Google Performance Max can report a strong return and still be the worst line in your ad account. It looks like it's working because it spends your budget on demand you already had — branded searches, repeat visitors, people who were going to buy anyway — and it hides where the rest of the money goes. For a small European ad budget, that isn't a campaign you optimise. It's a bet you probably haven't earned yet.

Performance Max is the fully automated campaign type Google pushes harder than any other. For most small European advertisers it is also the wrong default, and the reasons are hidden in plain sight.

Direct answer: Performance Max bundles all of Google's inventory into one automated campaign and switches off most of the controls you'd use to steer it. Its reported conversions often credit demand you already owned, so the dashboard flatters it. On budgets under roughly €1,000–3,000 a month, it rarely earns its keep.

What Performance Max quietly switches off

Performance Max runs one campaign across Search, Shopping, YouTube, Display, Gmail, Maps, and Discover at the same time, with Google's machine learning deciding who sees your ads, where, and at what bid. That is the pitch: one campaign, every surface, no manual work.

Read what it removes and the pitch reads differently. You don't choose keywords — you hand over assets and audience "signals" and let the system infer intent. For its first three years it showed you almost nothing about which searches or placements it bought. You get one blended budget across channels you can't rank against each other. The trade is real: you gain reach and hands-off automation, and you give up the ability to see and steer where your money goes.

For a founder who already can't tell whether their ads work, handing over the last of the controls is the wrong direction to walk.

The number that flatters it

Start with the metric everyone trusts and Performance Max quietly games: reported conversions.

PMax has a habit of absorbing branded and existing-intent searches — the cheapest, most certain conversions in your account — and booking them as its own. An Optmyzr study of 503 accounts in February 2025 found keyword overlap between Search and Performance Max in 91.45% of them, and search-term overlap in over 97%. When the two competed for the same query, Search posted the better conversion rate far more often than PMax did. The machine wasn't finding new customers. It was standing in front of the ones already walking through the door and taking the credit.

That is why the dashboard looks good. A marketer who pulled every client off PMax told Digiday the incrementality "came back less than 10% — in other words, it didn't produce incremental revenue." The return was real on the screen and close to fictional in the bank.

A conversion you were always going to get is not a conversion worth paying a machine to reclaim.

Where the rest of your money goes

The branded searches are the flattering part. The rest of the budget is the expensive part.

Left alone, Performance Max pushes spend into Display and YouTube placements — including mobile games and low-rent apps — that generate "engagement" nowhere near a buying decision. On a large ecommerce account that noise averages out. On a small budget it's fatal.

Search Engine Land documented a small retailer whose Google rep steered them into Performance Max on a new account. They spent over $3,000, saw clicks as expensive as $50 each, and got a single purchase in return. Rebuilt as a properly tracked Shopping campaign, the same account booked 56 new customers in two weeks at $53 each. The money in the first version didn't disappear. It went somewhere they couldn't see and would never have chosen.

The 2025 transparency updates are real — and not enough

Google did open the box a little, and it's worth being fair about that. Since early 2025 you can add campaign-level negative keywords (up to 10,000 by 2026), pull a full search-terms report with a column showing which channel triggered each query, and see channel-level performance. Three years ago none of that existed.

Here's the catch. It's transparency without steering. You can finally see that money went to a junk placement or an irrelevant search, but you still can't tell the algorithm to move budget off an underperforming channel. Watching the black box is not the same as driving it. And the advertiser who needed those controls most, the one with the smallest budget, is the least likely to have the hours to police a search-terms report every week.

The budget maths European founders can't skip

Performance Max charges tuition before it teaches. It needs volume to learn — practitioners put it around 30 to 50 conversions in the first 30 days — and Google's own guidance is a daily budget near 10 times your target cost per conversion. Run the sum on a €50 lead and you're looking at roughly €500 a day before the system is even trained.

Below about €1,000 to €3,000 a month, most campaigns never generate the conversions PMax needs, so it keeps guessing with money you can't spare. Europe adds a second tax. Performance Max runs on Google's audience signals and your first-party data, and EU consent rates leave both thinner than they are in the US. Weaker signal, smaller budget, more guessing.

Performance Max does earn its place in the right account: real product inventory Google can match to shopping intent, enough conversion history to train on, and a Search campaign already maxed out on qualified keywords. That describes a mature online store. It doesn't describe most owner-operated European businesses running a form and a modest budget.

The operator's test before you switch it on

Before you — or your agency — turn Performance Max on, four questions decide it:

• Are you already logging 30-plus tracked conversions a month? If not, the system can't learn, and you'll fund the guessing.

• Do you sell products with clear shopping intent, or are you lead-gen with a contact form? Lead-gen with weak signals is exactly where PMax bleeds.

• Have you excluded your brand terms? Otherwise it resells searches you already own and calls it growth.

• Can anyone tell you which query and which placement produced last month's leads? If not, you're paying for a number, not a result.

There's a tell for the ones who don't run it themselves. If an agency runs your Google Ads and the monthly report is all green arrows with no answer to "which search term produced this enquiry," Performance Max is often where the black box is hiding. I want results, not reports is the correct instinct — make them show you the mechanism, not the dashboard.

We don't make vibes-based decisions with a client's ad budget, and we won't hand a small one to a system we can't steer. Automation is something you earn with data, not something you switch on to skip the work. Performance Max isn't a black box you optimise — it's Google asking you to stop asking questions, and a small budget can't afford the tuition. Track from click to client, not from click to dashboard.

FAQ

Is Performance Max good for small budgets?

Usually not. Performance Max needs roughly 30 to 50 conversions a month to train its bidding, and Google suggests a daily budget near 10 times your target cost per conversion. Below about €1,000 to €3,000 a month, most accounts never reach that, so the system spends while it guesses.

Does Performance Max cannibalise brand search?

Often, yes. A February 2025 Optmyzr study of 503 accounts found keyword overlap between Search and Performance Max in 91% of them. PMax tends to absorb branded and existing-intent searches and report them as its own conversions, which flatters the return. Exclude your brand terms so it has to earn new demand.

Did the 2025 updates fix the black box?

Partly. You can now see a search-terms report, view channel-level reporting, and add campaign-level negative keywords. But you still can't tell the algorithm to move budget off a weak channel. It's transparency without control — you can watch where the money goes, not steer it.

Should I let my agency run Performance Max?

Only if they can tell you which query and which placement produced each lead, and why PMax beats a structured Search campaign for your specific account. If the report is all green arrows and no answers, that's the tell. Ask for the mechanism behind the numbers, not the numbers.

Plan. Build. Iterate.

That's the loop, and it applies to ad spend as much as to a website. If your Google or Meta budget is vanishing into a dashboard you can't read, book a 30-min call and get an honest read on where it's actually going. See how Ads Management runs inside one senior-led subscription — with every price on the Plans page.