Long exposure of a motorway at night with steady streams of white and red light trails

PPC Strategy

6 min read

2 Nov 2026

How to diagnose falling Google Ads impressions when demand hasn’t changed

A fall in Google Ads impressions usually triggers the same sequence of questions from leadership: has demand dropped, has a competitor moved in, or has something broken in the account? Each of those has a different fix, and acting on the wrong one can cost months of performance. Over the past year a fourth explanation has become much more common. Demand is steady, the account is healthy and competition is broadly unchanged, but the number of ad opportunities Google offers against those searches is shrinking as more queries are answered by AI Overviews and AI Mode.

Google’s own results put this in context. In its third-quarter 2026 report, search query growth was the highest it had been in five years, while search revenue growth slowed, partly because AI Mode currently earns less from ads than traditional results. More searching with less monetisation is exactly the pattern that shows up in accounts as stable demand and falling impressions. This guide sets out how to confirm whether that’s what is happening to you.

Start by separating demand from supply

Impressions are the product of two things: how often people search for what you sell, and how often Google offers an ad opportunity that you’re eligible for and win. Most diagnostic mistakes come from treating those as one number.

For demand, use sources that sit outside the paid auction. Search Console impressions for your core non-brand queries are the most useful, because they come from the same searches your ads target. Google Trends is helpful for direction and seasonality, and Keyword Planner volumes give a rougher, rounded view. Compare like-for-like periods, ideally year on year, so that seasonality doesn’t distort the picture.

For supply, the paid metrics matter. If organic impressions for a set of queries are flat or rising while paid impressions on the same themes are falling, and your bids, budgets and targeting haven’t changed, the most likely explanation is that fewer ad opportunities are being created for those searches.

The Google Ads metrics that tell you why

Search impression share is the starting point, but the two lost impression share columns are where the diagnosis happens.

Search lost IS (budget) tells you how many eligible impressions you missed because spend ran out. If this rises, the problem is budget pacing, often because CPCs have risen, rather than supply.

Search lost IS (rank) tells you how often you were eligible but your Ad Rank wasn’t high enough. A rise here points to competitors bidding harder, lower expected click-through rate or quality, or thresholds increasing for the remaining positions.

The useful insight comes from reading these together with raw impressions. Impression share is calculated against the impressions you were eligible for, so if the total pool shrinks, your share can stay level even though impressions fall. Stable impression share with falling impressions, alongside stable demand, is one of the clearest signs that the auction itself has got smaller.

Top and absolute top impression rate are the next check. When fewer ads are shown on a results page, the remaining positions are more contested, which tends to push up average CPC and shift where you appear. The Auction Insights report helps rule out a competitive explanation: if overlap rate and outranking share for your main competitors haven’t moved much, a new or more aggressive competitor is unlikely to be the cause.

Finally, segment everything by device and by campaign type. AI-led results behave differently on mobile and desktop, and Search, Shopping and Performance Max campaigns can be affected to different degrees.

How ads in AI Overviews and AI Mode currently work

It’s worth being precise here, because the details vary by market and change frequently.

According to Google’s help documentation, ads can appear above or below AI Overviews in all markets where AI Overviews are available, including the UK, using normal auction ranking. Ads within AI Overviews are more limited. At the time of writing they’re available in English in a list of countries that includes the US, Canada and Australia but not the UK, and Google excludes sensitive verticals such as finance, healthcare, alcohol and gambling.

Google has also said that ads in AI Mode are expanding internationally, with the UK among the markets planned for Q4 2026. In both surfaces, ads are matched using the query and the content of the AI response. Existing Search, Shopping and Performance Max campaigns are eligible automatically, but Google recommends broad match or AI Max for Search, because exact and phrase match keywords will match far fewer of these queries.

The practical limitation is reporting. You can’t choose AI Overview or AI Mode as a placement, and performance there is currently reported alongside other top-of-page ads with no separate breakdown. That means you can’t directly measure how much of your impression loss has moved into AI surfaces. You have to infer it from the pattern described above, and from search terms that look longer and more conversational than your usual queries.

What to do with the diagnosis

If the evidence points to a shrinking auction, the decisions are more strategic than technical. Cutting bids to protect efficiency can reduce visibility further on the positions that remain. Restructuring a healthy account in response to a market-wide change rarely helps.

The more useful questions are about eligibility and measurement. Is it worth testing broad match or AI Max in controlled campaigns so that you’re eligible as AI Mode ads arrive in the UK? Is your conversion data good enough for Smart Bidding to value those broader queries properly? And are you tracking business outcomes closely enough to tell whether the traffic you’re still getting is more or less valuable than before?

Explaining it to leadership

Leadership teams generally respond better to a clear framework than to platform terminology. A short monthly view covering three lines tends to work: demand (organic impressions or Trends for core themes), paid supply (total impressions and impression share) and commercial outcome (leads, sales or revenue from paid search). When the first and third are stable and only the second is falling, the conversation shifts from “what’s wrong with the account?” to “how do we position for a changing search results page?”, which is a much more productive discussion.

It also helps to be honest that some of this can’t yet be measured. Google hasn’t released placement-level reporting for AI surfaces, and until it does, part of the analysis is informed inference rather than hard data. Saying so openly builds more trust than presenting a confident explanation that later turns out to be incomplete.

If you’d like a second opinion on what your own impression data is showing, Propel is always happy to take a look.