The State of UK Google Ads Spend: Q3 2026

Key takeaways

• At least 69% of active accounts have a conversion-tracking fault serious enough to distort the numbers they optimise on, up from 62% last quarter, and that is a floor because the API cannot see Consent Mode, web Enhanced Conversions or tag errors.
• The median account put 25% of its search spend into keywords that converted nothing or cost more than three times its own average cost per action, a fall of almost a quarter from 34% in Q2.
• In the 20 accounts measurable in both quarters the median change was zero points, with nine improving and seven getting worse, so the book-level fall is not yet a turnaround in the same accounts.
• Around £1 in every £6 of measured search keyword spend bought no conversions at all, down from roughly £1 in every £5 last quarter, and that measure is only as reliable as each account’s own conversion data.
• The median account’s ads showed in fewer than one in three eligible searches, losing 40% to ad rank and 18% to capped budgets, with none of the 22 measurable accounts clean on both leaks.
• Like for like across 21 accounts, cost per click fell 15% year on year while conversion rates rose 67% and cost per action fell 49% on spend that was up 3%; the blended 20% rise in click prices is a mix artefact, not a market price.
• Performance Max took 39% of spend, up from 33%, and reported 5.2 against 2.2 for Search across nine like-for-like accounts, on the same tracking that is faulty in most accounts.
• The median ecommerce account reported a return of 2.5 times spend, up from 1.4 last quarter, with the bottom quarter reporting about 90p back for every £1 spent and the top tenth reporting 6.3.

Every quarter we take an anonymised look across the live UK Google Ads accounts in our audit programme. We are not hunting for reasons to spend more. We are looking for money that is already leaking. This is the third report in the series, the first covered Q1 2026, so we now have three quarters of the same method applied to the same book. We report medians rather than averages throughout, because spend is concentrated: the largest five accounts carry 58% of the spend we analysed and the largest ten carry 89%. You can read the Q2 2026 edition for the baseline.

The short version for Q3 2026 (July to September): budgets edged up and the money worked harder. Wasted search spend fell at book level, from a third of search budget to a quarter. Visibility improved too, with less lost to ad rank than last quarter. One number moved the wrong way, and it is the number everything else rests on: the conversion-tracking fault rate rose again, to at least 69% of active accounts. Over £480,000 of live UK spend sits behind these figures, on top of 3,000+ audits since 2017.

Source: PPC Geeks, Q3 2026 (July to September). Figures are floors or medians. See the methodology below.

UK Google Ads spend: the headline findings

For Q3 2026 (July to September), across the live UK accounts we analysed:

• At least 69% of active accounts have a conversion-tracking fault serious enough to distort the numbers they optimise on. Up from 62% in Q2 2026, and still the floor.
• Wasted search spend fell by almost a quarter in three months. The median account now puts 25% of its search spend into keywords that do no real work, down from 34% last quarter.
• The median account’s ads show in fewer than one in three eligible searches. Ad rank takes 40% of eligible impressions, roughly twice what capped budgets take (18%).
• Like for like, cost per click fell about 15% year on year. Conversion rates rose 67%, so efficiency improved.
• About 39% of UK Google Ads spend goes to Performance Max, and its reported return still grades its own homework.
• The median ecommerce account reports a return of only about 2.5 times spend.
• We analysed over £480,000 of live UK spend in a single quarter, on top of the 3,000 or so audits we have run since 2017.

Finding 1: The tracking problem got worse, not better

In Q2 we reported that at least 62% of active accounts had a conversion-tracking fault serious enough to distort the numbers they optimise on. Same method, same probe, one quarter later: 69% of the 26 active accounts we could judge. That is not a plateau and it is not progress. More than two in three accounts in our sample are steering by instruments that are wrong, and most of them are letting automated bidding optimise on the same wrong data.

As always, treat that figure as a cautious floor rather than a ceiling. The API cannot see Consent Mode configuration, web Enhanced Conversions or tag errors on the site itself. Those are three of the most common ways tracking breaks, and none of them are in this count. The honest number is higher than 69%. We just cannot prove how much higher from the outside, so we publish the floor.

Two details explain most of what we see. In 73% of accounts the primary conversion action still uses last-click attribution, so everything above the final click goes uncredited. And in 54% of accounts, site activity is counted as a conversion: page views, scrolls, button clicks and similar signals sitting in the same column as actual sales and enquiries. More than half of the accounts we analysed are telling Google that someone browsing is worth the same as someone buying.

