For an ecommerce store, Performance Max is really Shopping with everything else bolted on. The product feed does most of the work, and how you structure the campaign around it decides whether PMax helps or just spends.
The feed is the campaign
In a retail PMax campaign, your Merchant Center product feed is the engine. Titles, images, prices, and product types are what Google matches to queries and shoppers, far more than your text assets.
A clean, well-optimized feed will outperform clever ad copy every time, so that's where the effort belongs.
Setup, step by step
- Connect Merchant Center to Google Ads and fix feed errors before you spend a cent.
- Create a PMax campaign with your feed, and add asset groups themed by product category.
- Use listing groups to split products the way you'd split Shopping, by category, brand, or margin.
- Add strong images and video per asset group, so the non-Shopping surfaces have something to work with.
- Set a conversion goal with value, ideally real revenue, so PMax optimizes to money not orders.
Structure for margin, not just sales
Default PMax treats every sale as equally good, but your margins don't. Group products so you can steer budget toward the ones that actually make money:
- Separate high-margin and low-margin products into different asset groups or campaigns.
- Split your best sellers out, so they don't get buried under the long tail.
- Send margin-aware conversion values where you can, not just order revenue.
Rein in the brand and the junk
Retail PMax loves to spend on brand searches and cheap placements. Apply a brand exclusion so you're not paying to capture people already looking for you, and use placement exclusions to cut the worst Display and app inventory.
Both are buried in the settings, and both save real money on a store with volume.
Where ecommerce PMax breaks
- A messy feed, so Google shows the wrong products for the wrong queries.
- One giant asset group, so best sellers and dead stock share a budget.
- Optimizing to conversions instead of value, so PMax chases cheap, low-margin orders.
- Ignoring brand exclusions, so the campaign looks efficient on borrowed brand demand.
Measure against your own numbers
PMax will report healthy ROAS. Reconcile it against your store and your margins, because a campaign that's profitable on revenue can be losing money once you account for what each product actually earns. Margin is the number PMax can't see for you, so it has to be the number you watch most closely.
Get the conversion setup right first
None of the structure helps if the campaign optimizes to the wrong thing. Make sure purchase conversions carry real revenue values, that you're not double counting across GA4 and the Google tag, and that refunds flow back so ROAS isn't inflated.
A retail PMax fed bad conversion data will confidently spend toward the wrong products.
When to split into multiple campaigns
Asset groups give you some separation, but sometimes you need real budget control, and that means separate campaigns:
- Split brand and non-brand when exclusions alone don't give you clean separation.
- Split by margin tier when high and low margin products need different ROAS targets.
- Split hero products out when you don't want them competing with the long tail for budget.
The tradeoff is learning. Each campaign needs enough conversions to exit the learning phase, so don't split so finely that nothing gets the volume to optimize on.
For ecommerce, Performance Max lives and dies by the feed. Optimize the feed, structure for margin, cut the brand and the junk, and Shopping-with-everything-bolted-on starts pulling its weight instead of just spending.
Want a stronger data analyst role or a raise? Grab the FREE Product Analyst Playbook and get the exact roadmap to your next offer.
