Performance Max for small businesses: when it works and when to switch it off

Performance Max is one campaign that runs across Search, YouTube, Display, Gmail, Discover and Maps. You give it assets and a goal, and Google decides where and to whom the ads show. That's the appeal and the risk in one sentence. When the goal is clear and the data is clean, it scales well. When either is missing, it finds cheap clicks and reports them as success.

I've had it go both ways in accounts I run, so this is how I decide.

Condition 1: it knows what a result is worth

For Ginyaki, a restaurant group in Islamabad with its own ordering app, whose Google account I ran in 2023, the first Performance Max campaign was set to maximise conversions. It returned 1.25 in conversion value for every unit spent. The second was set to Target ROAS, optimising on GA4 purchase events that carried the real order value. It returned 4.27, and on its own produced 292 tracked orders worth Rs597,030.

Same business, same products, same platform. The difference was that the second campaign was told what an order was worth, so it went looking for people who place bigger orders rather than any order at all. If your tracking sends purchases without values, or sends a "conversion" for a page view, Performance Max has nothing useful to optimise on. Fix that first or don't run it. The Ginyaki Google Ads case study has the figures.

Condition 2: there's enough data to learn from

Google's own guidance is that a campaign needs a steady flow of conversions to come out of learning. In practice I want to see around 30 a month on the conversion that matters before I trust Performance Max with a meaningful share of the budget. Below that it stays in learning for weeks and makes decisions on noise.

For a service business that gets ten enquiries a month, Search is almost always the better home for the budget. Search lets you see the exact queries you're paying for and cut the ones that are wrong. Performance Max shows you much less.

Condition 3: it's buying new customers, not your own

Left to itself, Performance Max will happily show ads to people searching for your brand name, and count those sales as wins. They would have happened anyway. I exclude brand terms from it so the campaign is judged on new demand, and I check the placement report. Spend going to apps and low-quality sites means the campaign is counting junk as success.

When I switched it off

For Boatastic, a yacht hire business in the UAE, Performance Max was the first campaign in the account. It reached people cheaply, with 724 clicks and 787 video views at AED0.10 each, but at a 3.55% click-through rate and with enquiries coming from elsewhere. The people who wanted a boat trip were typing it into Google. A search campaign built around three ad groups and bid on a target cost per lead ran at a 15.23% click-through rate and brought 111 enquiries at AED56.47. The budget moved to search and stayed there. The Boatastic case study covers it.

That isn't Performance Max failing. It's the wrong tool for a high-ticket service where demand is expressed through search and the conversion volume is too low for it to learn.

The short version

Run Performance Max when purchases are tracked with real values, when there are enough of them each month for it to learn, and when brand searches are excluded so it has to find new customers. That usually means ecommerce, food ordering and app-based businesses with volume.

Use Search when enquiries are the result, volume is modest, and you need to see what you're paying for. Most local and professional service businesses fall here.

And whichever you run, judge it on orders or qualified leads in your own records, not on the conversions column alone. The Google Ads service page explains how I set that up.