---
title: "Google Performance Max: What It Is and How It Works"
description: "Google Performance Max explained: how asset groups and signals steer one budget across Google, why campaign ROAS hides losing products, and how to fix it."
canonical: "https://www.productmetrics.io/blog/google-performance-max"
pageType: article
language: en
publisher: "Product Metrics"
author: "Berend Vrakking"
datePublished: 2026-11-02
dateModified: 2026-11-02
image: "https://www.productmetrics.io/og/blog/google-performance-max.png"
---

> Content index: https://www.productmetrics.io/llms.txt

# Google Performance Max: What It Is and How It Works

## Key takeaways

- Performance Max lets Google move one budget between its ad channels. Audience signals and search themes guide where it starts; they never limit who sees your ads.
- Campaign ROAS hides products: in our worked example, a campaign at a ROAS of 5 earns €750 while its bestsellers lose €1,050.
- Give each margin band its own campaign and a target above its break-even ROAS, 1 ÷ contribution margin: a 25% margin needs at least 400%.

Google [Performance Max](https://www.productmetrics.io/glossary/performance-max "Performance Max is a goal-based Google Ads campaign type that uses automation to reach customers across Google advertising channels.") is a Google Ads campaign type that spends one budget across Search, Shopping, YouTube, Display, Discover, Gmail and Maps, and lets Google decide where each euro goes.

For a retailer, the hard part comes afterwards: knowing which products it spent on.

In the worked example below, a campaign reporting a [ROAS](https://www.productmetrics.io/glossary/roas "Return on ad spend (ROAS) is the revenue your ads generate for every euro you spend on them.") of 5 loses €1,050 on the bestsellers that take 60% of its budget.

I checked each Google claim against Google's own pages on 2 November 2026.

## How does Google Performance Max work?

Performance Max works from ingredients, not finished ads: you supply products, creative and a conversion goal, and Google assembles ads for each channel and moves the budget to wherever it predicts the next conversion.

Bidding follows the goal you set, such as purchase value with a [target ROAS](https://www.productmetrics.io/glossary/target-roas "Target ROAS is the average conversion value per unit of advertising cost that you ask Google Ads to aim for.").

The structure has three levels. The campaign holds the budget and the goal. An asset group is a creative brief for one theme. Assets are the materials Google mixes into ads.

| Part of an asset group | What it holds | Where Google uses it |
|---|---|---|
| Products | A listing group of Merchant Center products | Shopping ads and product-based ads on other channels |
| Text | Headlines, long headlines, descriptions | Search ads, Display and Discover cards |
| Images and logos | Product, lifestyle and brand images | Display, Discover, Gmail |
| Videos | Your own video files | YouTube and Shorts placements |
| Signals | Audience signals and search themes | A starting point for Google's targeting |

Say you sell outdoor gear. Before Performance Max, your boots ran in Shopping, brand terms in Search and a YouTube video in its own campaign.

Now one campaign can show the same boots as a Shopping listing, a Search ad or a Discover card, depending on where Google expects the sale.

Weak or missing assets limit where the campaign can serve well. Supply no video, and Google has far less to work with on YouTube.

### Do audience signals limit who sees your ads?

No, audience signals don't restrict delivery: Google describes them as suggestions that "help Google AI optimize for your selected goals" ([Google's help page on how Performance Max works](https://support.google.com/google-ads/answer/10724492)).

Search themes work the same way. Both tell Google where to start looking, not where to stop.

So build signals from your best customers, such as a past-purchaser list, rather than a vague interest category. A tight-looking signal doesn't mean a tight audience.

Asset groups split best where the message changes. One group for premium running shoes with fit-and-cushioning copy and a second for clearance with price-led copy makes sense. One group per shoe does not.

When an asset looks weak, check its serving status and impressions before deleting it. An asset that barely served hasn't failed creatively; it hasn't had a fair chance yet.

