Performance Max for Ecommerce in 2026: The Complete Guide

If you have been running Google Ads for any length of time, you will have noticed that Performance Max is no longer optional. Since Google replaced Smart Shopping campaigns in 2022, PMax has become the default format for most ecommerce advertisers — and according to Google, PMax campaigns now account for the majority of Google Shopping ad spend globally. Yet it remains the campaign type that is most poorly understood and most poorly managed.

This is the guide I wish I had had when I started working with it. No unnecessary jargon, no empty promises. Just what actually works for ecommerce in 2026. If you want to see how Performance Max fits into a complete Google Shopping strategy, the companion piece covers the full structural picture.

---

Key Takeaways

• Performance Max distributes budget across all seven Google channels simultaneously. In well-optimised ecommerce accounts, 60-80% typically goes to Shopping, the highest-converting channel (Google Ads Help, 2024).
• Segment asset groups by product category or margin from day one. Mixing all products in one group prevents the algorithm from associating coherent messages with specific products.
• The algorithm needs 30-50 conversions per month before activating Target ROAS. Setting an aggressive tROAS too early restricts bids and prevents the system from learning properly (Google Smart Bidding documentation, 2024).
• Creative quality matters more than you think. If you don’t upload a video, Google auto-generates one from your images — and those auto-generated videos reduce performance on YouTube and Discover.

Contents

1. What is Performance Max and why is it different?
2. How PMax actually works
3. The structure of a Performance Max campaign
4. Asset Groups: how to organise them for ecommerce
5. Audience signals: your biggest lever of control
6. Search Themes: the control most advertisers ignore
7. PMax + Standard Shopping: the hybrid strategy
8. The mistakes that cost the most money
9. How to tell if your PMax is working
10. When NOT to use Performance Max

---

What is Performance Max and why is it different?

Performance Max is the only Google Ads campaign type that runs simultaneously across all seven Google channels. According to Google Ads Help, PMax uses Google AI to optimise bids, placements, and creative combinations in real time — a fundamentally different model than any channel-specific campaign type.

Performance Max (PMax) is a Google Ads campaign type that uses artificial intelligence to distribute your budget automatically across all seven Google channels simultaneously: Search, Shopping, YouTube, Display, Discover, Gmail, and Maps.

Unlike a traditional Search or Shopping campaign, you do not control which channel your ad appears on, which audience sees it, or when. You give Google your creatives, your product feed, and your objectives — and the algorithm makes the rest of the decisions.

This has two implications:

• The good one: The system can find conversions in channels you would never have tested manually.
• The complicated one: If you do not configure it properly from the start, the algorithm optimises towards what is easiest to convert, not necessarily what is most profitable for your business.

You can consult the official Performance Max documentation from Google for the full technical specification.

---

How Performance Max actually works

PMax is more transparent than most advertisers realise, once you understand its three-step decision loop. According to Google’s Smart Bidding documentation, the system needs a minimum of 30-50 conversions per month to exit the learning phase — though real tROAS reliability only kicks in at 150+/month (see Smart Bidding: when it works and when it fails) — a threshold that shapes everything from campaign structure to budget minimums.

What PMax does under the hood is a three-step process that repeats continuously:

1. Decides where to show your ad — which channel, which type of user, which time of day.
2. Decides what to show — which combination of images, videos, headlines, and descriptions from your asset groups to assemble for that specific user.
3. Decides how much to bid — based on the estimated probability of conversion and your ROAS or CPA target.

All of this happens in milliseconds, millions of times per day. The algorithm learns from the conversion data it accumulates, so the first 30-45 days are critical: this is when the system is in its learning phase and needs sufficient traffic to stabilise.

One thing I’ve observed across accounts: PMax doesn’t behave like a single campaign. It behaves more like a portfolio of channel strategies held together by a shared budget and a shared conversion objective. The algorithm will sacrifice Shopping efficiency to test YouTube if your feed has gaps. Understanding that dynamic changes how you diagnose underperformance.

The real budget split

In well-optimised ecommerce PMax campaigns, the typical distribution is as follows. These figures vary by sector, catalogue, and feed quality — they are directional benchmarks based on real account management, not figures published by Google:

• 60-80% in Shopping — the highest-converting channel for ecommerce
• 10-20% in Search
• 5-15% in Display, YouTube, and Gmail

If you see Display or YouTube consuming more budget than expected, it is usually a signal that the product feed has issues or that your creatives are not competitive enough.

---

The structure of a Performance Max campaign

A well-structured PMax campaign requires deliberate decisions at both levels. Google Ads Help notes that correct campaign structure directly affects how quickly the algorithm accumulates the conversion volume needed to optimise efficiently.

Campaign level

At campaign level you define your conversion objective, bidding strategy (Maximise conversions or Maximise conversion value, with or without a tROAS target), budget, locations, and languages.

