Guide · Updated
How does the Facebook ads learning phase work?
Short answer
The learning phase is the period after you launch an ad set or make a significant edit, while Meta's delivery system works out how to deliver it. It usually ends after about 50 results in the week after the last significant edit; each significant edit restarts that week. An ad set unlikely to get there shows Learning limited.
While an ad set is learning, its results are less stable and its cost per result is usually higher, so judge it on the week rather than the day. Meta's own advice comes down to three habits: avoid unnecessary edits, avoid running too many ads and ad sets at once, and set a budget large enough to get enough results. Start with the three posts below, then go to the question that matches what your Delivery column shows. For the terms used here, see the Meta ads glossary.
Start here
- Why is my Facebook ad set stuck in the learning phase?
How to tell whether an edit keeps restarting the ad set's week or it can't reach about 50 results, with seven fixes.
- How do I fix learning limited on Facebook ads?
The six causes Meta lists for Learning limited, the fix for each, and how much budget it takes to get out.
- Which edits reset the learning phase on Facebook ads?
Meta's list of edits that restart learning, what doesn't, and how to edit a live ad set without a reset.
How many conversions does an ad set need to exit learning?
- How many conversions does a Facebook ad set need to exit learning?
What counts as a result toward Meta's 50, the budget that implies, and the separate rule for Shops ads.
- Do I really need 50 conversions a week to exit learning?
Why 50 a week is a stability guideline rather than a hard gate, and what moves an ad set through.
- My campaign exited learning but still loses money, why?
Why leaving learning proves stable delivery, not profit, and the margin checks to run before scaling.
How do I get more results so an ad set can exit learning?
- Should I run fewer ad sets on a small budget?
How splitting a small budget across many ad sets starves each one, and how to consolidate.
- Which optimisation event should I choose on Meta?
A matrix that maps business outcomes to Meta optimization events by the signal each one needs.
- Should I optimise for add-to-cart instead of purchase?
When optimizing for add-to-cart is an honest way out of Learning limited, and when to switch back.
How do I scale without restarting learning?
- How do I scale a Meta ads budget without resetting learning?
A ramp cadence for raising budgets, and when to edit an ad set rather than duplicate it.
- How do I scale a winning ad without resetting learning?
How to move a tested winner into scaling without throwing away what it has learned.
- Does Advantage+ campaign budget skip or reset the learning phase?
Why Advantage+ campaign budget doesn't shorten learning, and which edits restart several ad sets at once.
Should I turn off a campaign that is still learning?
- Should I turn off a Facebook ad set that's still learning?
Five checks, in order, that tell a learning ad set from a losing one, and what turning it off costs.
- What does it cost to kill a campaign during learning?
What serial restarts cost, and a patience rule that still protects the budget.
- How do I know if a new campaign is failing or just volatile?
How to read days 1 to 7 by cumulative volume rather than daily swings.
- My new campaign has zero conversions, what do I check?
A 48-hour triage for zero conversions: check the event and the offer before the creative.
Sources: About the learning phase, About learning limited, Significant edits and learning phase (checked 1 October 2026).
Last updated