Performance Max Campaign: Google's AI-Powered Advertising Format

Performance Max campaigns are taking over more and more budget in Google Ads – and not without reason. Advertisers achieve with PMax on average 18% more conversions at comparable cost-per-action than with classic Smart Shopping campaigns. Those who use the AI-driven campaign format correctly benefit from cross-channel reach and automated optimization. Those who let it run blindly quickly lose control over budget and performance in the SEA area.

What is a Performance Max campaign and how does it differ from other Google Ads formats?

A Performance Max campaign is a goal-based campaign format that bundles all Google channels in a single campaign. Performance Max was introduced in 2021 and completely replaces Smart Shopping as well as Local campaigns. Instead of managing individual campaigns for Search, Shopping, Display, YouTube, Discover and Gmail, the Google Ads AI takes over cross-channel delivery automatically.

The central difference to classic campaign types: Advertisers provide assets (texts, images, videos) and define target values – the AI then decides itself where, when and to whom ads are delivered. This offers significant efficiency potential, but at the same time severely limits direct control options.

How does Google Ads AI work within a PMax campaign?

The AI behind Performance Max is based on Google's machine learning models, which continuously learn from signals such as search queries, user behavior, location data and device types. The AI optimizes automatically across all channels and inventory sources in real time – and combines the provided assets into ad variants that are permanently tested and optimized.

Crucial is: The Google Ads AI learns from conversion data. The more and the better conversion signals are available, the more precisely the system can optimize. Advertisers with fewer than 30-50 conversions per month should use PMax with caution, as the system cannot deliver reliable optimization without sufficient data basis.

What is the difference between PMax and Smart Shopping?

Smart Shopping was the direct predecessor of Performance Max for the e-commerce sector. Smart Shopping campaigns were completely replaced by Performance Max by September 2022. The essential difference: Smart Shopping was limited to the Google Search Network and YouTube, while PMax plays on all available Google channels and processes significantly more asset variants.

Which channels does a PMax campaign play on – and how does Google distribute the budget?

Performance Max automatically plays on all six Google channels: Google Search, Google Shopping, YouTube, Display Network, Discover and Gmail. The budget distribution is completely taken over by Google Ads automation – the advertiser cannot specify percentage shares per channel.

Per auction, more than 300 signals are evaluated, including device, location, time of day, search query and user interests. In practice this means: Google decides in real time on which channel an impression has the highest conversion probability and switches the ad accordingly.

The 6 channels of Performance Max at a glance:

• Google Search: Text ads in search results – based on AI-generated keyword targeting
• Google Shopping: Product ads with image and price – requires a linked Merchant Center feed
• YouTube: Video ads (in-stream, in-feed) – requires video assets; without own videos, Google automatically generates videos from images
• Display Network: Banners and responsive ads on millions of partner websites
• Google Discover: Native ads in the personalized Google feed on mobile devices
• Gmail: Ads in the promotional inbox view – appear as email-like message

What are asset groups – and how do you structure them correctly?

Asset groups are the central control element within a Performance Max campaign. An asset group bundles all creative elements – texts, images, videos – for a specific topic group, product category or target audience. Asset groups should be segmented thematically – for example by product category, seasonality or target audience.

Sensible structuring is one of the few real control levers that advertisers have in PMax. Those who bundle all products in a single asset group give the AI too few signals for precise delivery – and thus waste a significant portion of the optimization potential.

How does Google Ads automation distribute the budget across channels?

Budget distribution in PMax is fully automated – and at the same time one of the biggest criticisms of the format. In practice, PMax campaigns often play a large portion of the budget on YouTube and Display – even if the ROI there is lower than in the Search area. Those who want to prevent this must counteract with brand exclusions and negative keywords at account level.

How do you set up a Performance Max campaign correctly?

The setup of a PMax campaign largely determines whether the system receives meaningful signals from the start or optimizes blindly. The most important rule is: More preparation saves learning phase and budget.

A complete setup follows these six steps:

1. Set up conversion tracking precisely: PMax needs conversion data as a foundation. Without valid tracking, the AI optimizes on wrong or incomplete signals. How conversion tracking is set up correctly in e-commerce is explained in our guide to Google Ads Conversion Tracking.
2. Define campaign goal and bidding strategy: "Maximize Conversions" is suitable for getting started without historical data. "Target CPA" or "Target ROAS" is recommended only from at least 30-50 conversions per month. More on bidding strategies in Google Ads.
3. Segment asset groups thematically: At least one asset group per product category or target audience – the more specific the grouping, the more precise the AI optimization.
4. Provide assets completely and with high quality: Google evaluates asset quality and prefers ad combinations with the highest quality signal. All available asset slots should be completely filled to give the system maximum combination possibilities.
5. Add audience signals: Own customer lists, website visitors, similar audiences – the AI uses these as a starting point but deviates from them in the learning process.
6. Set negative keywords at account level: PMax ignores campaign-related negative keywords. Exclusions must be entered at account level or via a manual request to Google support.

What assets do I need for a Performance Max campaign?

The quality and completeness of assets is one of the most important success factors in a PMax campaign.

PMax automatically directs traffic to the most relevant pages of a website – the quality of landing pages therefore directly influences campaign success. How SEA landing pages are optimally structured and why they differ structurally from SEO landing pages is explained in our in-depth article.

What audience signals should I add?

Audience signals are not classic audience targeting – they merely give the AI a starting point for optimization. The system can and will go beyond the stored signals if it recognizes higher conversion probabilities elsewhere.

