Performance Max in 2026 - what we know now after two years of proper data

Performance Max launched to significant controversy. Agency practitioners worried about the loss of control. Google promised better results through AI optimisation. Neither position turned out to be entirely right. PMax has settled into a role that is genuinely valuable for some advertisers and genuinely problematic for others, and the determinants of which camp you fall into are now reasonably well understood.

Where Performance Max consistently delivers

Ecommerce with large product catalogues and high conversion volume is where PMax performs most reliably. The AI has enough signal to optimise effectively, the cross-channel reach captures shopping intent across YouTube and Search simultaneously, and the automated creative testing at scale does things that manual campaign management cannot replicate. Retailers with 50 or more conversions per month and well-optimised product feeds consistently report PMax delivering incremental volume that Standard Shopping alone did not capture.

Remarketing within PMax also works well. The ability to maintain brand presence across Search, Display, YouTube, and Discover for audiences that have already expressed intent - through site visits or Customer Match lists - at relatively low cost makes PMax a practical remarketing layer for advertisers who previously managed separate remarketing campaigns across multiple campaign types.

Where Performance Max still frustrates

Brand safety and placement control remain genuine limitations. Despite improvements to the placement exclusion tools, PMax still serves on placements that many brand-conscious advertisers would prefer to avoid. The transparency on exactly where ads serve, though improved from launch, is still less granular than what standard Display or Search campaigns provide.

Low-volume accounts remain challenged. PMax needs conversion data to learn. Accounts generating fewer than 30 conversions per month do not give the algorithm enough signal to optimise effectively. In these accounts, Standard Search campaigns with manual or Target CPA bidding typically outperform PMax because the simpler system makes better decisions with limited data.

The 2026 update on controls

Google has added negative keyword functionality, brand exclusions, URL exclusion rules, and improved channel-level reporting to PMax following consistent advertiser feedback. These additions have addressed some of the most significant control gaps from launch. The platform is meaningfully more manageable than it was at launch, and the improvements make the transparency case against PMax less compelling than it was two years ago.