Meta's learning phase ends when an ad set can deliver stably, which Meta says usually occurs after about 50 results in the week after the ad set's last significant edit, per its help page About the learning phase as of November 2026. We call this the 50-event rule. If your ad set cannot reach that pace, Ads Manager shows Learning limited. The practical consequence is budget: to get 50 results in 7 days, an ad set needs roughly 7.14 times its cost per result in daily spend.
This page covers what Meta publishes about learning, which edits reset it, how to size budgets against it, and how Google's Smart Bidding learning period differs. Every rule is copied from Meta Business Help or Google Ads Help. The budget math is arithmetic on Meta's published threshold; Meta itself does not publish a budget formula.
What Meta says about the learning phase.
- Definition. The learning phase is the period when the delivery system still needs to learn how an ad set may deliver and perform. It starts when you create a new ad or ad set or make a significant edit.
- Status. The Delivery column reads Learning while an ad set is in the phase.
- Exit. Ad sets exit as soon as they can deliver stably, usually after about 50 results in the week after the last significant edit.
- Cost. Meta says that during learning, ad sets are less stable and usually have a higher CPA.
- Shops ads exception. For Shops ads, Meta requires a minimum of 17 purchases through your website and 5 through Meta for learning to complete.
Meta's wording shifts slightly between pages. The learning phase page says about 50 results; the Learning limited page says about 50 optimization events. They mean the same thing: the event you chose as the ad set's performance goal, counted from the last significant edit, not from launch. Ads Manager has a Last significant edit column you can add to see that date.
Learning phase budget math.
The formula follows directly from the 50-event rule. Fifty results in seven days is 7.14 results a day. Multiply by what each result costs.
| Target CPA | Results per day needed | Minimum daily budget | Minimum weekly budget |
|---|---|---|---|
| $10 | 7.14 | $71.43 | $500 |
| $25 | 7.14 | $178.57 | $1,250 |
| $50 | 7.14 | $357.14 | $2,500 |
| $100 | 7.14 | $714.29 | $5,000 |
| $200 | 7.14 | $1,428.57 | $10,000 |
If the budget column looks out of reach, Meta's own guidance points at the other levers: combine ad sets so one ad set gets all the results, or change the optimization event to one that happens more often. A $100 purchase CPA might be a $20 add-to-cart CPA, which needs about $143 a day instead of $714. That trade gets the ad set out of learning but optimizes for a weaker signal. Work out what a result is worth first with our break-even CPL calculator or the break-even ROAS formula.
| Optimization event | Assumed CPA | Results per week at $350/day | Reaches about 50? |
|---|---|---|---|
| Purchase | $100 | 24.5 | No, likely Learning limited |
| Initiate checkout | $40 | 61.3 | Yes |
| Add to cart | $20 | 122.5 | Yes |
| Lead (form) | $35 | 70 | Yes |
The CPAs in that second table are examples to show the math, not benchmarks. Use your own account's cost per result.
What counts as a significant edit.
Meta's significant edits page says every edit has some effect on delivery, but only a significant edit sends an ad set back into learning. It splits edits into two groups.
| Edit | Restarts learning? |
|---|---|
| Any change to targeting | Yes, always |
| Any change to ad creative | Yes, always |
| Any change to the optimization event | Yes, always |
| Adding a new ad to the ad set | Yes, always |
| Pausing the ad set for 7 days or longer | Yes, once you unpause |
| Changing bid strategy | Yes, always |
| Changing the disclaimer on social issue, election or political ads | Yes |
| Budget amount | Depends on the size of the change |
| Bid control, cost per result goal or ROAS goal amount | Depends on the size of the change |
| Ad set spending limit amount | Depends on the size of the change |
| Advantage+ campaign budget redistributing spend | No |
| Adding a new ad set to an Advantage+ campaign budget campaign | No, other ad sets are unaffected |
Meta does not publish a percentage threshold for budget changes. Its only example is that raising a budget from $100 to $101 is unlikely to restart learning, while changing it from $100 to $1,000 may. Any specific percentage you read elsewhere is a rule of thumb, not Meta's rule. With Advantage+ campaign budget, Meta warns that changing the campaign budget or the campaign bid strategy can push several ad sets back into learning at once.
Learning limited status.
