Performance Marketing

Performance Max for eCommerce: Feeds, Brand Exclusions and New-Customer Goals

Run Performance Max on profit: a feed with margin labels, brand exclusions so it stops buying your own name, and new-customer goals valued from real LTV.

TL;DR
  • Performance Max is only as good as the inputs you control: the product feed, what it's allowed to bid on and what you tell it a conversion is worth.
  • Add margin and performance labels to the feed, keep branded searches out with brand exclusions, and use a new-customer goal valued from your own repeat-order margin.
  • Judge results on new customers and store margin, using the channel performance and search terms reports, and test incrementality before you scale.

What does Performance Max actually do for an online store?

Performance Max (PMax) is a single Google Ads campaign type that serves across Search, Shopping, YouTube, Display, Discover, Gmail and Maps. For a store, the core of it is the product feed from Google Merchant Center: it powers Shopping ads, dynamic remarketing and the product formats on other placements.

You don’t choose keywords or placements. You choose three things:

  1. The inputs: your feed, your creative assets and audience signals.
  2. The boundaries: which products, brands, searches and URLs are in or out.
  3. The goal: what counts as a conversion and what each one is worth.

Left on defaults, PMax tends to find the cheapest conversions it can: people who already know your brand, recent site visitors and existing customers. Those conversions make reported ROAS look strong while adding few new customers. Every section below is about pushing it toward the conversions you’d actually pay for.

One change to know about: since October 2024, PMax no longer automatically takes priority over Standard Shopping campaigns for the same products in the same account. The campaign with the higher ad rank serves. That makes it practical to run both side by side, for example to protect specific products.

Why is the product feed the most important input?

In a retail PMax campaign, the product feed is your targeting. Google matches searches to products largely from titles, descriptions, product types and attributes. A thin feed means fewer relevant matches, whatever your bid.

Feed basics that pay back:

  • Titles that match how people search: brand, product type, key attribute (size, material, color), in that order of importance for your category.
  • Accurate GTINs where your products have them.
  • Clean product types and categories, so reporting and listing groups make sense.
  • Price and availability that match the landing page. Mismatches can get products disapproved. The same consistency matters for your on-page product schema.

On Shopify, the Google & YouTube app syncs products to Merchant Center. On WooCommerce and Magento, it’s usually an extension or a feed tool. Either way, use Merchant Center feed rules or a supplemental feed for anything the platform can’t send.

Use custom labels to bid on profit

Merchant Center offers five custom label fields (custom_label_0 to custom_label_4) that shoppers never see. They let you group products by what matters to your business:

Custom label Example values Used for
custom_label_0 margin-high, margin-mid, margin-low Separate ROAS targets by margin tier
custom_label_1 bestseller, core, long-tail Protecting budget for proven products
custom_label_2 new-launch Giving new products their own budget
custom_label_3 seasonal-q4, evergreen Turning seasonal groups on and off
custom_label_4 price-under-50, price-50-plus Spotting low-AOV products that can’t pay back

A PMax campaign has one ROAS target. If your margins differ by category, products with different break-even points shouldn’t share a campaign. We explain how to set those targets in break-even ROAS. A practical structure for most stores between $1M and $25M:

  • One campaign per margin tier, each with a target ROAS above its break-even ROAS.
  • A separate campaign for products that get few impressions, so they get a fair test instead of being starved by bestsellers.
  • Excluded products with no realistic path to profit, such as low-price items that can’t cover shipping.

Don’t split further than your conversion volume supports. Each campaign needs enough conversions for the bidding to learn, and many small campaigns usually perform worse than a few well-fed ones.

How do brand exclusions keep PMax off your branded traffic?

Branded searches convert at a high rate and low cost. If PMax can bid on them, it will, and your ROAS will look excellent. Most of those buyers were going to find you anyway, through organic results or a branded Search campaign.

Brand exclusions fix this. You add your brand to a brand list and apply it to the PMax campaign. Three details matter:

  • Scope. In PMax, brand exclusions apply to Search, Shopping and YouTube search inventory only. They don’t affect Display, Gmail or non-search YouTube placements.
  • The Shopping option. PMax offers a setting to “Allow Shopping ads on searches that mention excluded brands.” Leave it off if you want branded Shopping traffic handled elsewhere, for example by a Standard Shopping campaign, and on if you don’t run one.
  • Negative keywords. PMax now supports campaign-level negative keywords, and Google raised the limit to 10,000 per campaign in 2025. Use them for irrelevant searches and for brand misspellings a brand list might miss.

