Performance Max for Ecommerce: A Practical Setup Guide

Performance Max works well for ecommerce when the product feed is strong, the conversion data is reliable and brand search is kept out of it. Get those three right and the rest of the setup is about steering it towards the products that make you money and checking it every week.

This is how I set it up for ecommerce clients, including what I did on Ghost when I added Shopping and Performance Max to an account that had only been running search.

Start with the feed

Performance Max for retail is built on your product feed. The feed decides which searches and placements your products are eligible for and what the ads look like. A weak feed caps the whole campaign, and no amount of bid strategy tweaking fixes it.

Before launching, I check:

Titles. Lead with what the product is and the attributes people search for, such as brand, product type, colour, size and material. Titles written for the website are often too short or too clever.
Images. Clean main images on a plain background, with lifestyle images added as additional images where the feed allows.
Product type and categories. Accurate categorisation helps both Google’s matching and your own reporting later.
Price and availability. These need to match the site exactly. Mismatches lead to disapprovals.
Disapprovals and warnings. Clear these in Merchant Center before spending anything.
Custom labels. I use these to tag products by margin band, best sellers, seasonality or stock depth, so I can group and report on them.

On Mira Mikati, moving Google Shopping into Performance Max with feed-led optimisation was a core part of the Google work. The feed came first, the campaign second.

Structure asset groups by product theme and margin

An asset group is a set of creative (headlines, descriptions, images, video) tied to a group of products. I structure them around two questions: what does this product range look like to a customer, and how much margin does it make?

In practice that means:

Theme. Group products that share an audience and a message, such as a category or collection, so the creative actually matches what is being shown.
Margin. Keep high-margin and low-margin products apart where you can, so budget is not spread evenly across products that contribute very different amounts of profit.
Hero products. If a few products drive most of the revenue, give them their own asset group or campaign so they are not starved by long-tail products, or do not swallow the whole budget without you noticing.

I would rather have three or four clear asset groups than a dozen thin ones. Each needs enough conversion volume for Google to learn from.

Add audience signals

Audience signals tell Google where to start looking. They are suggestions, not targeting limits, but they shorten the learning period. The ones I use most:

Customer lists from Shopify, especially past purchasers
Website visitors and cart abandoners
Custom segments built from the search terms people use to find your category
Interests that match your customer, used sparingly

Past purchaser lists are usually the most useful signal for an established store.

Protect brand search

This is the step that most often goes wrong. Left alone, Performance Max will happily bid on your brand name. Those clicks were likely to convert anyway, so they inflate Performance Max’s reported ROAS and make it look better than it is.

What I do:

Run a dedicated brand search campaign so you control brand ads and can see brand performance on its own.
Exclude brand terms from Performance Max using the brand exclusion settings available at campaign level.
Check search term insights regularly to confirm branded queries are not leaking through.

On Mira Mikati, defending the brand on Google Search sat alongside Performance Max for exactly this reason.

Bid for margin, not just revenue

A pound of revenue on a 70% margin product is not worth the same as a pound on a 20% margin product. If you bid purely on revenue, Performance Max will chase whatever sells most easily, regardless of profit.

There are a few ways to correct for this:

Conversion value adjustments. Google Ads lets you set rules that raise or lower the value of a conversion based on factors such as location, device or audience. These are useful when certain customers are worth more.
Margin-aware structure. Splitting campaigns or asset groups by margin custom label and setting different ROAS targets for each.
Profit-based conversion values. Passing gross profit rather than revenue as the conversion value, where your tracking setup supports it.

The structural approach is the simplest to start with and easy to explain to a client.

Soto Gardens: adding Shopping and Performance Max to a search-only account

When I took on Soto Gardens, a UK ecommerce garden retailer, the Google account was search only. There was no Shopping and no Performance Max. Part of the work was adding Shopping and Performance Max, alongside consolidating fragmented Meta and Google campaigns into prospecting and retargeting structures.

In 2025, Soto reached £515,415 in tracked revenue from £120,212 total spend, a 4.29× blended ROAS, with purchases up 109% year on year. The full numbers are on the Soto Gardens results page.

The lesson from that account is that Performance Max did not replace search. Brand and high-intent search stayed in place, and Shopping and Performance Max added product-level coverage that the account had been missing.

What I check every week

Spend by product. Is budget concentrating on the products you expect, or drifting into low-margin lines?
Search term insights. Any brand leakage, or irrelevant themes pulling spend?
Feed health. New disapprovals, out-of-stock products still getting spend, price mismatches.
Asset performance. Low-rated assets to replace, especially images and video.
Shopify reconciliation. Does Google-reported revenue line up sensibly with what Shopify recorded? This is the number I report against.
Placement mix. Where the channel reporting allows it, check how much spend is going outside Shopping and Search.

When Performance Max is the wrong choice

I do not use it everywhere. It is a poor fit when:

Conversion volume is very low, so it cannot learn
Purchase tracking is unreliable or recently broken
The feed is thin, inaccurate or poorly categorised
You need strict control over exactly which search queries you appear on
Brand search is your main source of revenue and you cannot exclude it cleanly

In those cases a standard Shopping campaign plus well-structured search is often the better starting point. You can move into Performance Max once the foundations are there.

There is more on how I approach Google for online stores on my ecommerce paid media page.

Want a second opinion on your Performance Max setup?

If you are not sure whether Performance Max is taking credit for brand sales, or whether your feed and structure are holding it back, I can take a look. Request a free account audit and I will tell you what I would keep, change or switch off.

Questions people ask

Is Performance Max good for ecommerce?

It can be, particularly for stores with a well-maintained product feed and reliable purchase tracking. It works less well when the feed is thin, conversion data is low or tracking is unreliable.

Does Performance Max cannibalise brand search?

It can pick up branded queries if you let it, which flatters its reported ROAS. Excluding your brand terms from Performance Max and running a separate brand search campaign keeps the two apart.

How should I structure asset groups in Performance Max?

Group products by theme and by margin, so that each asset group has creative that matches the products it serves and budget can be steered towards the products that make you the most money.

When should I not use Performance Max?

Avoid it, or keep it small, when you have very low conversion volume, a poor feed, broken tracking, or a business where you need tight control over exactly which searches you appear on.