ABO vs CBO — Why We Test in ABO and Scale in CBO

Campaign Budget Optimisation? Only After You've Mastered ABO

Most brands running Meta ads jump straight to CBO. They think it's "the latest and greatest." But the playbook's backwards. If your account isn't profitable, CBO will scale the wrong thing faster.

Here's how to actually use these two budget models to your advantage.

What ABO and CBO Actually Do

ABO (Ad Set Budget Optimisation) hands budget control to you. You set a daily or lifetime budget per ad set. Meta delivers ads to reach your goal within that budget. Simple. Manual. Conservative.

CBO (Campaign Budget Optimisation) hands budget control to Meta. You set one campaign budget. Meta automatically shifts spend between ad sets based on performance. Efficient. Autonomous. Aggressive.

The trap is thinking CBO is universally "better." It's not. It's faster at scaling what's already working, but it's faster at burying what's underperforming too.

Why Testing Demands ABO

When you're testing creative, CBO works against you.

Here's the scenario: You launch a power brief with 6 distinct ad concepts. Each goes into its own ad set. In CBO mode, Meta watches performance for 24–48 hours. The moment one ad set shows a small edge (even if it's noise), Meta starves the others and floods spend into the "winner."

That winner might just be a fluke. It had the best first hour. By the time you have meaningful data, you've wasted budget and missed learning what the other 5 concepts could really do.

This is called the winner-takes-all problem in testing. CBO solves it by learning what resonates, but it does that learning on your dime. You lose creative signal.

ABO forces you to keep budgets equal. Each ad set gets the same daily spend. Six ad sets at $100/day each gives you real comparative data. After 7–10 days, you can actually see which creative concept performs. Then you pause the losers and scale the winner.

The data is cleaner. The learning is faster. The winners are real.

When to Flip to CBO

Once you know what works, CBO becomes your acceleration tool.

Let's say testing revealed one creative concept converts at 1.8x ROAS and another at 1.2x ROAS. You want to scale the 1.8x version hard. You're confident in that ad set.

Move both into a CBO campaign and give Meta $5,000 daily budget. Meta will allocate 70–80% to the higher-performing ad set and 20–30% to the lower one (keeping it for learning signal). Your profitable spend grows faster.

This is the graduate pattern: Test at equal budgets in ABO. Identify winners. Move winners into CBO and scale.

The mistake is flipping to CBO during testing. You get fast scaling of noise.

The Ecom Republic Playbook

Our testing structure looks like this:

Testing Phase (ABO, 7–14 days):

• One campaign per audience/product

• 3–6 ad sets per campaign (one per creative concept)

• Equal daily budgets ($100–$300/ad set depending on niche)

• Run until you have 50+ purchases per ad set

• Track ROAS, CPA, click-through rate per creative

Scaling Phase (CBO, ongoing):

• Move 2–3 winning ad sets into one CBO campaign

• Set daily budget equal to 1.5x–2x your testing budget

• Let Meta optimise spend allocation between winners

• Keep 1 "control" ad set for fresh creative testing within the same campaign

Real example: A skincare subscription brand tested 6 concepts. After 10 days, 2 concepts hit 2.5x ROAS, 3 were breakeven, 1 was a dud. We paused the dud and the breakeven ones, moved the 2 winners into CBO with $2,000/day budget. In the next 14 days, that campaign generated 40% more sales at similar ROAS because Meta efficiently allocated between the two proven winners.

The third winning ad set? We kept it separate to test new angles. One of those new tests became the next ad set to graduate into the CBO campaign three weeks later.

ABO Decisions: The Cost Cap Advantage

Inside ABO, you have another choice: Lowest Cost vs Cost Cap.

Lowest Cost tells Meta, "Optimise for my goal (conversions, purchases, whatever), spending as little as possible per outcome."

Cost Cap tells Meta, "Get me outcomes at a max price I set. If you can't hit that price, pause."

Cost Cap is underrated. Most brands don't use it in testing because they think it'll limit data. It does. But it limits bad data.

If you set Cost Cap at $50 (your target CAC) and an ad set can't hit that in 48 hours, you get a signal: this creative/audience combo isn't viable at your cost target. Pause it. Don't let it burn $2,000 over seven days to confirm what you knew in day two.

This is especially powerful in subscription or high-CAC niches where every failed test is expensive learning.

The Numbers: Why Testing Structure Matters

One of our jewellery clients tested a rebrand campaign across 4 creative concepts in ABO. Budget $400/day ($100 per ad set).

After 7 days:

• Concept A: 1.9x ROAS, $62 CAC

• Concept B: 1.4x ROAS, $88 CAC

• Concept C: 0.8x ROAS (paused)

• Concept D: 1.1x ROAS, $105 CAC

Moving only Concept A into CBO would've cut off future learning. Instead, we:

1. Paused Concept C (clear loser)

2. Moved Concept A into a CBO campaign ($2,000/day)

3. Kept Concept B, D in the original ABO for continued testing

4. Launched 4 new concepts in a fresh ABO testing campaign

In 30 days, the CBO campaign on Concept A hit $89k spend at 1.8x ROAS. The new testing campaign found Concept E (2.1x ROAS) which we'll graduate next week. Concept B became a "retargeting only" creative (worked at 1.4x but not at scale).

None of that happens if you CBO-first.

The Trap: Turning Off Non-Winners

Here's the biggest mistake we catch: brands turning off underperforming ads in ABO testing.

If you're running 6 ad sets equally and one is at 1.0x ROAS while others are at 1.8x, resist the urge to pause it. It's still contributing learning signal. Turning ads off mid-test fragments your data. The winning ads' learning shifts. CPC changes. You introduce noise you can't separate from real creative performance.

Run ABO tests through their full window (at least 7 days, ideally 10–14). Pause only after the test ends, based on cumulative data.

The exception: If one ad set is catastrophically underperforming (0.4x ROAS or worse) and you're seeing it damage account health (click quality declining, high CPM, low relevance score), pause it and reallocate budget to others. But this should be rare.

Putting It Together

Use ABO for creative testing, market testing, and audience testing. Use CBO for scaling proven winners. The sequence looks like:

1. Equal-budget testing in ABO (7–14 days, one campaign per hypothesis)

2. Identify top performers (usually 1–2 out of 4–6 concepts)

3. Graduate winners into CBO campaign ($2k+/day budget)

4. Spin up new ABO testing campaign for next round of concepts

5. Keep a small allocation in ABO for evergreen testing (new hooks, new audiences)

One more thing: never move a single ad out of an ad set into CBO. Ads work as a portfolio within an ad set. Moving one ad strips it of its context. The winning ad set stays intact. That's the unit you graduate.

This is how you avoid the spray-and-pray approach to scaling. Test methodically. Graduate winners. Reinvest in testing. That's the flywheel.

Book a Growth Diagnostic Call to review your current campaign structure and see where you can tighten testing or unlock scaling potential. Book your Growth Diagnostic Call to discuss your account.