Asset group level

An asset group is the basic unit of PMax: it contains your creatives, your products, and your audience signals. Within each asset group you include:

• Up to 15 headlines (max. 30 characters each)
• Up to 5 descriptions (max. 90 characters)
• Up to 20 images in multiple formats: landscape, square, portrait
• Up to 5 videos (minimum 10 seconds)
• The final URL
• Audience signals
• Product groups from the feed

If you do not upload a video, Google auto-generates one from your images. The quality of those auto-generated videos tends to be poor — see mistake number 5 below.

---

Asset Groups: how to organise them for ecommerce

The most common mistake I see in ecommerce accounts is having a single asset group with all products and all creatives mixed together. The result is that the algorithm cannot associate coherent messages with specific products. According to Google Ads guidance, the system needs sufficient volume per asset group to optimise correctly — fragmentation slows learning, but so does over-mixing unrelated products.

Best practice is to segment asset groups by product category or by margin.

Example structure for a fashion store

PMax Campaign — Women’s Fashion

• Asset Group: Dresses (products + dress creatives, headlines focused on dresses)
• Asset Group: Shoes (products + footwear creatives, headlines focused on footwear)
• Asset Group: Accessories (products + bags and jewellery creatives)

When the ad, the landing page, and the products in the asset group are aligned, relevance improves — and performance follows.

How many asset groups do you need?

Monthly spend	Recommended structure
< €1,000	1 campaign, 1-2 asset groups per main category
€1,000-€3,000	1 campaign, asset groups by category + margin
€3,000-€10,000	Separate campaigns by category or margin
> €10,000	Multiple campaigns with dedicated asset groups

Watch out for over-segmentation: fragmenting too much distributes conversion volume across many campaigns and slows the algorithm’s learning.

---

Audience signals: your biggest lever of control

Audience signals are the primary way you guide the algorithm’s learning direction. According to Google’s Customer Match documentation, uploading a first-party customer list as an audience signal accelerates PMax’s learning phase more than any other single configuration step.

Audience signals are how you tell Google “start looking here.” They are not fixed targeting — the algorithm takes them as a starting point and then expands towards similar users.

The most powerful signals for ecommerce, from highest to lowest impact:

Customer Match — Upload your database of real buyers. This is the most valuable signal because you are telling the algorithm exactly what kind of person buys from you. In 2026 you can also use it to exclude existing customers from acquisition campaigns. See the Customer Match requirements in Google Ads for minimum list thresholds.

Website visitors and buyers — Remarketing segments from your own traffic: users who visited product pages, abandoned the basket, or have already purchased.

In-Market segments — Users Google has identified as actively interested in your category. Useful when you do not have enough first-party data.

Custom segments — Audiences based on search keywords, visited URLs, or apps used. Very useful for capturing users searching for competitor brands.

Do not limit yourself to a single signal type per asset group. Combine Customer Match, website visitors, and In-Market in each asset group. The more information you give the system, the faster it learns.

---

Search Themes: the control most advertisers ignore

Search Themes are one of the most underused configuration options in PMax. According to the Google Ads 2025 update notes, the limit was raised to 50 Search Themes per asset group — double the previous cap — specifically to give advertisers more control over which search intent the algorithm prioritises during learning.

In 2026, PMax allows up to 50 Search Themes per asset group. These are suggested keywords that indicate to the algorithm which search terms are relevant to your products.

They do not work like keywords in a Search campaign — they do not exclude traffic or guarantee impressions on those searches. But they do help the algorithm orient itself during the learning phase. Use them for business-specific terms the system might take time to discover on its own: product names, niche categories, technical terms.

I’ve found Search Themes most useful for niche products with low search volume. Without them, the algorithm tends to default to broader, higher-volume terms in your category — often the same generic queries your competitors dominate. Adding 10-15 very specific Search Themes pulls the system towards the long-tail intent where CPCs are lower and conversion rates are higher.

---

PMax + Standard Shopping: the hybrid strategy

Since late 2024, PMax and Standard Shopping campaigns compete on equal footing in the auction. Google confirmed this change in their official ads blog, ending the era of automatic PMax priority. This opens the door to a hybrid strategy that many advertisers are using profitably in 2026.

The idea is to use PMax for scale and Standard Shopping for surgical control in specific situations:

Situation	Tactic
Products with low volume in PMax	Standard Shopping with Maximise Clicks to force data collection
Brand searches	Standard Shopping with manual CPC to control cost
Low-margin products	Standard Shopping with a bid cap
Stock clearance	Standard Shopping with aggressive bids and defined dates

For this strategy to work without the campaigns competing against each other, apply negative keywords at account level and clearly segment products to avoid overlap. The full Google Shopping campaign structure guide covers product group segmentation, custom labels, and bidding strategy decisions in depth.

---

The mistakes that cost the most money

After auditing dozens of ecommerce accounts, these are the mistakes that appear most frequently. Most of them are structural — set wrong at launch and compounding every day after.

1. Treating PMax as “set it and forget it” — The algorithm learns, but it does not make strategic decisions. You need to review performance by asset group, search terms in the insights report, and asset quality at least every two weeks.

2. Not optimising the feed before launching — The Merchant Center feed is the foundation of everything. A feed with weak titles, generic descriptions, or low-quality images undermines performance regardless of how well the campaign is configured. You can validate your feed in Google Merchant Center.

