Google will let Search advertisers run A/B tests across multiple campaigns at once starting in September, comparing budgets and ROI targets in a single experiment instead of one campaign at a time. The change extends AI Max for Search, the automation layer Google has been pushing into campaign management. Paid Media Editor Anu Adegbola reported the update for Search Engine Land.
The current one-click AI Max experiment tests a single campaign. Multi-campaign A/B testing lets an advertiser scale a budget or ROI target change across an entire account and see the aggregate effect before applying it broadly. That distinction matters for accounts managing dozens of campaigns, where a change validated on one campaign does not reliably predict what happens when the same change runs everywhere.
Google is also loosening a constraint that had discouraged some advertisers from testing AI Max at all. AI Max experiments will now support brand and location controls, meaning advertisers can measure the feature’s impact without disabling the guardrails that keep spend inside approved brand terms or geographic markets. Previously, running an AI Max experiment meant stripping out those controls for the test window, an unacceptable tradeoff for businesses with contractual or regulatory limits on where their ads can appear.
Performance Planner is getting a parallel expansion. Advertisers will be able to forecast how a bidding or budget change affects existing campaigns, then apply Google’s suggested changes with one click. That shortens the distance between running a forecast and acting on it, removing a manual step that previously separated planning from execution.
None of this changes what AI Max does to a campaign. It changes how much evidence an advertiser has before turning it on. Google has spent the past year folding more automated decision-making into Search campaign structure, and the pattern with automation features has been consistent: advertisers who validate a change against their own account data before scaling it get materially different results than advertisers who accept the default rollout. The new testing tools formalize that validation step rather than leaving it to individual account managers.
The brand and location control support is the more consequential change of the two for regulated or franchise advertisers. A national brand with regional franchisees, or a company operating under trademark restrictions in certain markets, previously could not test AI Max without exposing itself to the exact risks those controls exist to prevent. Removing that barrier means a wider set of advertisers can now generate their own performance data on AI Max instead of relying on Google’s account recommendations or peer anecdotes.
Google’s announcement does not include independent measurement of how much incremental performance these tools actually surface, only a description of the testing mechanics themselves. Advertisers will need their own experiments to know whether multi-campaign testing changes their conclusions relative to single-campaign tests.
Search teams with multiple active campaigns should hold off on account-wide AI Max budget changes until the multi-campaign testing tool ships in September, then run a controlled comparison before adopting Google’s suggested settings at scale.
Search Engine Land reported this update, written by Paid Media Editor Anu Adegbola, on August 20, 2026.