The following signals provide particularly valuable impulses for Google Ads AI:

• Own customer lists: Existing customers as a signal for similar new customers
• Website visitors (remarketing lists): Especially cart abandoners and product page visitors
• Similar audiences: Automatically created by Google based on own customer data
• In-market audiences: From Google's own interest category system

How remarketing audiences are strategically built and used is explained in our article on retargeting and how it works.

How do you control and optimize a PMax campaign after launch?

The learning phase of a Performance Max campaign lasts typically 6 weeks. During this time, no significant interventions should be made – budget, bid or asset changes reset the learning phase and significantly extend the period until stable performance.

After the learning phase, the actual Performance Max optimization begins. The rule is: The levers in PMax are limited, but effective – if you know them.

What levers exist for Performance Max optimization?

Performance Max optimization differs fundamentally from the optimization of classic campaigns. Direct keyword control, placement exclusions at campaign level or manual bid adjustments per device do not exist. Nevertheless, effective levers exist:

• Continuously improve asset quality: Google shows in the asset report which combinations perform "best", "good" or "poor". Weak assets should be replaced by stronger ones – ideally with new image motifs and text formulations.
• Further segment asset groups: If an asset group is too heterogeneous, it should be split – for example by seasonality, price segment or purchase intention.
• Refine audience signals: Upload new customer lists, make remarketing segments more precise and keep them current.
• Add negative keywords at account level: Exclude brand keywords of competitors, block irrelevant search queries.
• Use URL exclusions: Exclude certain landing pages from delivery – for example pages without conversion goal such as imprint pages or internal filter results.
• Set seasonal adjustments: Google offers the possibility to temporarily adjust conversion rate expectations – for example during sales campaigns or holidays.
• Regularly check conversion goals: Are the stored conversion actions still current and correctly weighted? Outdated or double-counted micro-conversions significantly distort AI optimization.

What to do if PMax burns budget without conversions?

If a PMax campaign delivers no or hardly any conversions after the learning phase, typically three causes are responsible: insufficient conversion tracking, too generic asset groups or too low budget. The daily budget should correspond to at least 3-5 times the target CPA so that the system can test enough auctions. A daily budget of €10 with a target CPA of €50 does not work structurally.

How do you evaluate a Performance Max campaign sensibly?

Reporting in PMax is one of the biggest criticisms of the format. The limited transparency is not an oversight – it is intended to prevent advertisers from making suboptimal manual interventions based on individual data. Nevertheless, there are ways to conduct meaningful analyses.

Which KPIs are really meaningful in PMax?

Not all available KPIs in PMax are equally reliable. View-through conversions – a conversion that someone performed after seeing an ad without clicking – are counted by Google by default and can greatly overestimate measured performance.

Why is reporting so difficult in PMax – and what helps?

In the search terms report, no individual search terms are shown, but only aggregated categories. This makes precise analysis of the search queries that trigger PMax significantly more difficult than with classic Search campaigns. The following workarounds help with analysis:

• Use UTM parameters consistently: Provide each asset group with its own UTM parameters to track channel-specific performance in Google Analytics 4. How UTM parameters are set up correctly is explained in our guide to UTM parameters and campaign tracking.
• Regularly evaluate asset report: The only direct performance insight at creative level – shows which asset combinations perform strongest.
• Use placement report: Which websites and apps received impressions and clicks? Irrelevant placements can be excluded at account level.
• Evaluate insights tab: Google shows aggregated audience categories and search topics there – helpful starting point for asset optimization and audience refinement.

What mistakes do advertisers make most frequently with PMax campaigns?

Most PMax campaigns do not fail because of the technology – they fail because of wrong expectations and typical setup errors that disadvantage the system from the start.

The six most common mistakes at a glance:

• No valid conversion tracking: PMax without clean tracking is like driving a car without a dashboard. Before PMax is started, conversion tracking should be fully checked and validated.
• All products in a single asset group: The algorithm receives too few signals for precise delivery. At least one asset group per product category is standard.
• Budget too low for meaningful optimization: Those who start PMax with €5 daily budget do not receive reliable results. The minimum is at least 3 times the target CPA as daily budget.
• Interventions during the learning phase: Every significant change in the first 6 weeks resets the learning phase. Patience in the initial phase is not an option, but mandatory.
• Not calculating out view-through conversions: These massively overestimate actual performance – especially with YouTube and Display inventory.
• No negative keywords at account level: Without exclusions, PMax can bid on brand terms of competitors or completely irrelevant search queries – this costs budget without any value.

For whom is a Performance Max campaign worthwhile – and for whom not?

Performance Max is not a universal format for every advertiser. PMax campaigns perform particularly strongly for e-commerce companies with complete product feed and sufficient conversion data – and significantly weaker for service providers with long sales cycles and low conversion volume. Those who want to additionally set up separate Shopping campaigns for e-commerce campaigns and optimize Google Shopping ads will find the most important levers in our article.

Conclusion

Performance Max campaigns are one of the most powerful formats in Google Ads – and at the same time one of the most difficult to control. Google Ads AI takes over decisions that were previously made manually: channel selection, budget distribution, bid adjustment and creative optimization. This saves time and opens up reach that classic campaigns do not achieve.

At the same time: PMax only works as well as the foundations you give it. Clean conversion tracking, thematically structured asset groups, realistic budgets and strategic exclusions are not optional optimization measures – they are the prerequisite for Google Ads automation to work meaningfully at all.

You want to professionally set up, control and evaluate Performance Max campaigns – without putting budget into a black box? We are happy to discuss this in a free initial consultation, non-binding and tailored to your specific use case.