### Where did Performance Max come from?

Google launched Performance Max to all advertisers on 2 November 2021, as the successor to Smart Shopping and Local campaigns ([Google's launch announcement for Performance Max](https://blog.google/products/ads-commerce/performance-max/)). The reporting retailers asked for came later.

| Date | Change | Source |
|---|---|---|
| November 2021 | Global launch to all advertisers | Google launch announcement |
| 2022 | Smart Shopping and Local campaigns upgraded to Performance Max | Google launch announcement |
| April 2025 | Full search terms reporting; channel performance reporting announced | Google Ads blog, 30 April 2025 |

At launch, Google reported an average 13% increase in total incremental conversions at a similar cost per action for advertisers using Performance Max, from its own July–September 2021 data. It is Google's own average across its advertisers, so test it against your own account.

## Performance Max vs Standard Shopping: which should you use?

Use Performance Max for reach across Google's channels when the campaign has enough conversions to learn from, and Standard Shopping when you need tight control over specific products and bids.

Many retailers run both, with each product in one campaign type only.

| Criterion | Performance Max | Standard Shopping |
|---|---|---|
| Channels | Search, Shopping, YouTube, Display, Discover, Gmail, Maps | Shopping results on Search, the Shopping tab and Images |
| Control unit | Asset group (creative, signals, products) | Product group inside an ad group |
| Keywords | No keyword targeting; search themes act as hints | No keyword targeting; negative keywords allowed |
| Bidding | Automated toward target ROAS or conversion goals | Manual CPC or target ROAS |
| Budget | One budget shared across all channels | Shopping placements only |
| Best fit | Strong feed, enough conversion volume | A few products that carry the account |

Three rows decide most cases. The first is the control unit: asset groups suit retailers who think in themes and landing pages; product groups suit those who think product by product.

Bidding is the second. Performance Max steers with targets rather than bids, so it needs data. In smec's State of Performance Max 2025 study of 4,000+ campaigns, 54% of campaigns with fewer than 30 conversions a month missed their ROAS target.

**Figure: Share of Performance Max campaigns below their ROAS target, by monthly conversions.** Campaigns past 1,000 conversions a month miss their ROAS target about a seventh as often as campaigns under 30.

![Bar chart of the share of Performance Max campaigns below their ROAS target by monthly conversions: 54% under 30, 20% at 90–150, 14% at 150–300 and 8% past 1,000.](https://www.productmetrics.io/images/blog/google-performance-max/pmax-below-target-by-conversions.webp)

| Conversions a month | Campaigns below their ROAS target |
| --- | --- |
| 0–30 | 54% |
| 30–60 | 33% |
| 60–90 | 27% |
| 90–150 | 20% |
| 150–300 | 14% |
| 300–500 | 10% |
| 500–1K | 9% |
| 1K+ | 8% |

_Source: smarter-ecommerce (smec), State of Performance Max 2025 (4,000+ campaigns); grouped by monthly conversions by Product Metrics_

The miss rate halves by 60–90 conversions a month and falls to 8% past 1,000.

Below about 30 a month, Standard Shopping's manual control is often the safer start.

Where ads run is the third. Performance Max finds demand you didn't know existed; Standard Shopping keeps intent high and reporting simple. A messy [product feed](https://www.productmetrics.io/glossary/product-feed "A product feed is a file or data source that lists your products, one per row, with attributes such as ID, title, price and availability.") makes both harder to steer, so check how [products are categorised in Merchant Center](https://www.productmetrics.io/blog/google-shopping-category) before splitting anything.

For example, a retailer with 2,000 products and a clean feed can let Performance Max spread them across channels.

A retailer with a handful of high-margin lines that each need their own bid gets that control directly in Standard Shopping.

## What should you set up before launching Performance Max?

Set up the feed, the [conversion value](https://www.productmetrics.io/glossary/conversion-value "Conversion value is the amount you tell Google Ads a conversion is worth, such as sales revenue or profit margin.") and the [custom labels](https://www.productmetrics.io/glossary/custom-labels "Custom labels are five attributes in a product feed, custom_label_0 to custom_label_4, that you fill with your own text to group products in campaigns.") before Performance Max gets any budget, because the campaign optimises toward whatever data you give it and can't repair bad inputs. Work through the steps in order.