This matters because automated bidding is obedient. Smart Bidding will buy more of whatever you tell it to value, efficiently and at scale. Feed it site activity and it will get very good at finding people who scroll. The account will look busy, the cost per action will look cheap, and the bank balance will disagree. Every other finding in this report rests on each account’s own conversion data, which is exactly why this one comes first.

Finding 2: Wasted search spend fell, but not in the same accounts

Here is the quarter’s best news. Across the accounts we could measure properly, the median account put 25% of its search spend into keywords that converted nothing or cost more than three times that account’s own average cost per action. Last quarter the same measure read 34%. That is a fall of almost a quarter in three months, and a genuinely useful move in the right direction.

The spread improved with the middle. Nine of the 20 measurable accounts, 45%, still run above 30% waste, down from 12 of 22 last quarter. So the problem is smaller, not solved: nearly half the accounts we could measure are still putting more than three pounds in every ten of search spend into keywords that are not pulling their weight.

The sharpest edge is the money that bought nothing at all. Spend on keywords that converted absolutely nothing came to 16% of the keyword spend we measured, down from 21% in Q2. Put plainly: around £1 in every £6 of measured search keyword spend bought no conversions whatsoever, against roughly £1 in every £5 last quarter. Better. Still not good.

Now the caution, because the book-level fall is not the whole story. In the 20 accounts we could measure in both quarters, the median change was zero points. Nine accounts improved and seven got worse. So the headline drop owes a good deal to which accounts were measurable this quarter, not to a sweeping clean-up across the same advertisers. We would rather report that plainly than sell a turnaround the panel does not support.

There is also a twist worth naming. This measure rests entirely on each account’s own conversion data, and Finding one says that data is faulty in at least 69% of active accounts, with 54% counting site activity as a conversion. Loose conversion definitions make dead keywords look alive, which means tidier tracking usually admits to more waste, not less. And the usual honest caveat applies: this covers search keywords only, and some of that spend defends a brand name or assists a sale that last-click reporting never credits. We would not call all of it pure waste. We would call a quarter of search budget worth a morning of anyone’s time.

Everyone wants to blame the auction. In the accounts we can compare properly, clicks got 15% cheaper year on year. The thing that actually got worse was the tracking, and that is the one problem nobody can blame on Google.

Dan Trotter, Head of PPC, PPC Geeks

Finding 3: Where your eligible impressions actually go

Last quarter the median account lost 50% of its eligible search impressions to ad rank. This quarter that figure is 40%. Google ran the auction, considered the ad, and left it out, because other advertisers’ combination of bid and ad quality won. Losses to capped budgets were unchanged at 18%. The improvement came entirely from the rank side.

Add it up and the median account’s ads appeared in 32% of the searches they were eligible for: fewer than one in three. This is not a rare failure mode. Eighteen of the 22 measurable accounts lose more than 30% of their eligible impressions to rank, and 15 of the 25 accounts with meaningful spend lose more than one impression in ten to capped budgets. Nothing inside the account screams about either of these, which is why they survive for years.

Two honest notes before anyone panics. First, eligible counts only the auctions Google considered the account for, and broad targeting widens that pool, so a low share can partly reflect wide match settings rather than weak ads alone. Second, losing to rank means outranked or priced out, a mix of bid and ad quality, which is precisely why it is fixable: tighter targeting, better creative and smarter bids all move it. Rank still takes roughly twice as much of your eligible impressions as capped budgets do. If the only thing you watch is the limited by budget flag, you are watching the smaller leak.

Put findings two and three together and the overlap is stark. Nine of the 22 measurable accounts, 41%, are wasting more than 20% of their search spend while also losing more than 10% of their impressions to capped budgets: paying for the wrong clicks and unable to afford the right ones at the same time. Fourteen of the 22 are wasting more than 20% while losing more than 30% to rank. Every one of the 22 has at least one of these leaks. Not a single account is clean on both.

One more finding that changes how you should audit. Across our accounts, waste and rank-lost visibility are statistically unrelated. Fixing one does nothing for the other. They are two separate leaks needing two separate fixes, and a review that checks only one is half a review. The win is still moving the money you already spend, not adding to it.

