What Is the Learning Phase in Facebook Ads?

When an ad set is new, or has just been changed in a way that matters, Meta’s delivery system still needs to learn how it may deliver and perform. It spends a stretch of time exploring before it settles. Meta calls that stretch the learning phase, and it expects performance to swing more while it lasts.

What happens during the learning phase?

Meta describes the delivery system as exploring the best way to deliver your ad set. While that exploration runs, ad sets are less stable and usually have a higher CPA, so early results are not necessarily a guide to how the ad set will perform later. Meta adds that the system never stops learning about an ad set; the phase simply ends once the ad set can deliver stably.

How many results does an ad set need to exit learning?

Meta says exit usually happens after about 50 results in the week after the ad set’s last significant edit. Shops ads follow a different rule.

Situation	Meta’s guideline
Standard ad set	About 50 results in the week after the last significant edit
Shops ads	At least 17 purchases through your website and 5 through Meta
Unlikely to reach about 50 optimization events in that week	Delivery status changes to Learning limited

What sends an ad set back into learning?

Only a significant edit. Meta’s list includes any change to targeting, ad creative or optimization event, adding a new ad, pausing the ad set for 7 days or longer, and changing bid strategy. Changes to budget, bid control, cost per result goal, ROAS goal or ad set spending limit may or may not count, depending on how large they are.

How do you check where an ad set stands?

1. Go to Ads Manager and select the ad set you want to review.
2. Find the Last significant edit column; if you don’t see it, click Columns and add it.
3. Check the Results column for the number of results since that edit.

What does Meta recommend while an ad set is learning?

Meta’s best practices for this period are:

• Wait before editing. Editing an ad, ad set or campaign during learning resets it and delays optimization.
• Avoid unnecessary edits. Only edit when you have reason to believe the change should improve performance.
• Avoid high ad volumes. With many ads and ad sets, the system learns less about each one; combining similar ad sets combines their learnings.
• Use realistic budgets. A very small or inflated budget gives the system an inaccurate signal, and frequent budget changes can restart learning.

Meta also says you shouldn’t try to avoid the learning phase completely. Testing new creative and marketing strategies is how performance improves over time.

When should you judge an ad set’s costs?

Meta suggests evaluating cost per optimization event after the initial learning phase or since your last significant edit, once the ad set has stable delivery. It recommends waiting for around 50 optimization events, while noting that some ad sets stabilize earlier.

FAQ

Is a higher cost per result during learning normal?

Yes, according to Meta. It says ad sets in the learning phase are less stable and usually have a higher CPA, and results from this period aren’t necessarily indicative of future performance. Meta suggests waiting for around 50 optimization events before judging cost per optimization event.

Does Advantage+ campaign budget make learning take longer?

Meta says no. The initial learning process for ad sets using Advantage+ campaign budget takes the same time as with ad set budgets, and ad sets do not re-enter learning as the campaign shifts budget between them.

What happens if an ad set can’t reach about 50 results?

If an ad set is unlikely to get about 50 optimization events in the week after its last significant edit, Meta changes its delivery status to Learning limited. Meta describes this as a sign that budget isn’t being spent effectively, not as a penalty.