Performance Max is Google’s most powerful campaign type – and its most misunderstood. It uses AI to automatically distribute your ads across Google Search, Shopping, Display, YouTube, Gmail, and Discover, optimising in real-time for conversions. When it works well, it is extraordinary. When set up poorly, it wastes your budget on irrelevant placements and cannibalises your existing campaigns.
What’s in This Article
Most Shopify stores either avoid Performance Max because it feels like a black box, or they launch it with default settings and hope for the best. Neither approach captures the full potential. The brands getting 5-8x ROAS from Performance Max are the ones who understand how to structure asset groups, feed the algorithm quality signals, and integrate PMax with their existing campaign structure.
Here is the complete setup and optimisation guide for running Performance Max campaigns that genuinely drive profitable sales for your Shopify store.
How Performance Max Actually Works (The Non-Technical Version)
Think of Performance Max as a single campaign that shows your products and ads everywhere across Google’s entire network. Instead of creating separate campaigns for Shopping, Search, Display, and YouTube, PMax handles all of them in one place. Google’s AI decides which channel, which ad format, and which audience to show your ads to – based on who is most likely to convert.
The AI learns from your conversion data. The more conversions it sees, the better it gets at finding similar buyers. This is why PMax tends to improve over time – the algorithm needs 4-6 weeks of data to optimise effectively. Do not judge a PMax campaign in its first two weeks; the learning period is real and necessary.
Your job is to give the AI the best possible inputs: a clean product feed (this feeds the Shopping placements), high-quality creative assets (this feeds Display and YouTube), strong audience signals (this tells the AI where to start looking), and accurate conversion tracking (this teaches the AI what success looks like).
Setting Up Asset Groups for Maximum Performance
Asset groups are the building blocks of Performance Max. Each asset group contains the creative assets (images, headlines, descriptions, videos) and the product listings that Google uses to create your ads. How you structure your asset groups determines how effectively Google spends your budget.
Create separate asset groups for each product category or marketing theme. Do not dump everything into one asset group – that is like putting every product in your store on a single shelf. A skincare store might have asset groups for: Cleansers, Serums, Moisturisers, Gift Sets, and Sale Items. Each group gets tailored creative that speaks specifically to shoppers in that category.
For each asset group, provide the maximum number of assets Google allows:
- Images (up to 20). Include a mix of product shots, lifestyle images, and square/landscape/portrait ratios. Google will test different images across different placements and keep the winners. More variety gives the algorithm more options to optimise.
- Headlines (up to 15). Mix keyword-rich headlines (“Best Vitamin C Serum Australia”), benefit-led headlines (“Brighter Skin in 14 Days”), and brand headlines (“Trusted by 10,000+ Aussie Women”). Keep them under 30 characters each.
- Long headlines (up to 5). These appear on larger ad formats. Include your strongest selling proposition in 90 characters or less.
- Descriptions (up to 5). Complement your headlines with specific benefits, social proof, and calls to action. “Free shipping over $80 AUD” and “4.8 star average from 2,000+ reviews” are effective descriptors.
- Videos (up to 5). If you have product videos, add them. If not, Google will auto-generate videos from your images – and these auto-generated videos are usually terrible. Even a simple 15-second product slideshow you make in Canva is better than Google’s auto-generation.
Audience Signals: Guiding the AI (Not Controlling It)
Audience signals in Performance Max are suggestions, not targeting restrictions. You are telling Google’s AI “start looking for customers who look like this” – but the AI will expand beyond those signals if it finds converting audiences elsewhere. This distinction matters because many advertisers treat signals like targeting and get frustrated when their ads reach unexpected audiences.
The most effective audience signals for Shopify stores are:
- Customer Match lists. Upload your customer email list. This tells the AI “find people who look like my existing customers.” This is your strongest signal because it is based on actual buyer data.
- Website visitors. Add your remarketing audiences. The AI will use these to understand who engages with your store and find similar users.
- Custom segments. Create segments based on search terms your customers use and websites they visit. For a skincare brand: people who searched “vitamin C serum,” “Australian skincare,” and visitors of competitor websites.
- In-market audiences. Add Google’s pre-built audiences for people actively researching purchases in your category.
The PMax + Standard Shopping Strategy
The highest-performing Google Ads accounts for Shopify stores do not use Performance Max alone. They use PMax alongside Standard Shopping campaigns in a complementary structure. This gives you the AI-powered reach of PMax plus the granular control of Standard Shopping.
The strategy is straightforward. Create a Standard Shopping campaign for your top 20% of products – your proven bestsellers where you want maximum control over bids, budget, and negative keywords. Then create a Performance Max campaign to cover everything else – your broader catalogue, new products, and long-tail discovery.
Standard Shopping takes priority over PMax for Shopping placements (when the same product is in both campaigns, Standard Shopping wins the auction first). This means your bestsellers get the control they need, while PMax handles the wider catalogue and discovers new converting audiences through Display, YouTube, and Discovery placements.
Also maintain a dedicated Brand campaign (Search) to protect your brand terms. PMax will bid on your brand keywords if you do not – and brand terms have artificially high ROAS that inflates PMax reporting. A separate brand campaign ensures accurate performance measurement.
Optimisation and Troubleshooting
Performance Max requires patience and specific optimisation tactics:
Give the campaign 4-6 weeks before making major changes. The learning period is real. Frequent changes reset the learning phase and prevent the algorithm from optimising. During the first month, resist the urge to change budgets, pause asset groups, or adjust ROAS targets.