**Set up the inputs before Performance Max gets a budget**

1. **Link and approve the feed.** Connect Google Ads to Google Merchant Center and clear every disapproval, so each product can compete.
2. **Match product IDs.** Check that the IDs in the feed are the IDs in campaign reporting, or your per-product analysis will have gaps.
3. **Send order value.** Use a purchase conversion that carries the order value, so ROAS reflects revenue rather than order count.
4. **Add custom labels now.** Create labels for margin band and performance segment, so the campaign split you'll want later already exists.
5. **Build asset groups by theme.** One asset group per theme, landing page or value proposition, never one per product.
6. **Record a baseline.** Note spend, conversions and contribution profit for the last 30 days, so you can tell a profit gain from a moved ROAS.

The conversion value matters most. If your tags miss purchases, Performance Max learns from part of your revenue; [setting up and checking enhanced conversions](https://www.productmetrics.io/blog/enhanced-conversions-vs-server-side-tracking) is worth doing before you trust the numbers.

If a product is missing from the feed view, fix its eligibility before judging its results. A product with no data hasn't failed. It never got to compete.

## Why can a high Performance Max ROAS hide losses?

A high Performance Max ROAS can hide losses because ROAS compares revenue with ad spend and ignores margin, and the campaign average blends products with very different margins.

A campaign can hit its target while its biggest spenders lose money on every order.

Here is a worked example with illustrative data. One campaign spends €10,000 and returns €50,000, a ROAS of 5.

Its bestsellers take 60% of the spend.

| Product group | Contribution margin | Ad spend | ROAS | Profit after ads |
|---|---|---|---|---|
| Bestsellers | 15% | €6,000 | 5.5 | −€1,050 |
| Mid-range | 30% | €3,000 | 4.5 | €1,050 |
| High-margin lines | 50% | €1,000 | 3.5 | €750 |
| Whole campaign | 21.5% | €10,000 | 5 | €750 |

**Figure: One Performance Max campaign at a ROAS of 5: profit after ads by product group.** The bestsellers have the highest ROAS, 5.5, and are the only group losing money, because a 15% margin needs 6.67 to break even.

![Bar chart, illustrative data, of profit after ads in one campaign at a ROAS of 5: bestsellers at ROAS 5.5 lose €1,050, mid-range products earn €1,050 and high-margin lines €750.](https://www.productmetrics.io/images/blog/google-performance-max/profit-behind-campaign-roas.webp)

| Product group (margin, ROAS) | Profit after ads |
| --- | --- |
| Bestsellers (15%, ROAS 5.5) | -€1,050 |
| Mid-range (30%, ROAS 4.5) | €1,050 |
| High-margin lines (50%, ROAS 3.5) | €750 |

Reference line: Break-even (€0).

_Source: Illustrative data; profit after ads calculated as ad spend × ROAS × contribution margin − ad spend_

Per €100 of ad spend, the gap is wider still.

At these averages, €100 on the bestsellers loses €17.50, while €100 on the high-margin lines earns €75.

Google moves budget toward the highest ROAS, so without margin data it moves budget toward the loss.

The other half of the blind spot is incrementality.

A product can look profitable because it captures brand searches and returning customers who'd have bought anyway: that is demand capture, not growth, and campaign ROAS can't tell the two apart.

### Which Performance Max reports show product performance?

Performance Max reports by asset group, asset and product, not by ad group. Querying ad-group resources returns no Performance Max data, because the campaign organises assets and targeting through asset groups ([Google's developer guide to Performance Max reporting](https://developers.google.com/google-ads/api/performance-max/reporting)).

| Report | Question it answers |
|---|---|
| Asset group performance | Which themes and audiences drive conversions |
| Asset performance and top combinations | Which headlines, images and videos serve and convert |
| Channel performance | Where each campaign delivers: Search, YouTube, Display and more |
| Search terms | Which searches triggered your ads |
| Product reporting (Merchant Center linked) | Which products earn spend and conversions |
| Merchant Center diagnostics | Whether weak results come from demand or from feed problems |

The channel and search terms views are recent. Google added full search terms reporting and announced channel performance reporting on 30 April 2025 ([Google's announcement of channel performance reporting](https://blog.google/products/ads-commerce/channel-performance-reporting-coming-to-performance-max/)).

They show where ads ran, not whether those sales were incremental.

