How the budget increment rule works
The idea is simple. A campaign’s delivery settles once the system has seen enough results to predict who is worth showing the ad to. A large change to the budget alters what the campaign is being asked to do, and delivery can be pushed back into a learning state, where performance is unstable and cost per result is unreliable. Raising the budget in small steps, with time between them, is meant to add spend without triggering that reset.
It is worth being clear about the status of this rule. It is a practitioner heuristic that circulated widely among advertisers, not a published platform specification, and the specific step size people quote has no official basis. Meta documents that significant edits — including budget changes — can restart the learning phase, but it does not publish a safe threshold. Treat the rule as sensible caution, not as a setting you are obeying.
Why the budget increment rule matters
The underlying risk is real even if the number attached to it is folklore. Doubling the daily budget of a campaign that had just stabilised is a genuinely good way to spend more money at a worse cost per result for several days, and on a small account those days matter.
The rule also enforces something more valuable than its arithmetic: patience and one change at a time. Stepping up gradually means you can see which increase started the trouble, and you keep the option of returning to the last level that worked.
Where the budget increment rule goes wrong
The first problem is treating it as a law. Advertisers refuse a sensible increase because it exceeds a number they read in a blog post, and growth stalls for no reason. What actually matters is how large the change is relative to the campaign’s own volume — a small campaign is disturbed by a change that a high-volume one would not notice.
The second is assuming small steps make scaling safe. Frequent small increases still add up, and if the audience is saturated or the creative has tired, careful stepping only delays the same decline.
The third is applying it where it does not belong. The concern is about delivery relearning, which is not the same as a temporary spend cap, a lifetime budget, or a campaign that has never left learning in the first place because it produces too few conversions.
How to act on it
Raise budgets in steps that are modest relative to what the campaign already spends, and give each step several days before judging it. Change one thing at a time, so that if cost per result moves you know why. Keep a note of the level that last performed well, since that is where you return if a step goes badly.
Judge each increase against the cost per result you can afford at the new volume rather than the figure the campaign produced when it was smaller. And if performance falls away no matter how gently you step, the problem is not the increment — look at frequency, audience size and creative age before adding any more budget, as covered under scaling.