Meta's About learning limited page says an ad set becomes Learning limited when it is unlikely to receive about 50 optimization events in the week after the last significant edit. The diagnosis can appear any time after you create the ad set or make a significant edit. Meta says it is not a penalty; it signals that budget is not being spent effectively because the system cannot optimize with the current setup. For Shops ads, it appears when the ad set has not produced 17 website purchases and 5 Meta purchases after 7 days.
| Cause Meta lists | Fix Meta suggests |
|---|---|
| Small audience size | Expand your audience |
| Low budget | Raise your budget |
| Low bid or cost control | Raise your bid or cost control (a tooltip appears on hover when this is the cause) |
| High auction overlap or too many ads at once | Combine ad sets and campaigns |
| Infrequent optimization event | Choose an event that happens more often, for example add to cart instead of purchase |
If a Learning limited ad set later gets enough optimization events since its last significant edit, Meta moves it from Learning limited to Active. You can add an Optimization events column under Customize columns to watch the count.
Google Ads Smart Bidding learning period.
Google's equivalent is the Learning status on Smart Bidding strategies, and the rules are different. Google's page About Smart Bidding learning says the status never lasts more than 7 days and that ads keep serving as normal during it. It shows after major structural changes: a bid strategy that was recently created or reactivated, campaigns, ad groups or keywords added to or removed from the strategy, or conversion goals or actions that changed significantly. Google says changes to budgets, targets, creatives or targeting do not trigger Learning status.
Google also says it can take 1 to 2 conversion cycles for performance to stabilize to a new ROI goal. Earlier versions of Google's help content and some trade coverage cited about 50 conversion events or 3 conversion cycles; the live page we opened in November 2026 does not include that figure, so we do not treat it as Google's current guidance.
| Question | Meta Ads | Google Ads Smart Bidding |
|---|---|---|
| Published threshold | About 50 results in the week after the last significant edit | No event count published; status lasts up to 7 days |
| Unit | Ad set | Bid strategy |
| Do budget changes reset it? | Large ones may | No, per Google |
| Do creative changes reset it? | Yes | No, per Google |
| Do targeting changes reset it? | Yes | No, per Google |
| Bid strategy change | Yes | Yes (new or reactivated strategy) |
| Stuck state | Learning limited | Not published |
| Stabilization note | Higher CPA during learning | 1 to 2 conversion cycles to stabilize to a new ROI goal |
Exceptions and distinctions.
- Learning never stops. Meta says the delivery system never stops learning; exiting the phase means delivery is stable, not that optimization ends.
- Count from the edit, not the launch. An ad set that launched a month ago and had its creative swapped yesterday is on day one.
- Don't avoid learning entirely. Meta says the learning phase is necessary and that testing new creative is essential. The goal is fewer, bigger edits, not zero edits.
- Ad volume dilutes learning. Meta says many ads and ad sets each learn less, and combining similar ad sets combines learnings.
- Tracking quality matters. If your pixel misses conversions, the ad set reaches 50 more slowly. A Conversions API setup recovers events the browser loses.
Learning phase checklist.
If you want this run for you, see our Facebook and Instagram ads service.
Questions people ask about the Meta learning phase.
How many conversions does Meta need to exit the learning phase?
About 50 results (optimization events) in the week after the ad set's last significant edit, per Meta Business Help as of November 2026. Shops ads need 17 website purchases and 5 Meta purchases.
How long does the Meta learning phase last?
Meta does not set a fixed duration. The window it measures is the week after the last significant edit, and the ad set exits once delivery is stable, usually after about 50 results.
How much budget do I need to exit learning?
About 7.14 times your target CPA per day (50 divided by 7). At a $25 CPA that is about $179 a day; at $50 it is about $357. This is arithmetic on Meta's threshold, not a Meta formula.
Does increasing my budget reset the learning phase?
It may, depending on the size of the change. Meta's example: $100 to $101 is unlikely to reset learning; $100 to $1,000 may. Meta publishes no percentage threshold.
What does Learning limited mean?
Meta judges the ad set unlikely to get about 50 optimization events in the week after its last significant edit. Meta says it is not a penalty. Fixes are combining ad sets, broadening the audience, raising budget or bid, or choosing a more frequent event.
Does pausing an ad set reset learning?
Pausing for 7 days or longer is a significant edit, and the ad set re-enters learning when you unpause it.
How long is the Google Ads Smart Bidding learning period?
Google says the Learning status never lasts more than 7 days, and performance can take 1 to 2 conversion cycles to stabilize to a new ROI goal.