Google announces AI Max experiments across Search campaigns

What changed

Brandon Ervin’s Google announcement introduces a single A/B test for budget or return-target changes across multiple Search campaigns, with rollout planned for September. Dan Taylor also shared the announcement. The original article separately describes AI Max experiments that retain brand and location controls, plus the ability to apply Performance Planner suggestions directly to campaigns.

These functions answer different questions. An experiment can compare outcomes under a defined change; a planning tool estimates what may happen. The convenience of applying a forecast does not turn that forecast into a measured result. Before making a portfolio adjustment, decide which business outcome the test should establish and which campaign settings must stay stable for the comparison to be useful.

As an illustrative case, an advertiser may want to raise investment across several non-brand campaigns while retaining an acceptable acquisition cost. The combined result is relevant to the budget decision, but it can conceal a shift toward one campaign or customer segment. Inspect the distribution of spend and outcomes as well as the total. An aggregate improvement does not mean that every participating campaign improved.

Why it matters for advertisers

Cross-campaign testing can bring the experiment closer to the decision a budget owner actually faces. Its value depends on defining the decision before reading the result. Establish the campaign set, intended change, success measure and conversion-lag allowance, then document other changes that could complicate interpretation. Preserve brand and location controls that are necessary for the business instead of relaxing them just to make a test easier to run.