Uğur Deniz İlhan
  1. Home
  2. Insights
  3. Growth Marketing
  4. How is the learning phase budget set for Meta ads?
Growth Marketing · Guide · 19 November 2026 · 6 min read

How is the learning phase budget set for Meta ads?

Ad sets launched on Meta with a low budget often get stuck in the learning phase. This guide covers how to calculate the minimum budget, what to do when the budget falls short and what resets learning.

Growth Marketing50 events a week
Short answerOn Meta, an ad set leaves the learning phase after about 50 optimisation events following its last significant edit. The minimum weekly budget is therefore 50 times the expected cost per event. If the budget falls short, the fix is not to force it but to combine ad sets, broaden the audience or pick a more frequent event.

What is the learning phase and why does it matter?

The learning phase is the initial period in which Meta’s delivery system learns whom to show an ad set to. According to Meta, an ad set leaves this phase once it reaches about 50 optimisation events after its last significant edit, and its status then shows as “Active”.

The optimisation event is the action the ad set aims for: a purchase, a form submission or an add to cart, for example. An ad set that is unlikely to reach that number within a week is labelled “learning limited”. Meta describes this not as a penalty but as a sign that the system cannot optimise well with the current setup.

How is the minimum budget calculated?

The calculation is three lines:

Weekly budget per ad set = 50 × expected cost per event

The daily budget per ad set is one seventh of the weekly amount; with more than one ad set, the amount is multiplied by the number of sets. The figures below were chosen only to show the calculation.

Input and resultValueHow it was found
Expected cost per event₺180From past campaigns or an estimate.
Weekly per ad set₺9,00050 × 180
Daily per ad setabout ₺1,2869,000 ÷ 7
Daily for two ad setsabout ₺2,5711,286 × 2

To run the calculation with your own figures, use the learning phase budget tool. The expected cost is an estimate and may not hold in the first days; that is why the calculated amount is a floor, not a target.

What if the budget cannot reach the threshold?

Keeping the budget below the threshold and waiting leaves the ad set in learning limited. These are the options Meta suggests:

OptionWhen it fitsWhat to watch
Combine ad setsWhen several small sets aim at the same goal.Combining is a significant edit; learning starts again.
Broaden the audienceWhen the audience is narrowly defined.Cost per event may change in a wider audience.
Choose an event that happens more oftenWhen purchases cannot get close to 50 a week.The system optimises for the chosen event; an add to cart, for example, may not turn into a purchase.
Raise the bid or cost controlWhen the bid limit is restricting delivery.Cost per event rises; compare it with break-even.

For how the last option affects profitability, see the break-even ROAS article.

Which changes reset learning?

According to Meta, pausing the ad set and changing the optimisation event, the audience or the creative count as significant edits. Bid strategy and budget changes may also count, depending on the size of the change. After a significant edit, the count towards 50 events starts again.

In our view it is therefore better to batch changes and make them in one go rather than spread them over several days. A small fix made every day can keep an ad set in learning indefinitely.

Our recommendation

Set the budget from the threshold calculation, not from a percentage or habit.

  1. Work out the cost per optimisation event from the last 30 days of data. If there is none, start with an estimate close to your sector and correct it at the end of the first week.
  2. Calculate the threshold and see how many ad sets your total budget can carry. If it cannot carry them, reduce the number of sets.
  3. If purchases cannot reach the threshold, choose the event one step higher in the funnel and track conversion to purchase separately.
  4. Make changes once a week, in a single time window.
  5. Do not judge an ad set’s performance before its learning phase is over.

Frequently asked questions

Is 50 events a fixed threshold?

No. Meta says “around 50”. That is why the calculation is used as a floor; the budget is planned a little above it rather than exactly at it.

Is learning limited a penalty?

According to Meta, no. Learning limited is a warning that the ad set is unlikely to get enough optimisation events with its current setup and that the delivery system cannot optimise well.

Does raising the budget reset learning?

It can. Meta states that budget and bid changes may count as significant edits depending on the size of the change. Small, infrequent adjustments are safer.

Sources

Read next

All insights

For a conversation, message on WhatsApp.

A short introduction is enough: your company and what you would like to discuss.

Open WhatsApp

+90 530 242 14 55