CBO

Under CBO, now labeled Advantage campaign budget in the interface, one budget sits at the campaign level and Meta's delivery system decides, essentially per auction, which ad set inside that campaign is likely to produce the cheapest result right now, then routes more of the campaign's spend there. Each ad set can still carry its own bid strategy, cost cap, bid cap, or minimum ROAS, and the system factors that constraint into where it is willing to send money.

The uneven-distribution complaint people run into is mechanical, not a bug: an ad set that gets a cheap early result reinforces itself, since more budget produces more data, which produces more confidence, which produces more budget. Pair a warm retargeting audience with a colder prospecting audience in the same CBO campaign and retargeting will typically win most auctions on a per-result basis, quietly starving the prospecting ad set of the money it needs to prove itself, even though prospecting was the actual growth lever the campaign existed to fund.

Minimum and maximum ad set spend limits exist specifically to correct that: a minimum guarantees an underdog ad set keeps getting funded even while it is losing the algorithm's real-time popularity contest, and a maximum stops one ad set from absorbing the entire budget. The tradeoff is real: set a floor on every ad set in the campaign and you have mostly rebuilt ABO, just with an extra layer of the algorithm still nudging things at the edges, which defeats the reason to choose CBO in the first place. Those limits also go stale. A set of minimums calibrated for three ad sets can badly distort delivery once you have paused one and are down to two, or once you have added a fourth without revisiting the math; nobody deliberately breaks CBO this way, it accumulates from campaigns being edited over months without anyone re-checking whether the floors and ceilings still make sense for the current ad set count.

Pacing type still applies at the campaign level under CBO the same way it does under ABO: Standard Delivery lets the system spread spend toward the best auctions across the day, while Accelerated Delivery spends as fast as possible and is reserved for narrow, time-boxed promotions rather than everyday scaling.

CBO earns its place once you already trust the ad sets in a campaign to be reasonably comparable and you want Meta doing the reallocation work instead of you: several creative concepts inside one audience, several lookalike percentages, or a scaling phase after manual testing already identified winners. It is the wrong tool when ad sets represent a hard operational split that needs to be preserved regardless of relative performance, which is exactly ABO's use case instead.

In reporting, check spend by ad set inside the campaign regularly rather than only looking at the campaign-level total; if one ad set is absorbing nearly all of the budget while the others get scraps, that is the moment to add a minimum spend limit, split the mismatched audiences into separate campaigns, or accept that the algorithm found a genuinely better pocket and let it run.