Google Performance Max for Ecommerce: When It Works and When It Does Not

Performance Max (PMax) is Google's all-in-one campaign type. One campaign can serve ads across Search, Shopping, YouTube, Display, Discover, Gmail and Maps, with Google deciding where each dollar goes. For ecommerce brands with a product feed, it has become one of the main ways to run Google Ads.

It can be very effective. It can also quietly spend your budget on people who were going to buy anyway. The difference usually comes down to inputs, structure and how you measure it.

How PMax works, in plain terms

You give Google a goal (usually purchases with a value), a budget, a product feed from Merchant Center, creative assets (headlines, descriptions, images, videos) and optional audience signals. Google then uses its bidding and prediction models to decide which channel, which product and which ad to show to each person. Audience signals are hints, not targeting. Google can and will show ads to people outside them.

Because so much is automated, the quality of what you feed the system matters more than any setting you control directly.

When PMax tends to work

• →A clean, rich product feed: accurate titles with the words people search for, good images, correct prices and availability, useful product types and custom labels.
• →Reliable conversion tracking with real purchase values, ideally with enhanced conversions set up so Google can match more sales.
• →Enough conversion volume for the bidding to learn. Accounts with very few monthly sales often see unstable results.
• →Healthy margins and a clear target for return on ad spend, so you can give the campaign a realistic goal.
• →A product range where shoppers compare options visually, which plays to Shopping and YouTube placements.

When PMax struggles

• →Weak or messy feeds. If the feed is thin, the campaign has little to work with and may lean on cheap Display or video inventory.
• →Very small budgets spread across a large catalog, so no product gets enough data.
• →Accounts where most conversions come from branded searches and returning customers. PMax can claim those sales and look great while adding little.
• →Businesses that need tight control over exactly which queries trigger ads, such as regulated products or very narrow niches.
• →Broken or inflated tracking, for example counting add-to-cart as a purchase. The system will optimize for whatever you tell it to.

Controls worth using

Google has added more control to PMax over time. Settings and names change, so check what is available in your account, but these are the levers we review in every audit:

• →Brand exclusions, so PMax does not compete with a separate branded Search campaign.
• →Negative keywords at the campaign or account level to block irrelevant or unwanted queries.
• →Search term and channel reporting, to see which queries and which surfaces are actually driving results.
• →Listing groups and multiple asset groups, to split products by margin, category or performance.
• →The new customer acquisition goal, to bid higher for new customers or focus only on them.
• →Final URL expansion settings, so Google does not send traffic to pages you did not intend.

A structure that holds up

There is no single right setup, but this one works for many ecommerce accounts:

1. 01A branded Search campaign with brand terms, kept separate and excluded from PMax, so you can see and control what brand traffic costs.
2. 02One or more PMax campaigns split by business logic: high-margin versus low-margin products, best sellers versus the long tail, or core categories. Each split should have a different goal or budget, otherwise it adds complexity without benefit.
3. 03Standard Shopping or non-branded Search campaigns where you need more control, for example to protect key product categories or test specific queries.
4. 04Custom labels in the feed (margin band, season, stock level) so you can move products between campaigns without rebuilding everything.

How to judge whether it is working

Do not judge PMax only on the ROAS it reports. Look at the business:

• →Is total revenue growing as PMax spend grows, or is it flat while PMax takes credit?
• →What share of PMax conversions come from new customers?
• →What happened to branded search volume and direct traffic? If they drop as PMax scales, it may be absorbing existing demand.
• →How do blended efficiency and new customer acquisition cost move month over month?

If you have the budget, a geographic holdout, where you run PMax in some regions and not in others for a few weeks, gives you a much clearer answer than any dashboard.

Creative still matters

Even in a feed-led campaign, assets shape where and how you show up. Provide a full set of headlines and descriptions, lifestyle and product images in the main aspect ratios, and at least one real video. If you do not upload video, Google may generate one automatically from your images, and you have less control over how it looks.

Key takeaways

• →PMax works best with a strong feed, accurate purchase tracking, enough conversion volume and realistic targets.
• →It struggles with weak feeds, tiny budgets spread too thin and accounts dominated by branded or returning demand.
• →Separate branded Search, use brand exclusions and negatives, and split campaigns by real business logic.
• →Judge PMax on total revenue, new customer share and blended efficiency, not on its own reported ROAS.