Meta learning phase: common questions
How long does the Meta learning phase last? +
The Meta learning phase ends when your ad set receives 50 optimization events within a 7-day rolling window. The clock does not tick in days - it ticks in events. At a typical e-commerce CPA of $40 and a $100 daily budget, you generate roughly 2.5 purchases per day, which means you need about 20 days to exit learning. Increase your daily budget to $285 or lower your CPA and you will exit in 7 days. The learning phase can also end earlier if Meta sees enough delivery data to stabilize CPMs, even without hitting the full 50-event threshold.
What does Learning Limited mean on Meta? +
Learning Limited means your ad set will never exit the learning phase at its current pace. Meta shows this status when your ad set does not receive enough optimization events - typically fewer than 10 per week. The most common causes are: daily budget too low relative to your target CPA, audience too small to deliver at scale, campaign too fragmented across too many ad sets splitting the same budget, or a bid cap set too aggressively. The fix is almost always to consolidate ad sets, increase budget, remove bid caps, or do all three.
What actions reset the Meta learning phase? +
The following actions reset your learning phase and restart the 50-event clock: editing your daily or lifetime budget by more than 25% in a single change, changing your bid strategy, changing your optimization event, changing your audience targeting significantly, pausing an ad set for 7 or more consecutive days, adding a brand-new ad to the ad set, and changing your campaign objective. Editing existing ad creative or copy causes a minor re-optimization period but is less severe than a full reset. Viewing reports, renaming campaigns, or downloading data has no impact on learning.
Should I use a Cost Cap or Bid Cap during the learning phase? +
Avoid Cost Cap and Bid Cap during the learning phase. These constraints tell Meta to stop delivering when the cost exceeds your cap, which often means delivery becomes erratic and you never accumulate the 50 events needed to exit learning. Run Lowest Cost (no bid cap) during the learning phase to let Meta spend freely and collect data. Once you have exited the learning phase and have a stable CPA baseline, then layer in a Cost Cap set at roughly 1.2x to 1.5x your target CPA to control efficiency at scale.
How many ad sets should I run to avoid fragmenting my budget? +
Each active ad set must independently reach 50 optimization events per week. If your monthly Meta budget is $3,000, that is roughly $100 per day. At a $40 CPA, you generate 2.5 purchases per day total. Running three ad sets means each gets roughly 0.8 purchases per day - which means all three will be permanently Learning Limited. The rule of thumb: run the fewest ad sets possible so each one can accumulate events quickly. Most accounts under $10,000 per month should run 1 to 3 ad sets. Add ad sets only when the account generates enough volume to support them.
