Performance Max asset groups: structuring for better signals

Performance Max (PMax) is Google's most powerful campaign type, running simultaneously across Search, Shopping, Display, YouTube, Gmail and Maps. The performance of any PMax campaign depends directly on the quality of the signals you feed into it. Asset groups are the primary vehicle for those signals. A poorly structured campaign with one generic asset group gives Google a confusing mixture of information. A well-structured campaign with thematically distinct asset groups gives the algorithm clear, actionable direction. This article explains what asset groups are, how to structure them for maximum signal strength, and how AdBrains AI technology automates this process for consistently better results.

What are Performance Max asset groups?

One asset group vs. structured asset groups

Impact on signal strength and campaign performance

One generic asset group

• All products/services mixed together
• Weak audience signals due to broad mix
• Ad Strength hard to optimise
• Google cannot learn thematically
• Low relevance per audience segment
• Conversion data fragmented across themes

Structured asset groupsRecommended

• Each group = one theme or audience
• Strong, targeted audience signals per group
• Ad Strength optimisable per group
• Google learns faster per segment
• High relevance: better CTR and ROAS
• Conversion signals bundled per theme

An asset group is the creative container within a Performance Max campaign. It holds all your creative elements: headlines, descriptions, images, logos, videos and calls-to-action. Each asset group also carries its own set of audience signals, the hints you give Google about who you want to reach. These can include custom intent audiences based on search terms, customer match lists, website visitor audiences and in-market segments.

Technically, you can run a PMax campaign with a single asset group. In practice, this only works well when your product or service offering is completely homogeneous. The moment you serve multiple product categories, audience types or intent levels, a single asset group becomes a bottleneck. Google cannot distinguish between themes and will optimise towards the combination that generates the most volume, which is not always the combination that delivers the best margin or the right audience profile.

Structuring asset groups by theme or audience segment gives the PMax algorithm the clearest possible picture of your business. The result is higher relevance, stronger Ad Strength scores and a Smart Bidding system that converges faster towards a profitable strategy.

Why structure is the foundation of strong signals

The quality of a Performance Max campaign is a direct function of the signals it receives. Audience signals are the most explicit way to guide Google, but the assets themselves, the texts, images and videos, are signals too. Google analyses asset content to understand the contexts in which they are relevant. If an asset group mixes headlines about online driving theory exams with images related to food safety certification, the internal signals contradict each other and the system learns more slowly.

For ToetsJeKennis.nl, an online platform for exams and courses, this principle is immediately applicable. A well-structured PMax campaign might include separate asset groups for driving theory preparation, food safety certifications, and corporate training programmes. Each group sends a coherent, unambiguous signal to the algorithm, leading to higher Ad Strength scores, more relevant impressions and ultimately a better ROAS.

The same principle applies equally in lead generation. At Clima-Active.nl, an air conditioning and heat pump installation company focused on quote requests, an asset group for "residential air conditioning installation" sends fundamentally different signals than one for "heat pump quote new build project". Separating them allows Google to learn two distinct customer journeys simultaneously and to route users to the most relevant creative at every touchpoint.

The five pillars of a strong asset group structure

Average performance improvement with correct asset group structure

Improvement vs. campaigns with one generic asset group

ROAS improvement

34%

CTR improvement

28%

Conversion volume increase

31%

Ad Strength score (Excellent)

62%

CPA decrease

22%

Effective asset group structuring rests on five core principles. Advertisers who apply these consistently give the PMax algorithm the best possible foundation for rapid learning and strong performance.

1. Thematic homogeneity: Each asset group represents one theme, product category or service. Never mix unrelated topics in a single group. The tighter the theme, the stronger the signal.
2. Targeted audience signals: Attach the most relevant audience signals to each group. Combine a custom intent audience built from specific search terms with an in-market segment and, where available, a customer match list or website visitor audience.
3. Consistent creative: All assets within a group should tell the same story. Headlines, descriptions, images and video must reinforce one another. This improves both Ad Strength and post-click user experience.
4. Dedicated landing pages: Connect each asset group to a specific, relevant landing page. A generic homepage for all asset groups reduces relevance and weakens the algorithm's ability to optimise effectively.
5. Regular evaluation and iteration: Review performance per asset group consistently. Remove POOR-rated assets, add new variants and test different audience signal combinations to keep the campaign improving over time.

At E-4motion.com, the webshop for new electric folding bikes, this structure is highly effective. An asset group for "electric folding bike for commuters" attracts a completely different audience than one for "electric folding bike as a gift". Different images, different messaging, different audience signals. By separating them, the PMax algorithm learns twice as fast and matches the right creative to the right user at the right moment.

Audience signals: the key to faster learning

Audience signals are frequently underestimated in Performance Max. They are not hard targeting restrictions: Google can and will serve ads outside these audiences if it identifies conversion opportunities. However, they provide an essential starting point for the learning process. The stronger and more relevant the audience signals, the faster the system converges on an effective strategy.

The most valuable audience signals for PMax include:

• Custom intent audiences: Built from specific search terms your potential customers use. These are the most powerful signals you can provide.
• Customer Match lists: Upload your own customer data for a strong first-party signal, especially valuable as third-party cookies disappear.
• Website visitor audiences: Segment by visitors to specific pages such as product pages or thank-you pages for a strong remarketing signal.
• In-market audiences: Google's own purchase intent segments. Most effective when combined with other signal types.
• Similar audiences (based on customer lists): Google's look-alike logic based on your own data often delivers strong new traffic at scale.

