Should I test different audiences in different ad sets?

Not on a small budget. Each ad set has to learn on its own, and Meta says that takes about 50 results in 7 days. Test different ideas as ads inside one broad ad set first. Test audiences later, when one ad already sells.

Is CBO better than ABO?

Neither is better everywhere. ABO gives each ad set a fixed budget, so a new idea gets a fair share. CBO lets Meta move the budget to whichever ad set gets the cheapest results. With only one ad set, the two work almost the same.

Should a new shop start with Advantage+?

Advantage+ Sales campaigns let Meta choose more of the setup, and they work best with sales data to learn from. A new account with few sales can start with a simple manual campaign: one broad ad set and automatic placements. Try Advantage+ once your pixel records sales every week.

A simple campaign structure for a small budget

Most new advertisers build too much. Three audiences, two campaigns, ten ads, all on RM 50 a day. It looks thorough. In practice, every part gets too little money to learn anything. A simple structure gives each ringgit a job.

The three levels

Every Meta campaign has three levels, and each one holds different settings.

What each level of a Meta campaign controls

The settings held at the campaign, ad set and ad level of a Meta ads campaign.
Level	What it holds
Campaign	The objective (Sales, Leads and so on), and the budget if you use a campaign budget
Ad set	Who sees the ads (location, ages), where they show, the result Meta looks for, and the budget if you set it here
Ad	The picture or video, the post text, the headline, the button and the link

Picking the objective is the lesson before this one: Pick the right campaign objective.

Why fewer ad sets win on a small budget

Each ad set learns on its own. Meta says an ad set needs about 50 results in 7 days to leave the learning phase. Split RM 60 a day across three ad sets and each one gets RM 20. None of them may get enough results to learn.

There is a second cost. When two of your ad sets reach the same people, Meta does not let them both bid for the same person. It picks one, and the others spend less than you planned. Meta calls this auction overlap. One broad ad set avoids it.

"Broad" here means only your location and a wide age range, with no interest list. Meta's system then finds the buyers inside that audience. The broad vs detailed targeting guide explains why this often beats hand-picked interests.

Where the budget sits: ABO or CBO

You can put the budget in one of two places.

ABO (ad set budget). Each ad set gets its own fixed budget. Good for testing, because a new idea cannot be starved by an older one.
CBO (campaign budget). One budget for the whole campaign. Meta moves money each day to the ad set that gets the cheapest results. In Ads Manager this is the Advantage+ campaign budget switch on the campaign (Meta's current name for CBO).

The common rule is ABO to test, CBO to scale. With a single ad set, the choice barely matters, because all the money goes to the same place. It starts to matter when you add a second ad set for proven winners. The CBO vs ABO guide covers the details.

Advantage+ or manual

Meta's Advantage+ Sales campaigns (called Advantage+ Shopping until Meta renamed them in 2025) hand more of the setup to Meta: who sees the ads, where they show, and how the budget spreads. They learn from your sales data, so they suit an account whose pixel records purchases every week.

A new shop with few sales can start manual and still keep it simple: one broad ad set, automatic placements, a budget on the ad set. Try Advantage+ once you have steady sales to feed it. The Advantage+ Sales guide walks through the setup.

How many ads in the ad set

Put 3 to 5 ads in your test ad set. More important than the number: each ad should give a different reason to buy. Two ads that differ only in colour teach you nothing. A price offer, a customer review and a "how it works" ad teach you which message sells.

If you have not picked your ideas yet, go back to Choose three angles.

A worked example: the Johor Bahru spa

Say you run a small spa in Johor Bahru, and you have RM 60 a day for a first-visit facial offer.

The structure that wastes money: three ad sets at RM 20 a day each. One for women 25 to 34 interested in beauty, one for women 35 to 44, one for a lookalike audience. Each ad set learns alone and slowly, and the three overlap.

The simple structure:

One campaign. Objective: Leads, because a booking is the result you want. Named "TEST Nov facial".
One ad set. Johor Bahru and nearby areas, ages 25 to 55, no interests. Budget: RM 60 a day on the ad set.
Four ads. A first-visit price, a short customer review, a "what happens in 60 minutes" ad, and a gift voucher idea for a partner.

After two weeks, you know which message gets bookings at the lowest cost. That winner later moves to its own scaling campaign.

A tree for the Johor Bahru spa test: one campaign named TEST Nov facial with the Leads objective, one ad set for ages 25 to 55 in Johor Bahru and nearby areas with no interests at RM 60 a day, and four ads, each giving a different reason to buy. A note below says not to split it into three ad sets at RM 20 a day each, because each one learns alone and slowly and the three overlap.

How AdPlay.ai sets this up

Launch by hand in AdPlay.ai has a Preset card with three choices:

Testing: one ad set with its own budget (ABO). On a ringgit ad account it fills in RM 50 a day, names the ad set by country (such as "Broad - MY") and starts the ads paused, so you can check them first.
Scaling: turns on Advantage Campaign Budget (CBO) and goes live at once, for proven winners.
Custom: you set everything yourself.

The Launch by hand page in AdPlay.ai with the Preset card set to Testing, and the campaign and ad set set to create new ones, starting paused

On Pro, Copilot can set up your first test instead. It makes one new campaign with one ad set: a broad audience in your country, ages 18 to 65, and the budget on the ad set. The steps are in Launch your test from AdPlay.ai.

Checklist

One campaign for the test.
One ad set, broad: location and ages only.
Budget on the ad set while I test.
3 to 5 ads, each with a different reason to buy.
No second ad set until one ad has proven it sells.

Your homework

Sketch your first test on paper. Write the campaign name and objective. Write one ad set: the location, the age range and the daily budget. Then list 3 to 5 ads, each with its reason to buy in one line.

Then, with Meta connected, open Launch by hand in AdPlay.ai (the button on this page, or Launch on any ad page). Pick your ads, click Continue and choose the Testing preset. Compare what it fills in with your sketch. You do not have to launch yet.