Search is still taking most of the budget in many Google Ads accounts. It is no longer taking the same share of the clicks.
That is the useful signal in a new Brainlabs analysis.
Across 31 Google Ads accounts, Brainlabs says Search campaigns accounted for 62% of clicks in January 2024. By August 2026, that share had fallen to 41%.
Search’s share of spend moved too, but less dramatically: from 86% to 71%.
Brainlabs separately looked at organic performance across 54 clients and reported a 10.5% decline in organic sessions after AI Overviews. AI-referral traffic replaced only about 5% of the lost sessions, although those AI referrals recorded a higher key-event rate: 13.4% versus 10.4% for organic.
Those numbers are worth paying attention to.
They are not an industry census.
They describe Brainlabs' own client datasets, with different samples for paid and organic analysis. European teams should treat them as a reason to audit their own traffic mix, not as proof that every advertiser is seeing the same shift.
The interesting gap is between spend share and click share
The paid-search finding becomes more useful when the two lines are considered together.
| Brainlabs metric | Jan 2024 | Aug 2026 | Change |
|---|---|---|---|
| Search share of Google Ads spend | 86% | 71% | -15 percentage points |
| Search share of Google Ads clicks | 62% | 41% | -21 percentage points |
Search click share declined six percentage points more than spend share over the period.
That does not tell us the cause.
It can reflect changing CPCs, campaign mix, Performance Max expansion, other Google inventory or differences between accounts.
But it does tell advertisers something important: budget share alone no longer describes how much traffic Search is producing.
A team looking only at spend allocation can miss a much larger change in click distribution.
Run a Search Click Mix Audit
NEMO would start by recreating the Brainlabs comparison inside the advertiser’s own account.
For each month, record:
- Search spend;
- total Google Ads spend;
- Search clicks;
- total Google Ads clicks;
- Search conversion value;
- Search conversions;
- Performance Max spend and clicks;
- other campaign-type spend and clicks.
Then calculate two shares.
Search Spend Share
Search spend divided by total Google Ads spend.
Search Click Share
Search clicks divided by total Google Ads clicks.
Do not assume they move together.
The gap between them can reveal whether Search is retaining budget while delivering a smaller portion of account traffic.
NEMO calls the difference in their movement Click Mix Divergence.
If Search spend share falls five points but Search click share falls fifteen, the media mix is changing more sharply at the traffic level than at the budget level.
That deserves investigation.
It is not automatically a problem.
Do not call a click-share decline a Search decline
A lower share of account clicks is not the same as fewer Search clicks in absolute terms.
This distinction matters.
Imagine Search generates 100,000 clicks one year and 105,000 the next. At the same time, Performance Max and other campaign types grow rapidly.
Search’s share of total clicks can fall even though Search itself grew.
So the audit needs both:
- absolute Search clicks;
- Search share of total clicks.
The same applies to spend.
Share metrics describe composition. Volume metrics describe scale.
A strong analysis needs both before concluding that Search activity has weakened.
The organic finding needs a separate denominator
Brainlabs' organic analysis uses a different client sample.
Across 54 clients, the agency reports organic sessions falling 10.5% after AI Overviews.
That is useful evidence of traffic pressure, but it should not be combined mechanically with the paid-search figures.
The paid dataset asks how click mix is changing inside Google Ads.
The organic dataset asks what happened to sessions after AI Overviews.
Those are related search trends, not one experiment.
The clean way to use them is as two separate diagnostics.
Paid teams should investigate changing campaign mix.
SEO teams should investigate whether impressions, rankings and demand remained intact while clicks declined.
AI referrals are not replacing lost organic traffic yet
The most revealing organic number may be smaller than the 10.5% decline.
Brainlabs says AI referrals replaced only about 5% of the lost organic sessions in its dataset.
That gives teams a useful metric.
NEMO calls it the AI Referral Replacement Rate:
AI referral sessions gained ÷ organic sessions lost
Suppose a site loses 20,000 organic sessions over a comparable period and gains 1,000 visits from ChatGPT, Gemini, Perplexity and other AI referrals.
Its AI Referral Replacement Rate is:
1,000 ÷ 20,000 = 5%
That makes the trade-off visible.
AI referrals may be growing quickly while still being far too small to compensate for declining organic traffic.
Growth rate alone can hide that.
Higher AI-referral conversion rates need context
Brainlabs reports a 13.4% key-event rate for AI referrals, compared with 10.4% for organic traffic.
That is interesting.
It does not mean AI traffic is more valuable overall.
A higher rate on a much smaller traffic base may still produce far fewer total conversions.
Teams should compare:
| Metric | Organic | AI referrals |
|---|---|---|
| Sessions | Measure | Measure |
| Key-event rate | Measure | Measure |
| Key events | Measure | Measure |
| Revenue | Measure | Measure |
| Revenue per session | Measure | Measure |
| Assisted conversions | Measure | Measure |
The important question is not simply which channel has the higher conversion rate.
It is whether higher-quality AI referrals are producing enough absolute business value to offset traffic lost elsewhere.
Separate visibility loss from click loss
For organic search, use Search Console before blaming AI Overviews.
Compare:
- impressions;
- clicks;
- CTR;
- average position;
- query demand;
- page groups;
- AI Overview exposure where measurable.
If rankings and impressions remain reasonably stable while clicks and CTR fall, a search-interface effect becomes more plausible.
If impressions fall sharply too, weaker demand or visibility may be contributing.
NEMO used the same logic in its French publisher traffic analysis: traffic loss and AI Overview exposure belong in the same investigation, but they should not be collapsed into a causal claim without controls.
Build one monthly search distribution table
The simplest way to make this repeatable is a monthly table.
| Measure | Previous period | Current period | Change |
|---|---|---|---|
| Search Spend Share | Measure | Measure | pp |
| Search Click Share | Measure | Measure | pp |
| Absolute Search clicks | Measure | Measure | % |
| Organic sessions | Measure | Measure | % |
| Organic CTR | Measure | Measure | % |
| AI referral sessions | Measure | Measure | % |
| AI Referral Replacement Rate | Measure | Measure | % |
| AI referral key-event rate | Measure | Measure | pp |
| Organic key-event rate | Measure | Measure | pp |
The table prevents several common mistakes.
A falling click share is not automatically falling click volume.
Growing AI referrals are not automatically replacing organic losses.
A higher AI conversion rate is not automatically greater total revenue.
And one agency’s dataset is not automatically the European market.
What European teams should take from the Brainlabs data
The Brainlabs numbers do not prove that Search campaigns are collapsing.
They do show why traditional channel reporting is becoming less useful when it stops at spend, clicks and last-click conversions.
Paid Search can keep a large share of budget while producing a smaller share of Google Ads clicks.
Organic traffic can fall while AI referrals grow from a much smaller base.
AI visitors can convert at a higher rate without replacing enough volume to offset lost search sessions.
Those patterns need different denominators.
That is the practical lesson.
Recreate the analysis with your own account and analytics data.
Track Search Spend Share and Search Click Share separately. Measure absolute Search traffic alongside both. Calculate how much lost organic traffic AI referrals actually replace. Then compare total business outcomes rather than celebrating a high percentage on a small channel.
Brainlabs has supplied a useful signal.
The stronger result is the one your own data can reproduce.