Something Inc.LoginSchedule a free consultation
ANALYTICS

Google DMA search changes: Europe is now your control group

Google rebuilt European search results today and told Reuters it is the largest quality reduction in the company's 29 year search history. Whatever you think of that claim, it just handed you a dated, geographically bounded natural experiment, and your current reporting will blend it into nothing.

ANALYTICSMEASUREMENTTREND READ

Google changed what European search results look like today, and it did something unusual on the way out the door. It pre-announced that the change makes the product worse. A Google official told Reuters the revamp represents the largest reduction in quality of service at the world's most popular search engine in its 29 year search history.

Set aside, for one minute, whether you believe that. The Google DMA search changes that went live on September 8, 2026 have a property that is far more useful to you than the argument about them: they stop at the EU border. Users outside Europe see none of it. There is a date, there is a boundary, and there is an untouched population running in parallel. That is a natural experiment, and search almost never gives you one.

Sep 8, 2026
date the revamped results went live, Europe only
460M EUR
fine issued by the European Commission in July 2026, roughly 534 million dollars
5%
of total worldwide turnover, the ceiling on periodic penalty payments for non-compliance
29 years
span Google says this quality reduction is the largest across, per its statement to Reuters

Most teams with European traffic will notice something in the October numbers, shrug, and attribute it to whatever core update is nearest. That is the failure worth preventing, and you have about three weeks to prevent it.

What actually changed in European results on September 8

The mechanics are narrower than the headline suggests, which is exactly why they are easy to under-react to. The change concerns how Google presents vertical search services, the comparison sites and specialist engines that the Digital Markets Act says Google must not disadvantage relative to its own equivalents.

Under the new arrangement one specialized search engine appears at the top of the results, with two more shown below it in reduced detail. Beneath those sits a carousel of hotels, airlines and restaurants. Placement is decided by Google's algorithm rather than by the prominence rules that governed the previous layout. The detail most people are skipping past is that real time prices and key features have been stripped out of the carousel listings.

SURFACEBEFORE SEPTEMBER 8AFTER SEPTEMBER 8DIRECTIONAL BENEFICIARY
Top of resultsGoogle's own vertical units held prominence by ruleOne specialized search engine placed first, algorithmically rankedComparison and vertical search services
Secondary slotsLimited standing for third party verticalsTwo further vertical search services, shown in reduced detailAggregators with existing authority
Travel carouselHotels, airlines and restaurants with live pricing and feature detailSame entities, real time prices and key features removedNobody, in the short term
Direct supplier listingsReachable through richer, price bearing unitsReached after an additional intermediary stepOnline intermediaries over direct suppliers

Nick Fox, Google's Senior Vice President of Knowledge and Information, framed the outcome as one that degrades the user experience for Europeans by boosting online intermediaries at the expense of local businesses, in the Reuters exclusive that carried the announcement. Google also says its own testing with millions of European users showed people retyping queries to find what they wanted, and it points to a 30 percent drop in free, direct booking traffic to European businesses following earlier DMA compliance changes. When Google tested a version of this layout in 2024, a test the European Commission then rejected, it reported hotels losing more than 10 percent of their traffic.

Google DMA search changes are an event and a lobbying position

Every number in the paragraph above came from the company that did not want to make this change, released on the morning the change shipped, sixty days after a 460 million euro fine, while the ceiling on further penalties sits at 5 percent of worldwide turnover. You would be careless to treat those figures as neutral measurement. You would be equally careless to treat them as fabricated.

Both things are true at once. Google has better instrumentation on European search behavior than any regulator, any comparison site and certainly any agency. Its claim that users retype queries more often is the kind of thing it can actually measure and the kind of thing that would be genuinely stupid to invent while under a compliance order. It is also a claim released in a format, an exclusive to a wire service, chosen to shape a policy debate rather than to inform your quarterly planning.

The professional response to a number like that is neither adoption nor dismissal. It is replication. Google has told you the direction it expects, the mechanism it blames, and the population it affects. Those are testable, and you have the data to test them on your own traffic, which nobody in this argument has an incentive to characterize honestly on your behalf.

THE DISTINCTION THAT MATTERSGoogle's 30 percent figure is evidence about Google's intent and its internal read, not a forecast for your account. Treat it as a hypothesis with a named source and a stated mechanism. Then go and see whether your own European traffic behaves the way that hypothesis predicts, because the answer is worth considerably more than the argument.

Europe is the cleanest control group search has been handed

Almost every change worth measuring in this industry arrives globally, gradually, and unannounced. Core updates roll for two weeks across every market at once. There is no untreated population, so causal attribution collapses into correlation and vibes, which is why so much of the field argues from anecdote.

This one is different in all three respects. It is dated to a single day. It is bounded by a regulatory border that maps almost exactly onto a field already present in your analytics. And the population on the other side of that border is unaffected by design, because Google has said so publicly and has every regulatory reason not to be quietly applying it elsewhere. If you sell into both Europe and North America, you have a treatment group and a control group that you did not have to construct, recruit or pay for.

The border is legible in your dataCountry dimension already exists in every analytics tool you run. Treatment is EU and EEA traffic, control is everything else. You are not building a segmentation so much as switching one on, and this is the rare case where a country split is a causal instrument rather than a reporting nicety.
The date is a single point, not a rolloutCompliance changes land on deadlines, not on gradual ramps. That gives you a clean before and after boundary, which is what makes a difference in differences comparison legitimate rather than decorative.
Seasonality is the confound to kill firstSeptember into October carries real seasonal movement, and it is not identical across regions. This is precisely why you need the non-EU control rather than a simple month over month EU comparison. The control absorbs the seasonality that would otherwise be read as effect.
Query mix is the confound most teams missThe change is concentrated in verticals where comparison engines and travel entities appear. If your EU and non-EU query mixes differ, and they usually do, compare within matched query classes rather than across whole markets, or you will measure your product catalogue instead of the change.

