CBO vs ABO: which Meta budget setup to use in 2026

The difference is who decides how the money is split. With CBO (campaign budget optimization) you set one daily or lifetime budget on the campaign, and Meta distributes it across the ad sets in real time. With ABO (ad set budget optimization) you set a budget on every ad set, and each ad set spends roughly what you gave it.

Neither is better in general. They answer different questions. ABO answers: how does each of my variants perform with equal money? CBO answers: where does Meta find the cheapest results right now? If you mix up the two questions, you get tests that prove nothing and scaling campaigns that stall.

Advantage campaign budget is Meta's current name for CBO. Meta has since moved the feature under its Advantage+ brand, and the Ads Manager label now reads Advantage+ campaign budget. According to Meta's product page, it distributes one budget in real time to the ad sets with the best opportunities, and all ad sets must use the same budget type, the same bid strategy and standard delivery.

Meta claims the feature can lower cost per action by an average of 4.6%. Treat that as Meta's own figure: the page does not show the study period or the sample. It is a reason to test CBO, not a guarantee.

In the streamlined campaign setup that Meta rolled out in 2025, the budget toggle for Advantage+ is on by default when you create a sales campaign. You can still switch to an ad set budget. If you keep CBO, you can set ad set spend limits (a minimum and a maximum per ad set) to stop Meta from starving a small ad set or dumping everything into one.

Ad set budget sharing is a hybrid between ABO and CBO. You keep a budget on each ad set, but Meta may move up to 20% of an ad set's budget to other ad sets in the same campaign when it expects better results there. The other 80% stays where you put it.

Per Meta's developer documentation, the rules are strict: daily budgets only, no lifetime budgets, at least two ad sets, the same bid strategy across ad sets, and no campaign budget. You can switch sharing off in a running campaign, but you cannot switch it on mid-flight.

For anyone who creates campaigns through the API or a tool: since API v24, the field `is_adset_budget_sharing_enabled` is mandatory whenever the budget sits on the ad sets. You must send true or false explicitly.

{
  "name": "TEST | Creatives | Sept 2026",
  "objective": "OUTCOME_SALES",
  "special_ad_categories": [],
  "is_adset_budget_sharing_enabled": false,
  "status": "PAUSED"
}

My take: switch sharing off for creative and audience tests, because a 20% shift is enough to blur a close result. Leave it on for evergreen ABO setups where you mainly want fixed splits per country and do not mind Meta smoothing the edges.

Use ABO when the goal is learning, and CBO when the goal is efficient volume. The table below covers the situations I see most in accounts, from small shops to budgets in the six figures per month.

Budget is the deciding factor more often than people admit. Meta needs about 50 optimization events per ad set within 7 days to leave the learning phase. That gives you a quick check for how many ad sets your budget can carry.

Weekly budget per ad set ≈ 50 × target CPA | Daily ≈ 7 × target CPA

Example: your target CPA is €30. One ad set needs roughly €1,500 per week, or about €215 per day, to exit learning on purchases. With €100 a day in total, splitting into five ABO ad sets gives each €20, and none of them will ever collect enough data. In that case run one ad set, or optimize for an earlier event such as add to cart until volume grows. If you are unsure what your target CPA should be, work it out from your margin with the ROAS calculator.

Yes. Go through these questions in order and stop at the first answer that fits.

1. Do you need a clean answer to a test question (which creative, which audience)? Use ABO with budget sharing off, one variable per test.
2. Does someone outside the ad account fix the split (per country, per product line, per client)? Use ABO, budget sharing optional.
3. Can your total budget fund about 50 conversions per week for each planned ad set? If not, cut the number of ad sets until it can, then continue.
4. Are the ad sets similar in audience size and funnel stage? If yes, use CBO. If no, split them into separate campaigns or use CBO with minimum spend limits.
5. Is this a sales campaign with broad targeting and a proven offer? Keep the default Advantage+ campaign budget and focus your time on creatives.

They can, depending on size. Meta defines a significant edit as pausing an ad set or changing the optimization event, audience or creative. On budget and bid strategy, Meta's help center only says a change may also be significant, depending on its magnitude. Meta publishes no percentage threshold.

The famous rule of never raising budgets more than 20% at once comes from agency blogs, not from Meta. The mechanism behind it is real: a large jump changes who the system can reach at what price, and results often wobble for a few days. The number itself is folklore.

• Raise budgets in steps and give each step a few days of data before the next one. Many practitioners use 15 to 30% steps, which is a sensible habit, not a Meta rule.
• High volume ad sets with hundreds of conversions per week tolerate bigger jumps because they relearn quickly.
• In CBO, a change to the campaign budget affects every ad set at once. In ABO you can scale only the winner.
• Bundle edits. Changing budget, audience and creative on three different days means three rounds of instability instead of one.
• Duplicating a winning ad set to scale it starts a brand new learning phase. Sometimes that is worth it, often it is not.

• Judging a CBO campaign after one day. Spend distribution in the first 24 to 48 hours is noisy.
• Putting retargeting and prospecting ad sets into one CBO campaign, so the small warm audience gets almost nothing.
• Running ten ABO ad sets on a budget that can feed two.
• Setting minimum spend limits so high that CBO becomes ABO with extra steps.
• Leaving ad set budget sharing on during a test and then wondering why the losing ad set spent less than planned.
• Reading the winner from platform ROAS alone. Meta, Google and TikTok each count conversions in their own favor, which is why we compare against true ROAS based on real revenue.
• Pausing ad sets inside CBO the moment one looks weak, which resets the balance for all others.

An approval-based agent proposes the budget move, explains it, and waits. In Marketing AIgency, agents review the Meta, Google and TikTok accounts, look into the numbers before they suggest anything, and put every proposal with its reasoning into an approval queue. Nothing runs until you approve it. If you want, you can set single action types such as small budget increases to autonomous within limits you define.

Three details matter for budget work. New accounts start in shadow mode, so the agents only watch and suggest at first. Guards check budgets and ads every 30 minutes, around the clock, which catches a runaway ad set long before the next morning. And every change goes into a log with the data and the reasoning, and its effect is measured after 7 days.

The agents steer on your company number, meaning real revenue or real leads from your shop, CRM, spreadsheet or first-party pixel, not on platform ROAS. That matters for CBO decisions. For a connected-car app that runs on the same agent system in 9 markets, Meta, Google and TikTok together claimed 2.2 times the revenue the company actually made in Germany in September 2026. Budget moves based on those platform numbers would have shifted money toward ad sets that only looked better.

In that setup, the agents made 316 proposals in the first 11 days. 101 were executed and 69 were rejected with a reason, and those rejections feed back into future proposals. Pricing is a fixed monthly fee with no share of your ad spend, see plans and pricing. If you are weighing this against an agency, our breakdown of Facebook ad costs shows what media itself costs in 2026.