Performance Max for small business: when it helps, when it hides

Introduction

The pitch is appealing: supply assets and a budget, and the system finds customers across search, display, video, maps and shopping without campaign management.

Sometimes that works well. Often it produces a campaign that spends steadily, reports conversions, and cannot be improved because nobody can see what it is doing.

1. Performance Max for small business depends on conversion data first

The system optimises toward whatever you tell it a conversion is. With nothing reliable to optimise toward, it optimises toward the wrong thing efficiently.

That means conversion tracking has to be working and meaningful before launch — a real enquiry or sale, not a page view or a click on a phone number that nobody answers.

If tracking is absent or counts low-value actions, this format will produce more of exactly that. It is the one prerequisite with no workaround.

2. Volume matters more than in a manual campaign

Automated bidding needs conversions to learn from, and a small business often does not supply many.

A campaign generating a handful of conversions a month gives the system very little signal, so it spends much of the budget exploring. Manual search campaigns on tightly matched keywords frequently outperform it at that scale.

Rough guide: if you are not producing conversions weekly, start with search and revisit this later.

3. Feed it real audience signals

Audience signals do not restrict targeting — they tell the system where to begin looking, which shortens the expensive learning period.

The strongest signal available to a small business is its own customer list, uploaded as a segment. Website visitors and people who have enquired before also help.

Generic interest categories are the weakest option and the one most often used, because it needs no preparation.

4. Set the exclusions before it spends

The controls exist and they are easy to miss during setup.

Add your negative keyword list at account level, exclude the locations you cannot serve, and set brand exclusions so the campaign is not buying searches for your own name that you would have received free.

Also exclude placements you do not want. Without that, some of the budget goes to app inventory and low-quality display space, which is where "cheap conversions that never call" tend to come from.

5. Supply real assets, not placeholders

The format assembles from what you give it, across formats you might not have considered.

Weak or missing images and video mean it leans on whatever it can generate or crop, and the output is often unrecognisable as your business. Supply proper images at the sizes requested, and at least one simple video, or the video inventory is filled automatically.

Poor assets are also a reporting problem: performance differences you cannot explain often come down to a badly cropped image running in a placement you never looked at.

6. Accept the reporting limits, and use what exists

You will not get the channel-by-channel and keyword-by-keyword clarity of a manual campaign.

What you can see is worth checking: asset group performance, the search themes report, and placement data. It is less than you want and more than most people look at.

If you need to know precisely which searches drove sales, that is an argument for running search separately rather than for arguing with the format.

7. Run it alongside search, not instead of it

The practical structure for a small budget: a manual search campaign on transactional keywords, and this format as an addition once the first is stable.

That way your highest-intent searches stay under direct control, with known costs and known landing pages, while automation works the wider inventory.

Watch for overlap — with brand exclusions absent, it will happily take credit for demand you already had.

8. Judge it on incremental customers

The evaluation question is not whether it reports conversions. It is whether total customers rose when it started.

Compare overall enquiry and sale volume before and after launch, not just the campaign's own reported figures. Automated campaigns are effective at capturing existing demand, which shows up as reported conversions without any increase in business.

If total volume did not move, you moved budget rather than growing.

Conclusion

Get conversion tracking honest before launching, and only use this format once you are producing conversions regularly enough to teach it something.

Feed it your customer list, set negative keywords, location, brand and placement exclusions before it spends, supply real assets, keep manual search running underneath, and judge it on whether total customers increased rather than on its own reported conversions.

Related reading

• How do you build a local growth plan for your small shop?

• Steps to create a Performance Max campaign for retail stores

• The funnel for a repeat purchase business is a loop

• The novice accountant's guide to understanding profitability reports for small stores

• Conversion tracking setup: what to measure before you spend

• What not to automate in a small business, and why it matters