Performance Max Best Practices eCommerce: 2026 Guide
In this guide:
- What Performance Max actually is (and what it isn't)
- The case for using PMax in eCommerce
- Performance Max best practices eCommerce: what to do before you launch
- Campaign structure: one campaign or many?
- Asset groups: not just a box-ticking exercise
- Audience signals: give the algorithm a head start
- What you can actually control in PMax
- How to read PMax performance honestly
- Frequently Asked Questions
Performance Max can work well for eCommerce, but not in the way Google's onboarding suggests. The businesses getting results from it are the ones treating it as a channel that needs tight constraints, not a set-and-forget system. Get the setup wrong and it will spend your budget efficiently on things that don't matter to your business.
We manage PMax campaigns across a range of Australian eCommerce accounts. What we keep seeing is that the setup decisions you make in the first week largely determine whether PMax is an asset or a budget drain. Most of the common mistakes happen before the campaign even goes live.
What Performance Max Actually Is (and What It Isn't)
PMax is Google's fully automated campaign type that runs ads across Search, Shopping, Display, YouTube, Discover, Gmail, and Maps from a single campaign. You give it assets, images, headlines, descriptions, a product feed, and Google's algorithm decides where, when, and to whom to show them.
Google's own Performance Max campaign overview covers the technical setup, but it won't tell you what actually works once a campaign is running. That's where Performance Max best practices for eCommerce come in.
On paper, that sounds efficient. In practice, it means you hand a significant amount of control to an algorithm that optimises for conversions as it defines them, not necessarily the conversions that matter most to your business.
It replaced Smart Shopping campaigns in 2022. If you ran Smart Shopping and it worked for you, PMax is the successor, but it comes with more complexity and less visibility into what's actually happening.
The learning curve can be expensive if you don't know what you're walking into. It's worth understanding how Google's bidding systems work before relying heavily on any automated campaign type.
The Case for Using PMax in eCommerce
PMax does have genuine strengths. When it works, it works because it has access to Google's full inventory and can find customers across multiple surfaces simultaneously. For eCommerce businesses with a solid conversion history and a well-structured product feed, it can find purchase intent in places that a standard Shopping or Search campaign wouldn't reach.
It also benefits significantly from scale. The more conversion data Google has to work with, the better the algorithm performs. If you're generating 50+ conversions per month, PMax is worth testing. Below that, the learning phase can take weeks, and the algorithm may never fully stabilise.
The Shopping placements within PMax tend to perform well for businesses with strong product imagery and competitive pricing. That's where a lot of the volume comes from in practice, and it's useful to keep that in mind when you're evaluating results.
Performance Max Best Practices eCommerce: What to Do Before You Launch
This is where most accounts go wrong. PMax isn't hard to set up. It's hard to set up well. Skipping the pre-launch checklist is how businesses end up with campaigns that spend confidently but produce nothing useful.
Sort your product feed first
PMax pulls from your Merchant Centre feed. If that feed has gaps, missing GTINs, thin product descriptions, or low-quality images, the campaign will underperform regardless of anything else you do.
Clean the feed before you touch campaign settings. Fix missing attributes, improve title formatting to match how people actually search, and make sure your images meet Google's quality requirements.
Have conversion tracking in place and verified
This sounds obvious, but we've seen accounts launch PMax before their tracking was properly confirmed. The algorithm needs purchase data to optimise toward. If it's receiving bad signals, it learns from bad signals, and then it doubles down on the wrong behaviour.
A home renovation client came to us with 18 months of campaigns running on unverified conversion tracking. Three months after we fixed it, conversions were up 41% and cost per acquisition was down 43%. The data quality was the entire problem, not the campaigns themselves.
If you're not sure your tracking is solid, read our guide on setting up Google Ads conversion tracking before going anywhere near PMax.
Set a sensible budget
Google recommends a daily budget at least 3x your target CPA, and in this case the advice is sound. PMax needs room to test.
A budget that's too tight relative to your CPA means the campaign spends the whole day in learning mode without ever gathering enough data to exit it. Underfunding PMax is one of the most reliable ways to make it look like it doesn't work when the real issue is it never had a chance to.
Campaign Structure: One Campaign or Many?
The default advice is to start with one PMax campaign. That's fine as a starting point, but it's not always where you'll end up.
If you have a large product catalogue with very different margin profiles, it's worth considering separate campaigns for your hero products versus your broader range.
PMax will find conversions wherever they're easiest, which sometimes means it spends heavily on low-margin products because those convert more easily on price. That's the algorithm doing its job. It's just not the job you wanted it to do.
You can use asset groups to segment product categories within a single campaign, but asset groups don't have their own budgets. If you need budget control at the product level, separate campaigns are the cleaner solution.
One thing to avoid: running PMax alongside a standard Shopping campaign targeting the same products. They'll compete with each other in auction, and PMax typically wins that fight, making the Shopping campaign redundant while also complicating your attribution.
If you're comparing the two, read our breakdown of Google Shopping ads vs Search ads to understand how the auction dynamics differ.
Asset Groups: Not Just a Box-Ticking Exercise
Asset groups are where your creative lives inside PMax. Most people treat them as a requirement to get the campaign live rather than a real opportunity. That's a mistake, because poor assets are one of the main reasons PMax underdelivers on placements beyond Shopping.