The combination of custom intent and customer match lists typically produces the strongest signals. For Clima-Active.nl, a customer list of previous quote requesters is an excellent signal for the heat pump asset group. Google quickly understands who the ideal enquirer is and searches for similar profiles across its entire network.

How AdBrains AI automates Performance Max asset groups

Manually tracking which asset groups are performing, which audience signals are providing the strongest direction, and which assets are dragging down Ad Strength is a time-intensive process at scale. This is precisely where AdBrains' own AI technology creates a structural advantage. Our systems continuously monitor the performance of every asset group and take automatic action based on data, not guesswork or limited bandwidth.

The AdBrains RSA improvement system analyses Ad Strength scores across all asset groups daily. Assets with a POOR rating are automatically identified and rewritten. Our AI uses content signals from the best-performing assets as input, ensuring new variants remain thematically consistent and push Ad Strength towards "Excellent". What would take a human manager weeks of attention, our AI executes every day without manual intervention.

The audience management automation module manages audience signals per asset group on a weekly basis. PROD audiences, incubator audiences and RLSA audiences are automatically created, updated and assigned to the correct asset groups. Signals are always current: an outdated customer list or a website visitor audience that no longer reflects current site segmentation is automatically refreshed.

AdBrains' server-side signal enrichment via our own sGTM infrastructure provides richer conversion signals to the algorithm. Every conversion is enriched with first-party data before being passed to Google, meaning the Smart Bidding system inside PMax receives far more detailed information about who is converting. This makes the audience signals in each asset group even more effective as a starting point for learning.

The multi-agent verification system ensures that no automatic change goes live without four independent AI agents approving the decision. Whether an asset group is being split, an audience signal is being replaced, or a new asset is being activated, every change passes four independent checks before execution. Structural errors in a PMax campaign can cost weeks or months of learning data, and this system is specifically designed to prevent that from happening.

Asset group structure versus product feed structure

A common source of confusion for e-commerce advertisers is the distinction between asset groups and product groups in the feed configuration within PMax. For Shopping ads, the product group setting determines which products from your Merchant Center feed are eligible. The asset group determines which creative assets and audience signals are paired with those products. They are complementary, not interchangeable.

For a webshop like Elletens.nl, which carries multiple product categories, the ideal PMax setup is segmented at both the asset group level and the product group level. The asset groups provide the right creative context and audience signals per category, while the product groups within those asset groups surface the most relevant products. Together, they give the PMax algorithm a maximally coherent and powerful signal.

Common mistakes in asset group structure

Signal strength through the Performance Max funnel

Funnel flow per stage with well-structured asset groups (percent mode)

Impressions reached

Broad Google inventory: Search, Display, YouTube, Gmail, Maps

Relevant clicks

Higher CTR through thematically aligned assets

Landing page engagement

Better relevance through targeted asset group signals

Conversions

Stronger Smart Bidding guidance through bundled signals

Average 2% conversion rate across all Google channels

Even experienced advertisers make the same structural mistakes with Performance Max asset groups. The following are the most common pitfalls and how to avoid them.

• Too many or too few asset groups: One asset group sacrifices thematic relevance. Dozens of asset groups fragment the learning process, leaving each group with insufficient conversion data. Start with three to six well-defined asset groups and expand only once each group generates enough conversions for stable learning.
• Generic audience signals: Broad in-market audiences without custom intent or customer data give Google very little direction. Invest in building quality customer lists and custom intent audiences based on real search data.
• Inconsistent assets: Combining headlines about product benefits with images of an empty office reduces internal coherence. Every asset must reinforce the theme of the group it belongs to.
• Same landing page for all asset groups: Google evaluates landing page relevance as part of its optimisation decisions. Using a generic homepage for every asset group is one of the most impactful structural errors an advertiser can make.
• No iteration after launch: PMax asset groups are not a set-and-forget structure. Regular evaluation of Ad Strength, asset performance ratings and audience signal effectiveness is essential for sustained success.

Frequently asked questions about Performance Max asset groups

How many asset groups should a Performance Max campaign have?

There is no single correct number, but a practical starting point is three to six asset groups, each built around a clearly defined theme or audience segment. Ensure each asset group receives enough budget and conversions for the Smart Bidding algorithm to learn effectively. An asset group generating fewer than thirty conversions per month will struggle to build statistical reliability. Expand only when existing groups perform stably and you have identified a new theme or segment strong enough to justify its own group.

What are the best audience signals for Performance Max?

The strongest signals are custom intent audiences built from specific, relevant search terms, combined with your own customer lists via Customer Match. Website visitor audiences segmented by high-intent pages such as product pages or thank-you pages are also highly valuable. Google's in-market audiences are useful as a supplementary layer, but provide too little specificity when used in isolation. The real power comes from combining multiple signal types to give Google a rich, multidimensional picture of your ideal customer.

How does Ad Strength affect the performance of my asset group?

Ad Strength indicates how complete and varied the assets in a group are. An "Excellent" Ad Strength means Google has sufficient creative variation to assemble the optimal combination for each individual auction. This directly leads to higher relevance and better CTR. Assets with a "POOR" rating are served less frequently and drag down overall group performance. Regular evaluation and replacement of POOR-rated assets is one of the highest-impact optimisation actions available in a PMax campaign.

Can I use negative keywords in Performance Max?

Yes, but the options are more limited than in standard Search campaigns. Negative keywords can be applied at the account level or, upon request to Google, at the campaign level. Because PMax operates across all channels simultaneously, careful negative keyword management is essential to prevent budget from being spent on irrelevant search terms. AdBrains runs daily automated search term mining specifically to maintain signal purity for PMax campaigns, ensuring that only relevant traffic reaches each asset group.