Mastering Performance Max: Strategy, Evolution & Best Practices

Next up:

The Evolution of Performance Max

2021

Saw the Beta launch for selected advertisers.

It focused on consolidating campaign management and testing AI-driven ad placement.

2022

Saw the global rollout & Smart Shopping migration.

PMax became Google’s default AI campaign type for eCommerce, replacing Smart Shopping and Local campaigns.

2023

Introduction of asset group reporting and brand exclusions.

This gave advertisers greater transparency and control over where and how PMax ads appeared.

2025 (to date)

Emphasis on lead quality, data integration and AI-driven creative tools

Reinforcing PMax’s dual role as both a conversion engine and a creative testing platform for all verticals.

Now the first of the 2 really important bits:

The Role of Data: Feeding the Algorithm

PMax’s power lies in machine learning, but its intelligence depends entirely on the quality of data it receives.

A. Conversion Tracking Foundations

Accurate, meaningful conversion tracking is a must have. PMax optimises around the goal you define, so define it properly.

• Use bottom-funnel events (purchases, qualified leads, bookings).
• Avoid vanity metrics (clicks, page visits) unless running an awareness campaign.
• Where possible, assign conversion values to help the system prioritise high-value actions.

B. Offline Conversion Tracking & CRM Integration

For especially B2B and lead-gen advertisers, PMax can now ingest offline data from CRMs (e.g. Salesforce, HubSpot). This bridges the gap between ad clicks and real sales.

How it works:

1. Capture the Google Click ID (GCLID) when a lead submits a form.
2. When the lead later qualifies or converts, upload that record back to Google Ads (manually or via API).
3. Google’s algorithm learns which leads became sales and optimises toward those patterns.

This closes the loop, turning PMax from a lead quantity tool into a lead quality engine.

C. Audience Signals & First-Party Data

PMax allows Audience Signals, which act as directional hints for the algorithm:

• Upload customer lists, remarketing audiences or high-value user segments.
• Use custom intent and search themes to indicate relevant topics or keywords.
• Pair this with Customer Match to help the AI find similar high-value prospects.

Now the second really important bit for PMax:

Creative Excellence: The New Competitive Edge

Creatives are not just assets, they’re the raw material for AI experimentation. In PMax, Google’s system mixes and matches your assets across placements to learn what performs best. This generates a lot of data and it is GOLD. At Altair we have our digital media team (our AI agents) pull these huge datasets apart to get the real insights to drive our clients forward.

A. Asset Groups

Organise creative assets into logical groups aligned with audience or product themes. For example:

• “Luxury Travel” vs “City Breaks”
• “Corporate Solutions” vs “SME Support”

Each asset group can contain unique messaging, images, and calls to action to improve relevance.

B. Best Practices for Creative Inputs

• Images: Use high-resolution visuals with minimal text. Include product shots and lifestyle imagery showing real people in context .
• Videos: Even simple, branded 15-second videos outperform static assets. Avoid relying solely on Google’s auto-generated videos, they tend to underperform by 25–40%.
• Headlines & Descriptions: Provide at least 5–10 of each to maximise testing combinations. Use different tones, benefit-led, action-driven, and emotional.
• Logos & CTAs: Ensure branding consistency across all assets. PMax will automatically adapt these across formats, but start with a strong design base.
• “Ad Strength” Metric: Aim for Excellent. Google’s analysis shows campaigns rated “Excellent” can drive up to 6% more conversions on average.

C. Generative AI in Creative

While Meta leans heavily on AI-generated visuals, PMax currently uses AI to reassemble and adapt your inputs, not fully generate new ads. However, this is likely to change, so stay on guard to stay on brand.

For now, treat AI as a co-creator, not the creator.

Lead Quality: From Volume to Value

A recurring theme is: optimise for quality, not just quantity.”

In lead-gen campaigns:

• Focus on deeper funnel goals (e.g. “booked demo” rather than “form fill”).
• Feed back qualified lead data from sales teams or CRMs.
• Use value-based bidding (tROAS or tCPA) with different conversion weights per lead type.
• Regularly clean your data: remove duplicate or spam leads to prevent AI bias.

Measurement, Reporting & Transparency

Early criticism of PMax centred around its opacity, but to be fair Google has steadily improved visibility:

• Search Term Insights: Shows which themes drive conversions.
• Asset Group Reporting: Breaks down performance by creative cluster.
• Conversion Paths: Reveals cross-channel interactions within PMax.
• Brand Exclusions: Lets you exclude branded keywords to focus on incremental reach.

We can now combine these with our digital media team's (AI Agents) pipelines for more granular reporting, a must for ongoing performance improvements.

To sum up, Google has been sunsetting massively important channels and absorbing them into PMax, this is not going away, it's only going to become more of a focus, so get on board and start testing.

Remember quality data and quality creative is absolutely central to a successful campaign.