52-account-structure-global

Use one campaign per objective with 3-5 ad sets covering broad or Advantage+, interest themes, lookalikes, and small retargeting, each with 2-3 ads. Run ABO with even budgets so every ad set gets enough data to conclude, and pause any ad set above 2x target CPA after 3 days.

Run 3-5 ad sets during testing, but only if each can reach roughly 50 optimization events per week at its budget. If the budget cannot fund that, consolidate into fewer ad sets or a single broad ad set rather than splitting budget thinner.

Use ABO during testing so each audience gets guaranteed spend and conclusive data. Switch to CBO only when scaling a proven winner, letting the algorithm allocate budget to the best performer.

Derive it from target CPA: multiply 50 optimization events by target CPA and divide by 7 days, or use the quick check of at least 5x target CPA daily. A $25 CPA implies roughly $180 per day per ad set.

Ad sets stay in learning when they cannot reach about 50 optimization events per week or when significant edits restart learning. Fix it by merging ad sets, optimizing for an earlier funnel event, or avoiding edits until learning completes.

Do not use it to score an existing account against audit checkpoints, configure pixels and conversion tracking, plan cross-channel budgets, or scale a proven winner. Those tasks belong to dedicated audit, tracking, media planning, and scaling skills.