CBO vs ABO: Should Your Meta Budget Live at the Campaign or Ad Set Level?
ImageFactory Engineering · Published 2026-06-18
If you're stuck on whether your Meta budget should sit at the campaign level (CBO, Advantage campaign budget) or the ad set level (ABO), the answer is simple: test with ABO, scale with CBO. When you want a clean read on whether new creative works, use ABO so every ad set gets equal spend. When you already have a proven winner and want the algorithm to scale it, use CBO. You're not picking one forever — the 2025-2026 standard is to use both, in sequence.
What's the actual difference between CBO and ABO?
ABO (Ad Set Budget Optimization) sets a separate budget for each ad set. Give Ad Set A $100/day and Ad Set B $100/day, and that's roughly what each spends — regardless of which performs better. Spend is equally isolated.
CBO (Campaign Budget Optimization, called Advantage campaign budget in Ads Manager) puts one budget on the whole campaign, and Meta's algorithm distributes it to whichever ad sets it predicts will perform best. If one ad set converts early, Meta pushes most of the budget there fast (AdsUploader). Meta's own data shows Advantage automation campaigns saw roughly 12% lower cost per purchase than manually managed ones (AdAmigo).
So when do I use ABO, and when do I use CBO?
The deciding question is: what are you trying to do right now?
ABO wins when you're testing new creative or new audiences. Equal, isolated spend gives you clean data on which variable actually won. It's also the right call when your creative hit rate is high (6-7 of 10 ads work), when daily budgets are small (roughly under $50), or when low conversion volume — like niche B2B — demands precise control (AdsUploader).
CBO wins when you already know what works and want to scale it. The algorithm is built for efficiency at volume, not isolation — bundle your proven winners into a CBO and let it grow them without micromanagement. Counterintuitively, when your hit rate is low and uncertain, CBO's automatic funneling toward winners can be more cost-efficient than ABO (Jon Loomer Digital).
Should new creative go in the same ad set or a new one?
This is the most common question we get over KakaoTalk, and the answer depends on your goal.
If the goal is testing, put it in a new ABO ad set. Cramming extra creatives into an existing ad set means Meta controls which one gets impressions — and your comparison gets dirty. Isolate creative as the only variable, keep targeting identical, give each ad set equal budget, and run for 3-4 days before judging (AdsUploader).
If the goal is scaling, it's a different story. Don't switch off or pull a winning ABO creative. Leave the ABO original running and duplicate it by post ID into a new CBO campaign — post ID duplication preserves social proof like likes, comments, and shares. The winner now works in both campaigns at once (AdsUploader).
Why does testing get messy inside CBO?
Because CBO doesn't split spend evenly across ad sets. If one ad set converts early, Meta dumps almost the entire budget there, and your other candidate creatives die before gathering enough impressions to prove themselves (AdAmigo). That makes CBO a poor tool for answering "which creative wins?"
If you absolutely need early fairness inside a CBO, set a per-ad-set minimum spend for about a week so every candidate gathers data, then remove the minimums and let the algorithm optimize. Keep it to 3-5 ad sets per CBO — more than that fragments the budget inefficiently (AdsUploader).
What does "change one variable at a time" really mean?
If you turn on Advantage+ Audience, Advantage+ Placements, and Advantage campaign budget all at once, you won't know what drove the lift. To isolate the CBO effect, run a side-by-side test — one campaign on manual budgets (ABO), one on CBO, same conditions — for 7-10 days and compare CPA and ROAS (AdAmigo). Always one variable at a time.
Checklist
- Define the goal first — testing or scaling? Testing means ABO; scaling means CBO.
- Test in a new ABO ad set — don't cram into an existing one. Creative is the only variable; targeting and budget stay equal.
- Give the test 3-4 days minimum — don't judge too early. Respect the learning window.
- Duplicate winners into CBO by post ID — leave the ABO original on. Preserve social proof.
- Keep CBO to 3-5 ad sets — more fragments the budget.
- Need CBO fairness early? Set a per-ad-set minimum for one week — then remove it and hand control to the algorithm.
- One variable at a time — don't switch on audience, placement, and budget automation together.
- Validate a CBO switch with a 7-10 day side-by-side — compare on CPA and ROAS before committing.
For how to grow the budget once you reach the scaling phase, see our ad budget scaling guide.
How does ImageFactory make this testing cheaper?
Whether you run CBO or ABO, the test only means something if you can hand the algorithm several distinct creatives to choose from. Loading a different creative into each ABO ad set, then duplicating winners into CBO, is ultimately a fight over creative supply. To test 3-5 fresh creatives a week, that supply has to be cheap.
ImageFactory makes it cheap. Upload one master or product cutout, generate multiple style variations, and adapt them to 1,400+ placements across 110+ platforms. Text and logos aren't re-rendered — the original pixels are preserved — safe zones are handled automatically, and distortion stays at 0-2%. It supports up to 2K/4K and 15 languages. That means you can produce both the candidate creatives for your ABO test and the per-placement variations you'll scale in CBO, all in one pass. Try it with the 14-day free trial.