If you have spoken to a Google rep in the last two years, you have been told to run Performance Max. It is the campaign type Google promotes most aggressively, and for understandable reasons. When it works, it works extremely well. When it does not, it is one of the most expensive ways to learn that lesson.
We get asked about Performance Max constantly. Should a Stroud roofer run it? A Cheltenham dentist? A Gloucester family law firm? The honest answer is that it depends, and the factors that decide it are not the ones Google's interface tells you about. Here is what actually matters.
What Performance Max actually is
Performance Max is a single campaign that runs your ads across every Google property at once. Search, Display, YouTube, Gmail, Discover and Google Maps, all from one campaign with one budget. You upload your creative assets, set your conversion goals, and Google's machine learning decides where, when and to whom to show your ads.
That is genuinely powerful. A well-built Performance Max campaign can find buyers in places a Search-only campaign would never reach. The problem is that it can also spend a frightening amount of money on placements that will never convert, especially in its first few weeks while the algorithm is still learning.
Where Performance Max genuinely shines
There is a specific profile of business that does very well on Performance Max, and it is worth knowing whether you fit it before you spend anything.
Performance Max works best when you have plenty of conversion data. The algorithm needs signals to learn from. If your account is generating thirty or more conversions a month, Performance Max has enough data to make sensible decisions. If you are generating three, it is essentially guessing.
It also works best when your conversion has an obvious monetary value. E-commerce stores, online bookings, lead forms with a known close rate. Anything where Google can clearly see the outcome of a click and learn to bid for similar ones.
Finally, it works best when you have strong creative assets. Performance Max is asset-hungry. It wants headlines, descriptions, images, videos and logos. If you only have a handful of stock photos and one tagline, the system will be working with one arm tied behind its back.
Where it burns money for small accounts
Most local service businesses do not fit that profile. They convert eight or ten times a month, not eighty. They generate enquiries that turn into customers through a phone conversation rather than an online sale. They have a logo and maybe a few photos from a previous job. For these accounts, Performance Max behaves very differently.
The first issue is data starvation. With low conversion volumes, Performance Max never moves out of its learning phase properly. It keeps experimenting with new placements and audiences, and each experiment costs you money. A Search campaign with manual keyword targeting would already be optimised at the volume your account produces. Performance Max is still trying things out.
The second issue is placement bleed. Performance Max distributes spend across Search, Display, YouTube and Discover automatically. For a local plumber, the bulk of the high-intent traffic is on Search. Performance Max will still spend a significant share of your budget on Display impressions in the corners of unrelated websites, because that is what the algorithm believes is optimal for your bid strategy. You cannot see exactly where the spend is going either, because Google's placement reporting is deliberately limited.
The third issue is brand cannibalisation. By default, Performance Max can bid on your own brand searches, and it will, because they convert. Your reporting will show Performance Max delivering wonderful conversion numbers. What it is not telling you is that those conversions would have happened anyway, for free, from organic search. Your overall paid spend goes up while your incremental revenue stays flat.
The audience signal myth
Google tells you that audience signals are a "suggestion" to help Performance Max find your ideal customers faster. In practice, the algorithm treats your audience signals as one input among many and will frequently spend the majority of your budget reaching people who match no signal you provided.
This is not a bug. It is how the system is designed. Google's machine learning has access to billions of signals you do not. It believes it can find converters you would never have targeted. Sometimes it is right. Sometimes it is wrong, and you only find out after the budget is gone.
For small accounts with limited conversion data, audience signals are nowhere near enough to keep Performance Max focused on the right buyers. Your campaign needs hard constraints, and Performance Max does not give you those by default.
Asset groups: where most accounts go wrong
If you are going to run Performance Max, the asset group structure determines almost everything about how well it performs.
A common mistake is creating one asset group with every product or service mixed together. The algorithm then averages everything out and learns slowly. The better approach is one asset group per distinct product or service, each with its own headlines, descriptions, images and audience signals.
For a kitchen company with a showroom in Cirencester, that might mean one asset group for "kitchen design," one for "kitchen installation" and one for "showroom enquiries." Each one tightly themed, each one with creative that matches.
It is more work to set up. It is also the difference between a Performance Max campaign that converts and one that spreads thinly across irrelevant searches.
Final URL expansion: turn it off
By default, Performance Max can send traffic to any page on your website that it thinks matches the search. This is called final URL expansion. For most local businesses it is a disaster.
You spend weeks building a beautiful landing page for kitchen extensions. Performance Max decides a visitor's search is closer to your "about us" page and sends them there instead. They bounce. You pay for the click.
Turn final URL expansion off. Force Performance Max to send traffic only to the landing pages you have approved.
When to use it (and when not to)
Performance Max makes sense for a small business when:
- Your account is generating thirty or more conversions a month already.
- You have a clear value associated with each conversion.
- You have multiple distinct services or products that warrant separate asset groups.
- You have invested in proper creative assets, including at least a few short videos.
- You are willing to commit to a 60 to 90 day learning period and not panic when early weeks look messy.
Performance Max does not make sense when:
- Your conversion volume is in single digits per month.
- You have one core service offered to a single local catchment.
- Your conversions are phone calls with no clear value attached.
- You cannot afford to spend 20 to 30 percent of your budget on data exploration.
For most of the local businesses we work with in Gloucestershire, the answer is the second list. A well-built Search campaign with proper conversion tracking and a robust negative keyword list will outperform Performance Max consistently at the volumes most local accounts produce.
The realistic verdict
Performance Max is a powerful campaign type that Google has spent years engineering. It is also, at the small end of the market, the campaign type that wastes the most money. The two facts are not contradictory. The technology rewards scale and data. If you have neither, you are subsidising other people's experiments.
If you are running Performance Max now and your results feel underwhelming, it is worth pulling the campaign apart and looking at where the spend is going. If you are being pushed into Performance Max by a Google rep and your account does not fit the profile above, push back. There is no obligation to take their advice, and your budget will thank you for it.
If you would like an honest second opinion on whether your current campaign structure is the right one, request a free Google Ads audit here. We will tell you exactly what is working, what is not and what we would change.