropped in some cases, though to a much lesser extent than impressions. Brodie Clark Consulting+1
  
• Third-party rank tracking tools are reporting anomalies: missing data, unexpected drops, odd shifts, especially for results beyond page 1. Search Engine Land+1
  

So far, the pattern suggests this isn’t just a seasonal content issue or a simple algorithm update – there appears to be something going on with measurement, reporting, and possibly what counts as an impression.

What are the root causes of the decline? (What the experts are saying)

Based primarily on two pieces of recent reporting (Brodie Clark’s “The Great Decoupling – Analyzing the impact of &num=100” and Search Engine Land’s “Google Search rank and position tracking is a mess right now”), here are the key issues believed to be behind what we’re observing.

What are the implications for stakeholders? 

For CMOs, Marketing Managers, Clients, this matters because:

• KPI interpretation risk: If impressions are seen to be dropping due to reporting changes or scraping reduction, not because your content or SEO is under-performing, then benchmarking against past periods will show misleading results. This could lead to a misinformed reaction – e.g. kill campaigns, reallocate budget – based on data artefacts rather than a true performance decline.
  
• Strategy adjustments: Emphasis may shift more towards clicks, conversions, engagement metrics, where possible, instead of impressions.Teams should also audit desktop and mobile trends separately to spot divergences.
  
• Tooling costs and accuracy: Rank trackers and third-party tools will need to adapt. Some may reduce how far past page 1 or 2 they reliably support, while others may introduced new pricing models to reflect the increased difficulty of tracking deeper SERP positions. If you rely on these tools for reporting, especially beyond the first page, expect some noise, gaps, or evolving feature sets as providers adjust.
  
• Client communication: Its important to give full transparency to avoid loss of trust. Clients should understand that some of what looks like “loss of visibility” may simply be measurement revisions or cleanup.
  

What we recommend doing now

Acting as the Director of SEO, here’s what I would suggest:

1. Audit your GSC time series data carefully
   
   • Segment by device (desktop vs mobile) to see where drop is worst.
     
   • Compare trends in impressions vs clicks vs CTR.
     
   • Check whether average position changes are consistent with change in impressions and clicks, or if it’s a reporting artefact.
     
2. Correlate with external / non-GSC signals
   
   • Use analytics tools (e.g. Google Analytics) for landing page traffic trends. If GSC impressions drop but landing page sessions hold steady, that suggests a reporting issue.
     
   • Look at rankings via trusted tools – allowing you to see which tools are affected. Use comparisons before/after 10-14 Sept.
     
   • Monitor conversion metrics: leads, signups, sales. If those are steady, the traffic (and prior impressions) are likely still there but not being fully reflected.
     
3. Adjust performance reporting temporarily
   
   • Set expectations for clients/leadership: this period is “noisy”.
     
   • Use broader windows and smoother data rather than currently volatile daily numbers.
     
   • Avoid reacting to short-term drops in impressions alone.
     
4. Speak with SEO tool providers / vendors
   
   • Ask how they are adapting to the removal of &num=100.
     
   • Get clarity about how their methods may now under-report ranking depth or miss certain SERP features.
     
5. Monitor updates from Google
   
   • Watch for official statements about changes in GSC reporting or SERP behaviour.
     
   • Keep an eye on community discussions; changes may roll out gradually or in phases.
     
   • Stay current on how Google treats impressions (what counts, what’s visible etc.).
     

Found’s perspective and what we are doing

At Found, here’s how we interpret this:

• We believe much of the drop in GSC impressions since ~10-14 Sept is likely due to measurement / reporting changes, particularly the removal of &num=100, which has exposed that a large amount of prior impression volume may have been inflated (or artificially high due to bot/scraper traffic or tracking tools).
  
• We do not believe (at this stage) this is a significant decline in SEO performance; rather, it’s a change in how metrics are collected/shown.
  
• Nevertheless, because metrics shape decisions, this is material: we are treating this as a risk to client confidence and campaign evaluation, and adjusting our reporting accordingly.
  

Specifically, we are:

• Flagging this issue in all client performance reports since mid-September.
  
• Shifting more weight onto conversion / business outcome metrics.
  
• Running internal tests to try to separate human vs bot traffic in GSC and analytics.
  
• Revalidating priorities: making sure that any drop in impressions that also correlates with drop in traffic/conversions is fully investigated; if not, we interpret that drift as reporting noise.