Skip to content
← Back to Bytes
Paid Media/Aug 28, 2026

Performance Max: When It Works and When It Hides the Waste

Performance Max is strong at harvesting demand that already exists and weak at telling you why something failed. Here is what to feed it and what to cut.

TL;DR

Performance Max is very good at collecting demand that already exists and poor at explaining itself. It rewards clean conversion data, strong creative and firm exclusions, and it punishes accounts with weak tracking by spending confidently on the wrong thing. Judge it on business outcomes, never on its own reporting.

Performance Max arrives in most accounts as a promise of simplicity: hand over the assets, hand over the budget, let the system decide where the ads run. The Growth Bully, a Malta performance marketing agency, runs it in plenty of accounts and it earns its place in a lot of them. It also quietly absorbs budget in accounts that were never set up to hold it accountable.

The honest summary is that Performance Max is excellent at harvesting demand that already exists and poor at telling you why a result happened. Both halves of that sentence matter. Treat it as a demand harvester with weak diagnostics and it behaves. Treat it as a strategy and it becomes the single hardest line in the account to explain.

What is Performance Max actually doing?

It is one campaign that places ads across search, shopping, display, video, discovery, maps and email inventory, choosing the combination itself. You supply assets, audience signals, a budget and a conversion goal. The system decides which channel, which creative and which audience each impression goes to, then optimises towards the goal you defined.

That last clause carries all the risk. The campaign will pursue whatever you set as the goal with total commitment, which is an advantage when the goal is a real business outcome and a serious problem when the goal is a form view, a page load or an event that fires more often than it should.

When does Performance Max work well?

When there is existing demand to collect, a clean conversion signal to chase, and enough conversion volume for the system to learn from. Retail catalogues, established brands with steady search interest, and accounts with accurate transaction data are where it consistently performs. It is very good at finding the people already looking for what you sell.

It also works well as a coverage layer over an account that already has deliberate search campaigns. In that arrangement the search campaigns own the terms you have decided to compete on, and Performance Max collects what falls outside them. The results are readable because you can still see which part of the account produced what.

When does Performance Max hide the waste?

When the conversion signal is weak, generous or wrong. The campaign optimises confidently towards whatever it is told counts, so an account counting newsletter signups and quote requests as one undifferentiated conversion will get a lot of newsletter signups. The reporting will look excellent right up until somebody checks the sales pipeline.

The second failure mode is brand absorption. Left unconstrained, Performance Max will happily serve against your own brand searches, claim conversions that would have happened anyway, and post a cost per acquisition that looks like a triumph. Nothing has been created. Existing demand has been repriced and re-credited, and the account now believes something untrue about where its growth comes from.

What should you feed a Performance Max campaign?

Better inputs, not more budget. The campaign only has your assets, your data and your goal to work with, and it will amplify whatever quality it finds there. Weak creative and a vague conversion goal will not be rescued by a larger budget, they will simply be distributed faster and further.

  • One clean conversion goal that maps to a real business outcome, with lower value actions either removed or valued separately.
  • Genuinely varied creative across formats, since the system cannot test a variety it was never given.
  • Accurate conversion values where different outcomes are worth different amounts, because a single flat value teaches it that every result is equal.
  • Audience signals built from real customers rather than aspirational demographics, used as a starting hint and not as a hard target.
  • A landing experience that holds up, because traffic quality cannot compensate for a page that loses people, as landing page structure covers in detail.

What should you exclude?

Brand terms, existing customers where repeat purchase is not the objective, and placements that consistently produce traffic without outcomes. Exclusions are the main steering mechanism you retain in an automated campaign, and an account that has set none has effectively agreed to accept whatever the system finds cheapest.

Brand exclusion is the argument worth having early. The counterargument is always that brand traffic converts well, which is true and beside the point, because it would have converted without the ad. Separating brand from everything else is what makes the rest of the account legible, and it is the first change we make in most Google Ads accounts we take over.

How do you read the reporting gaps?

Carefully, and with outside evidence, because the campaign reports less about its own internals than any other campaign type. Channel level detail is limited, search term visibility is partial, and asset group reporting tells you what performed without reliably telling you where. Judge it on outcomes you can verify independently.

A workable routine looks like this:

  1. Set a baseline before launching, covering total account cost per acquisition and total enquiry volume, so the campaign can be judged against the account rather than against itself.
  2. Watch total account results, not campaign results, for the first month, since the fastest way for Performance Max to look successful is to absorb conversions another campaign was already producing.
  3. Check brand and non brand separately, every month, without exception.
  4. Reconcile reported conversions against the CRM, because platform reporting cannot tell you whether an enquiry was any good. Our conversion tracking guide covers how that reconciliation should work.
  5. Review asset groups on creative quality rather than on placement mystery, since creative is the input you can genuinely control.

Should Performance Max be your only campaign type?

Almost never. As the only campaign in an account it removes your ability to diagnose anything, because every result arrives blended and unattributable. When performance drops there is no isolated variable to examine and no comparison campaign to reason against, so the only available response is guessing.

A more durable structure keeps deliberate search campaigns on the terms that matter commercially, keeps brand separate and constrained, and gives Performance Max the remaining space to do what it is genuinely good at. That structure costs a little more attention to run and it keeps the account explainable, which is the difference between managing spend and watching it. The same logic runs through how we build every digital marketing account and every lead generation programme, alongside the Meta side of the media mix.

If your Performance Max campaign reports well and your sales team has not noticed, those two facts are related. Show us the account and we will tell you which of those conversions were already yours.

Want more of this in your Google results? Add us as a preferred source.
Questions

The honest answers.

Is Performance Max better than standard search campaigns?

They do different jobs. Standard search campaigns let you decide which terms you compete on and show you exactly what you paid for, which makes them the right tool for commercially important keywords. Performance Max collects demand across every Google surface with far less visibility. Most healthy accounts run both, with search owning the deliberate terms and Performance Max covering the rest.

Why does my Performance Max campaign show a great cost per acquisition?

The most common reason is brand absorption. Unless brand terms and existing customers are excluded, the campaign will serve against people already searching for you and claim conversions that would have happened anyway. The second reason is a loose conversion goal that counts low value actions, such as newsletter signups, alongside genuine enquiries. Check both before trusting the figure.

Should I exclude brand terms from Performance Max?

In most accounts, yes. Brand traffic converts well because the person already decided to look for you, so paying for it inflates the campaign result without creating demand. Excluding brand keeps the campaign honest about what it is generating and makes the whole account readable. Keep brand in a separate, deliberately managed campaign where you can see its true cost.

How much conversion data does Performance Max need to work?

Enough for the system to learn a pattern rather than react to noise, which in practice means a steady flow of the same clean conversion action rather than an occasional one. Accounts with sparse or inconsistent conversion data get erratic results, because the campaign is optimising towards a target it cannot see clearly. Fix the tracking and the volume question before adding budget.

How long should I run Performance Max before judging it?

Give it a defined learning period of several weeks, then judge it on total account outcomes rather than on its own report. The key comparison is account level cost per acquisition and enquiry volume before and after launch, because a campaign that only redistributes existing conversions will look successful in isolation while the account as a whole has not moved.

Put it to work

Ready to get aggressive?

Stop reading about growth and start running it. Let us build the marketing engine that does the work for you.

Real results, tracked to the euro. Your growth plan in 30 minutes.Book a growth call →