Google needs a range of assets to serve your ads across different placements. Horizontal images for display, square images for discovery, videos for YouTube. If you don't supply a video, Google will auto-generate one from your images. Those auto-generated videos are almost always poor quality, and you have no control over what they look like.
Supply your own videos. Even a simple 15-second product or brand video is better than what Google creates automatically. You don't need a production budget. A clean screen recording, a product walkthrough, or a customer testimonial clip will outperform a Google-generated slideshow every time.
For copy, don't just repeat the same idea across different headlines. Google tests combinations of your assets, so give it real variation to work with: different angles, different value propositions, different calls to action. If all your headlines are variations of "Shop [Product] Online", you've given the algorithm nothing to test.
Audience Signals: Give the Algorithm a Head Start
You can't control who PMax targets, but you can give it audience signals as a starting point. Think of these as suggestions rather than hard targeting. Google uses them to guide initial learning, then expands from there.
The most valuable signals are your own first-party data: customer lists, website visitors, past converters. These tell the algorithm what your actual customers look like and give it somewhere sensible to start rather than going completely broad from day one.
Interest-based audiences and in-market segments can also help, but they're a weaker signal. Your own customer data is always better. If you're running Meta campaigns alongside Google, you likely already have customer lists you can upload directly.
What You Can Actually Control in PMax
Less than you'd like. But a few levers matter more than people realise:
- Negative keywords Google added account-level negative keywords that apply to PMax. Use them. Without negatives, PMax will match to branded terms, competitor searches, and irrelevant queries. Apply your existing negative keyword list from your Search campaigns as a minimum starting point. Our guide to negative keywords covers how to build and maintain a solid list.
- Brand exclusions You can exclude your own brand terms from PMax so it doesn't cannibalise your branded Search campaigns. This is worth doing almost universally. PMax claiming conversions on branded queries you would have gotten anyway inflates its reported performance.
- URL exclusions You can exclude specific URLs from receiving traffic. Useful for blocking account login pages, campaign-specific landing pages, or pages that don't belong in the ad experience.
- Placement exclusions Account-level placement exclusions apply to PMax display and video inventory. If you have categories of placements you know don't convert for your business, excluding them is worth doing.
How to Read PMax Performance Honestly
Reporting in PMax is limited, and that's a genuine frustration. You can see asset performance labels (Best, Good, Low) but not granular data on which placements, queries, or audiences are driving results. Google has added some search term visibility over time, but it's still partial compared to what you'd see in a standard Search campaign.
What you can do: compare PMax performance against your previous Shopping or Search campaigns on the same products, using the same conversion definitions. Don't just look at total conversions. Look at ROAS, average order value, and whether the sales being generated are the ones that actually matter for your margins.
Also watch for what's sometimes called the "PMax attribution problem." If you're running PMax alongside other campaigns, it tends to claim last-click credit on branded queries and retargeting touchpoints that other campaigns drove.
Factor that in when you're evaluating whether PMax is genuinely generating new demand or just taking credit for conversions that were already coming. Knowing how to tell if your Google Ads are working is especially important with a campaign type this opaque.
Following Performance Max best practices for eCommerce means reading the data critically, not just taking Google's reported numbers at face value.
Frequently Asked Questions About Performance Max for eCommerce
Does Performance Max replace Shopping campaigns in Google Ads?
For most eCommerce accounts, yes. PMax includes Shopping placements as part of its inventory, so if you're running it with a product feed it effectively absorbs what a standard Shopping campaign would have done. Some advertisers keep separate Shopping campaigns for specific products or tighter budget control, but in most cases PMax has taken over that role since Smart Shopping was retired in 2022.
How long does Performance Max take to learn?
Google's official learning period is 6 weeks, and that's roughly accurate in practice. During this time the algorithm is gathering data and performance will look inconsistent, sometimes good, sometimes not. Resist the urge to make big changes during this period, as each significant adjustment resets the learning clock and extends the instability.
Should I run Performance Max and Search campaigns at the same time?
Yes, they can coexist. The key is making sure PMax has brand exclusions applied so it doesn't compete with your branded Search campaigns. For non-branded search traffic, PMax and Search can both run. Just monitor for overlap in your search terms reports and review attribution data regularly to understand what each campaign is actually contributing.
What's a realistic ROAS target for Performance Max in Australia?
It depends on your category and margins. Most eCommerce clients we work with target somewhere between 3x and 6x ROAS, with lower targets in competitive categories like fashion and homewares, and higher in niche or higher-margin products.
Set your target ROAS based on what's genuinely profitable for your business, not what Google suggests as a default benchmark, which is often too conservative to be useful.
What happens if I don't supply a video asset for PMax?
Google will auto-generate one from your existing images. These auto-generated videos are almost always low quality and you have no control over how they look. Supplying even a basic 15-second video, a product showcase, a brand clip, a customer testimonial, will outperform anything Google creates automatically and gives you control over how your brand appears on YouTube placements.
Want to Know If Your PMax Campaign Is Actually Working?
PMax is one piece of a properly structured eCommerce Google Ads account. Getting it right takes more than following Google's default recommendations. It takes someone actually looking at the data regularly and making calls based on what the account is telling them.
We manage campaigns for Australian eCommerce businesses and can take a look at whether your current Performance Max setup is leaving money on the table.