Finding 4: The truth about rising click prices

“Google Ads is getting more expensive” is the industry’s favourite headline, and on the lazy numbers it looks true. Blend our whole book together and the average cost per click is 20% higher than a year ago. We are not going to quote that as a fact about the market, because it is not one. A blended figure mixes in new accounts, departed accounts and shifts between sectors and channels. It measures the composition of our book, not the price of a click. Treat it as an artefact to be explained, not evidence to be cited.

Compare the same accounts like for like, Q3 2026 against Q3 2025, and the picture inverts. Across the 21 accounts present in both years, cost per click fell 15%. Conversion rates rose 67%. Cost per action fell 49%. Spend was broadly steady, up 3%. The same advertisers paid less per click, converted far more of them, and halved what each result cost. If your own costs jumped this year, the honest first question is not what Google is charging. It is what changed in your account.

One seasonal note, because it will explain what some advertisers felt this quarter. In the 25 accounts we can compare across both quarters, spend rose 5%, conversions rose 3%, cost per click moved 1% lower and conversion rate slipped 4% against Q2. That is ordinary summer auction movement. It is not inflation and it is not a trend, and quarter-to-quarter price moves should never be read as either. The like-for-like year-on-year comparison is the only figure here that says anything honest about what clicks cost.

The longer arc supports the calm reading. Across the whole book, and remembering that the book changes shape each year, conversion rates sat around 3% in 2024, around 3% again in 2025, and around 5% this year. Cost per action was £18 in 2024, £24 in 2025 and £16 this year. Blended click prices did rise across those three years, from about 62p to about 69p to about 83p, and the cost of a result still ended up lower than it was two years ago. Efficiency has been outrunning click prices.

Finding 5: Performance Max still grades its own homework

Performance Max took 39% of all the UK Google Ads spend in our sample this quarter, up from 33% in Q2. Eleven of the 32 accounts run it. So roughly one account in three is absorbing close to two pounds in every five of the spend we analysed. Last quarter’s small dip in share did not become a retreat.

On reported numbers it remains the star of the show. Across the nine accounts running both channels in Q3 2026, Performance Max reported a return on ad spend of 5.2 against 2.2 for standard Search. Read as a headline, that looks like a settled argument.

The asterisk has not shrunk. Brand and Shopping cannibalisation flatter that reported return, because conversions that would have arrived anyway get counted as new. The biggest single category of its activity is the one Google labels as unknown. The platform cannot demonstrate that its conversions are incremental. And it rests on exactly the same conversion tracking that is faulty in at least 69% of active accounts, with site activity counted as a conversion in more than half of them. Nearly two pounds in five of the budget we analysed flows into a channel that marks much of its own homework. That is not an argument for switching it off. It is an argument for measuring it from outside the platform before you give it any more.

Finding 6: The ecommerce reality check

Across the 16 accounts tracking revenue, the median reported return in Q3 2026 was 2.5 times spend. Last quarter the same measure read 1.4. That is a real improvement in what the middle of the pack reports getting back, and it lands a long way below the glossy benchmarks the industry likes to quote.

The median hides a brutal spread. The bottom quarter of these accounts report getting back around 90p for every £1 they spend, a reported return of 0.9. The top quarter report 4.7 and the top tenth report 6.3. The distance between a well-run account and a poorly run one is not a few percentage points. It is the difference between a channel that funds the business and one that quietly drains it.

Read all of that carefully, though, because reported return is a floor, not a verdict. It ignores margin, repeat purchases and anything that happens offline or over the phone. It rests on conversion tracking that this report says is faulty in at least 69% of active accounts. Untracked revenue makes a healthy account look unprofitable, and loose conversion definitions make a weak one look fine. If your reported return sits anywhere near that 2.5 median, check the measurement before you condemn the marketing.

What this means for your Google Ads spend

Three quarters of the same method point to the same order of operations, and this quarter sharpens it.