Monitor your Search Terms report (available through the Insights tab). PMax will bid on search terms you may not expect – some will be irrelevant. Add negative keywords at the account level to block wasted spend on irrelevant queries. This is the most important ongoing optimisation task for PMax.
Check your asset performance ratings monthly. Replace any assets rated “Low” with new creative. The algorithm deprioritises low-performing assets, but removing them and adding fresh alternatives gives the AI new material to test.
If ROAS is below target, check these common issues: your product feed may have data quality problems (see our Google Merchant Center article), your conversion tracking may be inaccurate (verify in Google Ads), your audience signals may be too broad, or your budget may be too low for the algorithm to optimise effectively (PMax needs at least $50-100/day to learn properly).
Reading PMax Reports: Where the Data Actually Lives
The most common complaint about Performance Max is that it is a black box. It is less opaque than it looks. Google has quietly added most of the reporting operators asked for, it is just buried three clicks deeper than anywhere else in the interface.
Asset group reporting. Campaign, then Asset Groups, gives you conversions and value split by asset group. This is the single most useful view in PMax, because it tells you which slice of your catalogue the algorithm is actually finding buyers for. If one asset group is carrying 80% of the value, that is not a failure, it is a signal to split it out into its own campaign with its own budget rather than letting it subsidise the others.
Asset performance ratings. Inside each asset group, individual headlines, descriptions, images and videos are rated Low, Good or Best. Replace every Low asset each month. It is a 20 minute job that consistently moves campaign performance, because the algorithm can only combine what you have given it. Aim for at least 5 headlines rated Good or Best before you go looking for structural fixes.
Search themes and the search terms insight report. Insights, then Search Terms, groups the queries PMax served against into themes. You will not get the full query-level detail Search campaigns give you, but you will see the categories. Two things to look for: brand queries eating your budget (which inflate reported ROAS and mask weak prospecting), and irrelevant categories that need a negative keyword list applied at account level.
The brand traffic problem is the big one. Left alone, Performance Max will happily spend a large share of budget on people searching your brand name, then report a 12x return. Those customers were coming anyway. Apply a brand exclusion list at campaign level, run a separate brand Search campaign to catch that traffic cheaply, and watch what happens to your PMax ROAS. It will drop, often from 8x to 4x, and that lower number is the truthful one. Judging PMax on a brand-inflated ROAS is the fastest way to over-invest in a channel that is not actually acquiring new customers.
For the same reason, sense-check PMax against a channel-agnostic number rather than its own reporting. A marketing efficiency ratio that compares total revenue to total ad spend will tell you within a fortnight whether a PMax scale-up added revenue or simply reallocated it.
Scaling Performance Max Without Wrecking Your Blended Return
PMax scales differently to Meta. Push the budget too fast and the algorithm re-enters a learning phase, performance craters for a week or two, and most founders panic and pull it back before it recovers. Then they do it again a month later.
- Move budget in 20 to 30% steps, no more than once a week. A jump from $100 to $150 a day is usually absorbed. A jump from $100 to $400 will reset learning. Make the change early in the week so you can read a full 7 days before deciding anything.
- Adjust target ROAS by 10% at a time. Dropping your target from 500% to 300% in one move tells the algorithm to buy completely different traffic. Step it down, hold for 14 days, read the result, step again.
- Wait out the learning period properly. PMax needs roughly 2 to 6 weeks and at least 30 conversions to stabilise after a material change. Judging it at day 4 is judging noise.
- Add inventory before you add budget. More budget against the same 40 products just raises your CPC. Expanding the catalogue, adding new asset groups, or opening a new product category gives the algorithm somewhere new to spend.
- Change one thing at a time. Budget, target ROAS, asset groups and feed are four separate levers. Pull two at once and you learn nothing about either.
The seasonal wrinkle for Australian stores: PMax’s learning cycle does not fit neatly inside a Click Frenzy or BFCM window. If you plan to spend heavily in late November, your budget ramp needs to start in the first week of October so the campaign is out of learning and stable before the traffic arrives. Use seasonality adjustments in Google Ads to warn the algorithm about a known conversion-rate spike rather than letting it discover the spike on day one of the sale.
Two things to have working before you scale at all. Your feed needs to be clean, because PMax draws from the same Merchant Center data as Shopping and a weak feed caps the ceiling on both. And your Shopping campaign structure needs to be settled, so you can see clearly which channel a lift came from.
Then hold the whole account to one number. Platform-reported ROAS across PMax, Shopping, Brand and Meta will always sum to more revenue than your Shopify dashboard shows, because every platform claims the same conversion. Reconcile against blended ROAS monthly. If blended is flat while PMax reports a record month, the campaign is claiming credit for demand you already had.
The Compound Effect of Performance Max Mastery
When Performance Max is set up correctly and given time to learn, it becomes one of the most efficient acquisition channels available. The AI finds converting audiences you would never discover through manual targeting. It tests creative combinations you would never think to try. And it optimises 24/7 across every Google property, capturing demand wherever it exists.
One eCommerce Circle member restructured their Google Ads from a messy collection of campaigns into the PMax + Standard Shopping + Brand framework described here. Within 8 weeks, their overall Google Ads ROAS improved from 4.2x to 6.8x, and revenue from Google increased 42% on the same budget. The combined approach outperformed both PMax-only and Shopping-only strategies they had tried previously.
Google Ads strategy, including Performance Max, is a core coaching area inside eCommerce Circle. We help hundreds of Aussie Shopify founders structure campaigns, clean up feeds, and build the reporting that shows true channel profitability. If your Google Ads still feel like a black box, let’s talk.
