ABO vs CBO: Test in One, Scale in the Other
ABO is for testing, CBO for scaling, and most bad results come from setup, not the budget mode. When to use each on Meta, and how to move winners.
Jonathan Solomon · Updated
The short answer
Use ABO to test, CBO to scale. Ad set budgets give every audience, interest or event a fixed share of spend, so a test actually reads. Once a winner is clear, move it into CBO, now called Advantage+ campaign budget, and let Meta push spend toward what converts. Keep the optimization event the same from test to scale.
ABO vs CBO at a glance
| Dimension | ABO | CBO |
|---|---|---|
| Where the budget sits | On each ad set | On the campaign |
| Who decides spend | You | Meta |
| Best job | Testing audiences, events and creative | Scaling proven winners |
| Audience control | Full, per ad set | Shared; Meta favours early leaders |
| Main risk | Overspending on losers if nobody prunes | Starving ad sets before they can be judged |
| Effort | Hands-on; you move the money | Lighter day to day |
| Default for new campaigns | Opt in | On by default for Sales, Leads and App campaigns |
When ABO wins, and when CBO does
Choose ABO when
- You are testing a new audience, interest, event or creative.
- Each variant needs a guaranteed share of spend to produce a readable result.
- Spend is modest and you want to control exactly where it goes.
Choose CBO when
- You have proven ad sets doing the same job, with the same objective.
- There is enough volume for Meta to learn: its learning-phase guidance points to about 50 optimization events per ad set within 7 days.
- You want spend reallocated in real time without managing it by hand.
The third option
Run both, as stages. Test in ABO, graduate the winners into a CBO scaling campaign, and keep a small ABO lane running so the next winner is always being found. Budget mode is a stage decision, not an account-wide setting.
Refinity’s take
Most "CBO doesn't work" stories are setup problems. The usual culprit is a campaign with mixed objectives, or an untested audience dropped into a scaling structure and judged on the first few days. CBO did what it is built to do: it found the cheapest results and spent there.
So we test where the controls live. Audience targeting (interests, events, exclusions) sits at the ad set, which makes the ad set the unit of a test, and ABO is what holds each one's budget steady long enough to read. When a variable wins, it moves to CBO, where Meta's job is to scale it, not to pick it.
Structure first, budget mode second.
— Jonathan Solomon, Founder
— Jonathan Solomon, CEO / Accounts Manager
The setup check: five things to get right before choosing a budget mode
Budget mode only works on top of a sound structure. Run every campaign through these five checks first:
- One optimization event, test to scale. Every ad set in a campaign should optimize for the same event, and the scaling campaign should optimize for the event the test was judged on. Under CBO, ad sets chasing different events compete on unequal terms and the budget drifts to whichever is cheapest; Meta itself recommends ad set budgets when optimization goals are mixed.
- One variable per test. Each ad set in a test should differ by one thing: the audience, the interest, the event or the creative. Change two at once and a winner tells you nothing you can reuse.
- A budget floor per ad set. Give each test ad set enough spend to work through Meta's learning phase: roughly 50 optimization events within 7 days. Below that, the result is noise.
- A written graduation rule. Decide before launch what "winning" means (cost per result, volume and timeframe), so the move to scaling is a rule, not a mood.
- A clean scaling campaign. Promote winners into a separate CBO campaign that holds only proven ad sets with the same objective. Don't drop untested ideas into it.
If a campaign fails a check, fix the structure before touching the budget mode.
Why tests belong in ABO
CBO is designed to find the cheapest results early and concentrate spend there. That is exactly right for scaling and exactly wrong for testing. A variant that would win over a full week can lose its budget in the first two days because another ad set got cheap results first. ABO removes that competition: each ad set spends its own budget, so every variant gets a fair read.
Why scaling belongs in CBO
Once ad sets are proven and share an objective, the question changes from "which works?" to "how much can each absorb?" Meta can answer that faster than a person moving budgets by hand: campaign budget optimization distributes one budget across ad sets in real time, toward the best opportunities. That is the job it was built for.
The 2026 wrinkle: you may already be in CBO
Meta renamed campaign budget optimization to Advantage+ campaign budget, and it is now switched on by default when you create Sales, Leads and App Promotion campaigns. Owners who never chose CBO are often running it, and testing inside it, without realizing why some ad sets never spend. Check the campaign's budget setting before reading any test.
The misconception: "CBO is always better because Meta's algorithm knows more"
Meta's delivery system is good at allocating spend between options that already work. It is not built to give an unproven option a fair hearing. Treating CBO as a testing tool hands the choice of winner to whichever ad set happened to get cheap results first.