3. Mixing products with very different margins in the same asset group — The algorithm will optimise towards what converts most easily, not what is most profitable. Separate by margin from day one.

4. Setting an overly aggressive tROAS from the start — A very high tROAS target during the learning phase restricts bids so severely that the system cannot accumulate enough data to learn. The industry standard is to wait until you have between 30 and 50 conversions per month before activating tROAS — a figure Google also recommends in its Smart Bidding documentation. Start conservative and increase gradually.

5. Not uploading video — If you do not include video, Google auto-generates one from your images. These auto-generated videos tend to be low quality and reduce performance on YouTube and Discover. Upload at least one product video of 15-30 seconds.

6. Not using audience exclusions — In 2026 you can now exclude Customer Match lists. If you are not excluding existing customers from acquisition campaigns, you are paying to reconvert people who have already bought from you.

---

How to tell if your PMax is working

These are the indicators I review in every audit. No single metric tells the full story — you need to read them together.

• % of impressions in Shopping vs other channels — Should be 60-80% in Shopping for ecommerce. Per Google Ads Help, Shopping consistently shows the highest conversion rates for product-based campaigns.
• Asset quality — In the asset group report, any asset rated “Low” should be replaced.
• Conversion volume — You need a minimum of 30-50 conversions per month for the algorithm to optimise correctly.
• Search terms report (Insights) — Review which searches are generating conversions and add negatives where needed.
• ROAS by asset group — Identify which product categories are performing and which are not hitting the target. For a deeper look at the levers available once campaigns are running, the guide on how to improve ROAS on Meta and Google paid campaigns covers both channels.

---

When NOT to use Performance Max

PMax is not the solution for everything. There are situations where other campaign types are more appropriate:

• Very low budgets (under €500/month) — There is not enough volume for the algorithm to learn properly.
• Very small catalogues (fewer than 20 products) — Consider Standard Shopping for more control over which products appear and when.
• Long conversion cycle products — If your sales cycle runs to weeks, the PMax algorithm struggles to optimise correctly.
• When you need full transparency — PMax remains a partial black box. If the client requires detailed reporting by search term or placement, combine it with Search campaigns.

---

Frequently Asked Questions

How long does Performance Max take to exit the learning phase?

PMax typically takes 30-45 days to fully exit the learning phase. According to Google’s Smart Bidding documentation, the system needs a minimum of 30-50 conversions within the window to optimise bidding reliably. During this period, avoid major campaign changes — each significant edit can restart the learning clock.

Can I run Performance Max alongside existing Shopping campaigns?

Yes, and since late 2024 both campaign types compete on equal Ad Rank footing with no automatic PMax priority, per Google’s official confirmation. A hybrid strategy using PMax for broad reach and Standard Shopping for specific product control is a valid setup in 2026. Use account-level negative keywords to prevent overlap between the two.

How many asset groups should I create per Performance Max campaign?

According to Google Ads guidance, the right number depends on conversion volume. As a general rule, each asset group needs enough conversions to learn — fragmentation across too many groups slows the algorithm. For accounts spending under €3,000/month, 2-4 asset groups per campaign is typically the right balance between thematic organisation and volume per group.

Does feed quality really affect PMax performance that much?

Feed quality is the single biggest lever for Shopping-heavy PMax campaigns. Google Merchant Center scores your feed on title quality, image quality, and completeness. A feed with generic titles like “Men’s T-Shirt Blue” performs significantly worse than one with specific, keyword-rich titles. In accounts I’ve audited, feed optimisation alone has produced 20-35% improvement in Shopping ROAS before touching any campaign settings.

What’s the difference between audience signals and targeting in PMax?

Audience signals in PMax are suggestions, not restrictions. You tell the algorithm where to start, but it can and will expand beyond your signals to find new converters. This is fundamentally different from standard audience targeting, where you limit reach to specific segments. Per Google Ads Help, providing strong signals (especially Customer Match lists) dramatically shortens the learning phase and improves early campaign efficiency.

---

Making Performance Max Work for Your Ecommerce Store

Performance Max in 2026 is a powerful tool when configured with judgement. The difference between an account that wastes budget and one that scales profitably is not in the automation — it is in feed quality, asset group structure, the audience signals you provide, and ongoing supervision.

It is not magic. It is strategy applied to an AI system.

If you want to know how your Performance Max campaign is performing, I can do a free audit of your account. I review the structure, the feed, the asset groups, and the budget split, and give you a concrete report with the points to improve. Request your free audit here.

---

Related articles:

• How to structure Google Shopping campaigns (2026)
• Google Ads Quality Score in 2026: what actually moves the needle

---

Sources

1. Performance Max campaign overview — Google Ads Help
2. Smart Bidding minimum conversion thresholds — Google Ads Help
3. Asset group best practices — Google Ads Help
4. Customer Match requirements — Google Ads Help
5. Search Themes in Performance Max — Google Ads Help
6. Google Merchant Center feed quality guidelines
7. Performance Max updates 2024 — Google Blog