A weekly product check takes five steps:

1. Pull product reporting for the last 30 days.
2. Join product IDs to your Merchant Center feed, so each row carries price, category and labels.
3. Add contribution margin next to ROAS for your top 50 products.
4. Flag products with spend but no conversions, and run Merchant Center diagnostics on them.
5. Move budget only after those checks.

Doing this by hand works for 50 products. Across thousands, every week, it is where most Performance Max accounts stop looking, and where [seeing Performance Max spend per product](https://www.productmetrics.io/performance-max) saves the most time.

> **Get more performance out of Performance Max.** Give your best-performing products more priority, segmented on your own data. [Analyse your products for free](https://app.productmetrics.io/) Log in with Google, no credit card.

## How do you optimise Performance Max for profit?

Optimise Performance Max for profit by giving each margin band its own campaign with a target ROAS above that band's break-even, and by moving products between campaigns on their own results.

Google sets target ROAS per campaign, not per product, so the campaign split is how different products get different targets.

### Set a break-even ROAS per margin band

Break-even ROAS is 1 ÷ contribution margin, where contribution margin is revenue minus product cost, shipping, payment fees and expected returns.

A 25% margin breaks even at a ROAS of 4, a target ROAS of 400%; check yours with the [break-even ROAS calculator](https://www.productmetrics.io/break-even-roas-calculator).

| Margin band | Contribution margin | Break-even ROAS | Lowest target ROAS |
|---|---|---|---|
| Low | 15% | 6.67 | 667% |
| Medium | 30% | 3.33 | 333% |
| High | 50% | 2 | 200% |

In POAS terms, POAS = ROAS × contribution margin and break-even POAS is 1.0; [POAS vs ROAS](https://www.productmetrics.io/blog/poas-vs-roas) explains when to bid on profit instead. Set each band's target above its break-even line, with the profit you want to keep on top.

**Break-even verdict calculator.** Enter the price excl. VAT, cost of goods, other variable costs per order and your current ROAS to see whether ads pay for themselves.

- Contribution margin = (price − cost of goods − other variable costs) ÷ price
- Break-even ROAS = 1 ÷ contribution margin
- POAS = ROAS × contribution margin (1 is break-even)
- Left per €100 of revenue after ads = 100 × contribution margin − 100 ÷ ROAS
- Target ROAS (%) = 100 ÷ (contribution margin − profit kept), using fractions for margin and profit kept

Example (illustrative data):

| Input or result | Example |
| --- | --- |
| Price excl. VAT | €100.00 |
| Cost of goods | €60.00 |
| Other variable costs per order | €15.00 |
| Contribution per order | €25.00 |
| Contribution margin | 25.0% |
| Break-even ROAS | 4 |
| Current ROAS | 5 |
| POAS | 1.25 |
| Left per €100 of revenue after ads | €5.00 |
| Target ROAS to keep 0% of revenue as profit | 400.00% |

Verdict: Above break-even: ROAS 5 is over the 4 you need. After ads you keep €5.00 of every €100 of revenue. With no profit kept, target ROAS is 400.00%, equivalent to the break-even ROAS. A bidding target does not guarantee break-even; compare achieved return on the same revenue and cost basis.

_A product with a 25% contribution margin breaks even at a ROAS of 4, so a ROAS of 5 keeps €5 per €100 of revenue._

Try your own numbers in the [break-even ROAS calculator](https://www.productmetrics.io/break-even-roas-calculator).

Gross margin overstates what an order is worth. In fashion and footwear, returns alone can turn a healthy-looking margin into a thin one.

### Segment products and write the segment to a custom label

Segment products by ad clicks and return, then write each segment to a Merchant Center custom label, so a listing group can put each segment in the right campaign.

Product Metrics' [Product Segmentation](https://www.productmetrics.io/product-segmentation) does this: it segments every product on clicks and ROAS or POAS, and writes the segment to `custom_label_0` through a supplemental feed, only for the labels you approve.

| Segment | Clicks and return | Where it goes in Performance Max |
|---|---|---|
| Stars | High clicks, high return | Its own campaign, with more priority |
| Question Marks | High clicks, medium return | The catch-all campaign |
| Cash Cows | Low clicks, high return | The catch-all campaign |
| Dogs | High clicks, low return | A separate campaign with a stricter target, or excluded |
| Drainers | Low clicks, no return | The catch-all; mostly too little data to judge, so don't exclude them for low spend |
| Ghosts | No activity | Check the feed: they aren't getting impressions |

Then split campaigns by label, and give each product to one campaign type.