1. Fix the tracking first. At least 69% of active accounts have a serious fault, up from 62% last quarter, and 54% count site activity as a conversion. Every optimisation you make on top of that compounds the error.
2. Bank the waste you have already cut. The median account still puts 25% of its search spend into keywords that do no real work, and nine of 20 measurable accounts are above 30%. Move that money onto the keywords that demonstrably sell.
3. Chase rank, not just budget. The median account loses 40% of eligible impressions to ad rank against 18% to capped budgets, and the two leaks are unrelated, so fixing one does nothing for the other.
4. Ignore the CPC panic. Like for like, click prices fell 15% year on year while conversion rates rose 67%. Blended click-price headlines describe the mix of accounts, not the market.
5. Audit Performance Max from outside the platform. It took 39% of spend this quarter, up from 33%, and reports 5.2 against 2.2 for Search in like-for-like accounts. That number rests on the same distorted tracking as everything else.

How this compares with the wider market

Our figures are first-party, pulled from live accounts rather than from a survey of what advertisers think is happening, which is both their strength and their limit: they describe the live UK accounts we analysed, not the whole UK market. For a broader read on direction, the IPA Bellwether report tracks UK marketing budgets quarter by quarter and is the standard reference for whether spend is being revised up or down. On waste, independent estimates of how much search spend achieves nothing commonly land somewhere in the 15 to 40% range, which is a wide band because everyone defines waste differently. Our 25% median sits squarely in the middle of it, down from the upper half of the band last quarter. If you are planning on a number, the middle of that range remains the honest place to start.

About this data

These figures come from the live UK Google Ads accounts in PPC Geeks’ audit programme, analysed anonymously and in aggregate for Q3 2026 (July to September). That is 32 enabled accounts and 25 with spend in the quarter, depending on the measure, totalling over £480,000. Year-on-year comparisons are like for like across the 21 accounts present in both years. The waste figure covers the 20 accounts with enough conversion data to measure; the impression-share figures cover the 22 accounts running Search. The ecommerce figure covers the 16 accounts tracking revenue. The tracking probe judged the 26 accounts active in the last 30 days of the quarter. We name no client, and no one can identify a single account. We never publish a figure drawn from fewer than 8 accounts; where a quarter is too thin we widen the window and say so. Several of our numbers are deliberately cautious floors, and where the account data alone cannot reveal a fault, we have said so and given the conservative figure.

PPC Geeks is a UK PPC agency working across Google, Microsoft, Meta, Amazon and TikTok. If you would like an honest look at where your own account stands, tracking included, our free Google Ads audit does exactly that. No pressure to spend more. Just the truth about the money you already spend.

Frequently asked questions

Is Google Ads getting more expensive in the UK?

Not in the accounts we can compare properly. Across the 21 accounts present in both Q3 2025 and Q3 2026, cost per click fell 15% year on year, conversion rates rose 67% and cost per action fell 49%. A blended figure across our whole book shows click prices up 20%, but that mixes in new and departed accounts, so it measures the shape of the sample rather than the price of a click.

How much of my Google Ads budget is being wasted?

In the live UK accounts we analysed, the median account put 25% of its search spend into keywords that converted nothing or cost more than three times its own average cost per action, down from 34% in Q2 2026. Nine of the 20 measurable accounts were still above 30%. Some of that spend defends a brand or assists sales that last-click reporting never credits, so not all of it is pure waste, and the whole measure depends on each account’s own conversion data being sound.

Why don’t my ads show when my budget isn’t limited?

Because ad rank takes far more of your eligible impressions than budget does. The median account in this quarter’s data lost 40% of eligible search impressions to rank and 18% to capped budgets, and showed in just 32% of the searches it was eligible for. Rank is a mix of your bid and your ad quality, so tighter targeting, better ads and smarter bidding all move it.

Is Performance Max actually better than Search?

On reported numbers it looks it: across the nine accounts running both in Q3 2026, Performance Max reported 5.2 return on ad spend against 2.2 for Search. Those numbers come with heavy caveats, including brand and Shopping cannibalisation, a large unknown category, no proof that its conversions are incremental, and the same conversion tracking that is faulty in at least 69% of active accounts. It took 39% of the spend we analysed this quarter, up from 33%, so it is worth verifying outside the platform.

What return on ad spend should a UK ecommerce account expect?

Across the 16 accounts tracking revenue in our Q3 2026 sample, the median reported return was 2.5 times spend, up from 1.4 the previous quarter. The spread is wide: the bottom quarter reported about 90p back for every £1 spent, while the top quarter reported 4.7 and the top tenth 6.3. Reported return is a floor, not a verdict, because it ignores margin, repeat purchases and offline sales, and it rests on tracking that is broken in most accounts.