The three numbers Google DMA search changes will move

Direction is predictable from the mechanism even where magnitude is not. Three things follow from placing an intermediary above a supplier and removing price from the preview.

Impressions for direct supplier pages should hold or fall modestly, because the queries have not gone anywhere. Click through rate on those impressions should fall, because an interposed intermediary captures clicks that previously continued to the supplier. And here is the one almost nobody has said out loud yet: on the clicks that do arrive, conversion rate should fall too. Stripping real time prices and key features from the carousel removes the qualification step that used to happen before the click. Users who would once have self selected out on price now arrive, look, and leave.

If that holds, European teams are about to see a traffic decline and a conversion decline at the same time, which is the signature of a landing page problem. It will be diagnosed as one. Money will be spent on it. The cause will be a layout change in a market a thousand miles from whoever is running the test.

EU organic sessions, segmented on their own25%
EMEA regional rollup, with EU at 60 percent of the region15%
Global blended organic line, with EU at 22 percent of total6%
Global blended line against a 3 percent lift in other markets3%

Illustrative math, not measured data: a decline confined to EU organic traffic, shown as it appears at four reporting granularities. Percentages are the size of the drop as seen in each view.

Substitute your own regional weightings, because the shape is the argument rather than the figures. A material regional shock arrives at the executive dashboard as a rounding error, sits comfortably inside normal weekly variance, and gets explained away. The same problem shows up whenever a single blended line stands in for several different markets, which is the reason we keep arguing for honest error bars on search measurement rather than one confident number.

A twenty five percent decline in one market becomes a six percent wobble in a global line, and a six percent wobble is indistinguishable from a bad fortnight. The change did not hide. Your reporting hid it.

Instrument it before your October reporting locks in

This is a week of work at most, and its value decays fast, because the pre-period is finite and shrinking. Everything below assumes you already have country level data, which almost everyone does and almost nobody segments on.

01Freeze a clean pre-period todayExport EU and non-EU organic sessions, clicks, impressions, average position and conversion rate for the eight weeks ending September 7, split by country and by query class. Do it as a stored export rather than a live report, so a later interface change or data retention window cannot quietly rewrite your baseline. We have written before about keeping reporting redundancy against analytics outages, and the same discipline applies to a baseline you will still be citing in January.
02Split treatment from control at the country levelEU and EEA countries are the treatment group. Everything else is control. Resist the urge to use an EMEA rollup, because the UK and Switzerland sit outside the DMA perimeter and will dilute the exact signal you are trying to read.
03Match query classes across both groupsPick the query classes where vertical search services and travel entities actually appear, then compare those classes to themselves across the border. A whole market comparison measures your catalogue and your brand strength, not this change.
04Track conversion rate as a first class metric, not a footnoteThe price stripping effect shows up in conversion, not in sessions, and conversion is the metric most likely to be reported at global granularity only. If you measure one thing at country level this quarter, measure this, because it is where the mechanism and the money meet.

None of this requires new tooling, which is the point. It requires deciding, before the data arrives, what the comparison is. Decide afterwards and you will find the comparison that fits whatever story is convenient in October, which is how entire regional shifts get absorbed into a narrative about algorithm updates.

What to do in the next thirty days

If you have direct booking or direct purchase revenue in Europe, the intermediary channel you may have spent years disintermediating just got structurally promoted above you. That is not a reason to abandon direct, and it is certainly not a reason to panic in week one. It is a reason to check what your presence looks like inside the comparison engines that now sit above you, and to price that channel honestly against the traffic it is about to be handed. This is the same dynamic we walked through when Google placed a fixed list of partners inside AI Mode booking: when the surface promotes a fixed set of intermediaries, presence inside that set stops being a nice to have.

If you are a comparison or vertical search service, there is a live opportunity in the top slot, and it is now algorithmically allocated rather than assigned by rule. Slots decided by an algorithm are slots that can be earned, and the field is briefly unsettled while everyone works out what earns them.

And if you sell to local businesses, read Google's own framing carefully, because it is the most testable claim in the whole announcement. Google says the change hurts local businesses and helps intermediaries. If that is true in your vertical, it will be visible in your clients' EU numbers within six weeks, and you will be the only party in this argument holding independent evidence either way.

DO THIS NEXTBefore you close this tab, pull an eight week EU versus non-EU organic export split by country and query class, and save it somewhere that is not a live dashboard. That single export is the difference between knowing what this change did to you and guessing about it for the next two quarters. It takes twenty minutes today and cannot be reconstructed in November at any price. If regional segmentation is not something your current setup makes easy, that is the actual finding, and it is what our reporting and analytics work usually starts by fixing.

See where you are cited today

A free snapshot audit of your rankings and AI citations before we ever talk.

TT
Tyler TruffiMANAGING PARTNER, SOMETHING INC.

Tyler leads work at the intersection of SEO and generative engines at Something Inc., helping B2B brands get ranked and cited across every major AI engine.

Free consultation

Let us be the last SEO agency you ever work with

A 30 minute call and a free audit of your SEO and GEO position. You keep the findings either way.