Free tool
Learning-phase exit estimator
Will your ad set escape Meta's learning phase, and when? Drop in last 7 days of leads — get a verdict.
Want this running 24/7 on your account?
Free tool
Will your ad set escape Meta's learning phase, and when? Drop in last 7 days of leads — get a verdict.
Want this running 24/7 on your account?
When you create or meaningfully edit an ad set, Meta's delivery system has no performance history for it and spends the first stretch exploring — testing placements, audiences and times of day to work out who converts. Performance during this period is volatile and usually worse than the ad set's eventual steady state. The phase ends once the ad set has accumulated enough conversion signal, which Meta puts at around 50 optimisation events in a rolling seven-day window.
The cost of staying in learning is real: unstable CPA, delivery that lurches between audiences, and results you cannot make decisions from. The cost of restarting it repeatedly — by editing budgets or creative mid-phase — is higher still.
Paste seven comma-separated daily conversion counts (a minimum of three values works). The tool sums them, takes the mean and standard deviation, computes the coefficient of variation, compares the first three days against the last three to get a trend, and projects the weekly total at your current daily rate.
The five states, in the order they are tested: Exited — total is at or above 50. Stuck, unstable signal — coefficient of variation above 0.6, so your daily numbers swing too hard for Meta to optimise against. Stuck, low volume — the seven-day projection is under 25 events, half the target, so no amount of patience will get you there. Trending to exit — the last three days beat the first three and the projection clears 50. Still learning — everything else. Days to exit is the remaining events to 50 divided by your current daily average.
Stuck on volume is a budget or audience problem: roughly double the budget or widen the targeting, because an ad set projecting 20 events a week is not going to drift into 50. Confirm the money is there to move with the budget reallocator rather than raising the account total. Stuck on variance is an editing problem — freeze the ad set, stop changing budgets daily, and let a week run clean.
Trending to exit and still learning both mean the same thing in practice: hold steady and resist the urge to tune. Once you have exited, that is the moment to scale, and the saturation curve estimator tells you how far you can push before efficiency drops. A frequent cause of chronic low volume is two ad sets splitting the same conversions between them — the audience overlap calculator will size that.
Last updated August 31, 2026. All 17 calculators and AI helpers are listed on the free tools hub.
FAQ
Meta's published guidance is roughly 50 optimisation events in a rolling seven-day window per ad set, and that is the target this estimator scores against. The event has to be the one the ad set is optimising for — 50 landing page views will not exit an ad set optimising for purchases.
Because Meta is trying to learn a stable pattern, and unstable delivery gives it nothing to learn from. The estimator computes a coefficient of variation (standard deviation ÷ mean) across your daily numbers and flags the ad set as stuck when it exceeds 0.6, even if the seven-day total looks acceptable. Erratic volume usually traces back to budget edits, so stop editing before you add spend.
As little as possible. Budget changes, audience edits, creative swaps and bid changes can all restart the phase. Hold spend steady and let volume accumulate. If the estimator says you are stuck on volume rather than stability, the fix is a bigger budget or a broader audience — not more edits at the current level.