Expect reported ROAS to fall after you exclude brand. That’s the point: the number becomes closer to what the campaign actually adds. Run your brand terms in a dedicated Search campaign with its own, lower budget.

How do new-customer goals work, and which mode should you choose?

Google’s customer acquisition goal tells bidding to treat new customers differently from existing ones. There are three modes:

Mode What it does Available on
New Customer Value (“Bid higher for new customers”) Adds a value you choose to conversions from new customers Search, PMax, Shopping, Demand Gen
High Value New Customer Bids higher again for new customers you define as high value PMax, Search
New Customer Only Bids only for new customers Search, PMax, Shopping, Demand Gen

Google needs to know who your existing customers are. You define them with first-party data, such as a Customer Match list and your website tag, and Google can also build an auto-detected list from past purchases. Lists need at least 100 active members to be used. Keep yours refreshed, ideally through an automatic sync from your store or ESP, or recent buyers will be counted as new.

For most stores, New Customer Value is the safer start. New Customer Only can starve the campaign of conversions and push CPCs up, so test it rather than switching everything at once.

How much value to add: a worked example

Don’t pick the extra value by gut feel. Base it on what a new customer is worth after the first order. The numbers below are illustrative.

  • Average first order: $80, with a 40% contribution margin, so break-even ROAS on the first order is 1 ÷ 0.40 = 2.5.
  • Your cohort data shows new customers generate $20 more contribution margin from repeat orders in their first 12 months.
  • Margin is 40% of revenue, so $20 of margin is equivalent to $50 of revenue. Set the new customer value at $50, and bidding values a new customer’s first order at $130 instead of $80.

Be careful with reporting. Google shows the added amount in a “New customer lifetime value” column, and the “Original conv. value” column shows value before the adjustment. Calculate ROAS against your break-even using the original value, or you’ll be comparing an inflated number to a target built on real revenue. To get the repeat-order figure, see our guide to customer lifetime value.

The same framework has a retention side: you can also report on re-engaged lapsed customers. Keep the acquisition campaign focused on new buyers and let email and SMS handle existing ones, where they cost far less.

How do you read PMax results without being fooled?

Platform-reported ROAS is the least reliable number in the account. Use these instead.

  • Channel performance report. Under Insights and reports, it breaks results down by Search, YouTube, Display, Discover, Gmail, Maps and Search partners. The “Ads using product data” segment isolates feed-based ads. If most spend goes to placements that don’t produce new customers, that’s a signal.
  • Search terms report. Check what PMax is matching to. Branded or irrelevant queries become negative keywords.
  • New customer counts from the store. Compare new-customer orders in Shopify with what Google reports. If new customers in the store don’t rise when PMax spend does, the campaign is mostly harvesting.
  • Incrementality tests. Google offers experiments for PMax, including tests against Standard Shopping. Geo holdouts are another option. Both are forms of incrementality testing, and they answer the question ROAS can’t: what would have happened without the spend?

What should you check before scaling PMax?

  • [ ] Feed titles, GTINs, product types, price and availability are clean in Merchant Center.
  • [ ] Custom labels mark margin tier and product performance.
  • [ ] Campaigns are split by margin tier, each with a target above its break-even ROAS.
  • [ ] Your brand list is applied and the branded Shopping setting is a deliberate choice.
  • [ ] Brand terms run in a separate Search campaign.
  • [ ] The customer acquisition goal is on, with existing customers defined by a refreshed list.
  • [ ] The new customer value is based on repeat-order margin from your own cohorts.
  • [ ] ROAS is judged on original conversion value against store revenue and new-customer counts.
  • [ ] An incrementality test is planned before a major budget increase.

FAQ

Should I run Standard Shopping alongside Performance Max?

It can make sense. Since PMax no longer automatically takes priority, a Standard Shopping campaign can hold products you want to control closely, such as bestsellers or brand-sensitive items. Keep the product sets clearly separated so the two campaigns don’t compete for the same items.

How long should a new PMax campaign run before I judge it?

Give it enough time and conversions to learn, typically several weeks, and avoid changing targets or budgets daily during that period. Judge it on new customers and margin over the full period, not on the first few days of reported ROAS.

Why did my ROAS drop after adding brand exclusions?

Because branded searches were inflating it. Those buyers already knew you and would likely have bought through organic or a branded Search campaign. A lower ROAS with more new customers is usually a healthier campaign.

Turn this into a plan for your store

The Free Profit Audit scans your store, sizes your biggest profit leaks and builds a 90-day plan.

Free Profit Audit