Performance Max is no longer an experiment. In 2026, it accounts for the majority of Google Ads revenue, and for most scaling brands running Google Ads, it will be the dominant campaign type in their account whether they planned it that way or not. DSA campaigns are being migrated to AI Max. Smart Shopping has been absorbed. The platform is moving towards full automation, and the question is no longer whether to use Performance Max it is whether you know how to run it well.
Most brands that struggle with PMax make the same mistake: they assume that automation means the work is done for them. They launch a campaign, upload some assets, set a target ROAS and wait. The results are mediocre. They conclude that Performance Max does not work for their business.
The brands achieving 200 to 400 per cent better ROAS from PMax are doing something structurally different. They understand that automation does not replace strategy. It executes it and only as well as the inputs allow.
What Performance Max Is Actually Doing
Performance Max is a unified buying system that places ads across Search, YouTube, Display, Gmail, Discover and Maps, optimising dynamically across all placements based on a single conversion goal. The algorithm has access to Google’s intent signals, cross-channel behavioural data and real-time auction dynamics that no manual campaign can replicate.
But the algorithm does not know your brand. It does not know which audiences are high-LTV versus high-churn. It does not know that your best customers look nothing like your average customer. It does not know that there are brand safety contexts you want to avoid, or creative directions that damage your positioning even if they generate clicks.
All of that needs to come from you. And most brands fail to provide it.
Audience Signals: The Input That Changes Everything
Audience signals are not targeting in the traditional sense they are suggestions that orient the algorithm’s learning. Performance Max uses them as a starting point for finding converters, not as a hard constraint. But the quality of those signals determines how quickly the campaign exits the learning phase and how accurately it targets from that point forward.
Effective audience signals in 2026 are built from first-party data: your customer list, your purchaser list, high-intent site visitors, email subscribers who have demonstrated product interest. Feed the algorithm a customer match list of your best customers defined by LTV, not just purchase frequency and you give it a meaningful signal to optimise towards. Feed it a broad interest category, and you are providing almost no useful information.
The brands that see fastest results from PMax are typically the ones with the richest first-party data to contribute as signals. If you do not have clean CRM data, building it is a prerequisite not an optional enhancement.
Asset Groups: Where Strategic Control Lives
Asset groups are how you maintain brand and strategic control inside an automated campaign. Rather than running one asset group with a broad mix of creative, structure your asset groups around distinct audience intents or product lines.
A scaling e-commerce brand might run separate asset groups for: new customer acquisition, competitor conquest, high-LTV product lines and seasonal promotions. Each asset group has a distinct creative direction, a relevant URL target and audience signals specific to that intent cluster. This prevents the algorithm from blending incompatible audiences and diluting performance across contexts.
Each asset group should contain a diverse creative library at minimum, multiple headlines, descriptions, images in both horizontal and vertical formats (Google added native 9:16 support in late 2025) and video assets. The asset-level performance reports inside PMax now show which specific assets are contributing to conversions, which should drive your creative refresh decisions.
Conversion Data Quality Is Non-Negotiable
Performance Max’s algorithm is only as intelligent as the conversion data it receives. If your conversion tracking is incomplete if you are missing server-side events, if your attribution window is misaligned with your sales cycle, if you are optimising towards micro-conversions that do not correlate with actual revenue the algorithm will optimise towards the wrong outcomes.
This is where many brands undermine their own PMax performance without realising it. They have weak tracking, launch Performance Max, and interpret the mediocre results as a campaign problem. The campaign is usually fine. The tracking is the problem.
Google’s Enhanced Conversions passing hashed customer data to improve conversion matching should be implemented for any brand running PMax at scale. Combined with a properly configured GA4 and server-side event stream, this is what gives the algorithm the signal quality it needs to find your real customers reliably.
What You Can and Cannot Control
Performance Max gives you less keyword-level control than traditional Search campaigns. You cannot see exactly where your ads are appearing. You cannot exclude specific placements as granularly as in Display campaigns. This makes some advertisers uncomfortable.
The legitimate controls you do have are: brand safety exclusions via account-level placement settings, negative keyword lists (now available at campaign level in 2026), brand exclusion to prevent PMax from competing with your own brand search campaigns, and URL exclusions to prevent traffic from landing on pages you do not want to drive paid traffic to.
Use all of them. The instinct to let PMax run without guardrails is understandable it is an automated system, surely it knows what it is doing but the algorithm optimises for the conversion goal you set, not for your brand positioning or business model constraints. Those guardrails are your responsibility.
The Integration Principle
Performance Max performs best when it is part of a coherent account structure, not when it is running in isolation. A mature Google Ads account in 2026 typically combines PMax for broad, automated scale with targeted Search campaigns for high-intent branded and non-branded terms, YouTube campaigns for sequenced brand storytelling, and clear budget logic that prevents PMax from cannibalising lower-funnel campaigns.
AI runs the channels. You design the system that channels operate within. That is not a diminished role for the advertiser it is a more strategic one. The brands winning with Performance Max are not the ones that have surrendered control. They are the ones that have learned which inputs matter and built the discipline to provide them consistently.
Plethora Digital manages Performance Max and full Google Ads accounts for scaling brands across the UK and EU. Book a strategy call to understand how your current account structure compares to what is actually working.