Products that need their own bid go to Standard Shopping; the rest go to Performance Max. [Structuring Shopping and Performance Max by profit](https://www.productmetrics.io/blog/product-segmentation-google-shopping-pmax) covers the split, and [how to exclude products from Performance Max](https://www.productmetrics.io/blog/performance-max-for-ecommerce) the listing-group steps.

### Test whether the sales are incremental

Test incrementality once a quarter with a holdout, because attributed ROAS can't separate new sales from sales that were coming anyway.

Exclude brand terms, split new from returning customers, and judge the result on contribution profit.

[A Shopify merchant's guide to causal revenue](https://hello.quikly.com/blog/incrementality-testing) explains the holdout logic in merchant terms.

The [Product Metrics guide to incrementality testing](https://www.productmetrics.io/blog/incrementality-testing) applies it to product campaigns.

A lower reported ROAS can mean a better business when it reflects a deliberate shift toward new customers.

### Check three gates before scaling the budget

Scale Performance Max only when you can see each product's results, know each product's break-even and have tested incrementality.

Added budget spreads across whatever the campaign already spends on, including products you can't judge yet.

| Gate | Question | If the answer is no |
|---|---|---|
| Visibility | Can you see spend and conversions per product? | Build product reporting and join IDs to the feed |
| Economics | Do you know each product's contribution margin and break-even ROAS? | Finish the margin work first |
| Incrementality | Have you excluded brand terms or run a holdout? | Test before adding spend |

Once all three pass, scale in steps and review product reports weekly.

Change one thing at a time: re-segmenting constantly restarts learning and muddies every comparison.

## Final thoughts

Performance Max is good at finding conversions and blind to what they're worth. The setup works when spend migrates toward high-margin products week after week, and the clearest sign is contribution profit growing while reported ROAS holds steady or dips.

If ROAS climbs and profit stays flat, Google is getting better at harvesting demand you already owned.

The first move is small: one column of contribution margin in the product report. The products that deserve more priority usually show themselves.

If your weekly product report turns into a dashboard the whole team opens, [Pazi's guide to ecommerce dashboards](https://blog.pazi.ai/p/9c44efc9-69a6-47b4-97d7-6ba40bf86609/) is a useful reference for deciding what belongs on one screen. Keep yours simple.

Before you scale Performance Max on an account that has never had a structured review, it can help to [audit your ecommerce Google Ads campaigns](https://notfair.co/google-ads-audit/ecommerce), so structural problems surface while they're still cheap to fix.

## Frequently asked questions

### Can you add negative keywords to Performance Max?

Yes. Performance Max accepts negative keywords and brand exclusions, which stop it matching those searches on Search and Shopping inventory. They filter searches, not products: to take a product out of Performance Max, exclude it in the asset group's listing group.

### What happens when Performance Max and Standard Shopping include the same product?

The two campaigns compete on Ad Rank for that product, and Performance Max doesn't automatically win (Google Ads Help, campaign interaction). Give each product to one campaign type, so you know which campaign's numbers describe it.

## Keep reading.

- [Google Ads Campaign Types for Ecommerce: Search, Shopping and PMax](https://www.productmetrics.io/blog/google-ads-campaign-types-ecommerce)
- [Google Ads Negative Keywords: Match Types and a Profit Test](https://www.productmetrics.io/blog/google-ads-negative-keywords)
- [Google Ads optimisation for Shopping and PMax](https://www.productmetrics.io/blog/google-ads-optimize)

---

Written by Berend Vrakking, founder of Product Metrics. Last updated 2026-11-02.

1M+ products analysed · Nominated at the Dutch Search Awards for Best Measurement Campaign (2023) and Best Use of Data (2024)

HTML version: https://www.productmetrics.io/blog/google-performance-max