Quick answer. A Facebook campaign has three levels: campaign (objective), ad set (audience, budget, placements), and ad (the creative). Start with one campaign on your real objective, one or two ad sets, and three to five ads in each. Use campaign budget optimization so Meta spreads spend to the winners, and give it enough budget to exit the learning phase. Consolidate instead of fragmenting.

When I train new buyers on Meta, the first mistake is almost always structure. They build twelve ad sets, drop one ad in each, split the budget into tiny pieces, and then wonder why nothing learns. The platform never gets enough signal to do its job.

Structure is the part that decides whether the algorithm can work for you. Let me walk you through how Meta campaigns are actually built and give you a clean setup you can copy on day one.

What are the three levels of a Meta campaign?

Every Facebook or Instagram campaign is built in three layers, and each layer controls something different. Once you see what lives where, the whole tool makes sense.

  • Campaign: this is where you pick your objective, the goal you want Meta to chase. You also turn on campaign budget here if you want it.
  • Ad set: this is where you set the audience, the budget and schedule (if not set at the campaign level), the placements, and the optimization event like purchase or lead.
  • Ad: this is the creative itself, the image or video, the headline, the primary text, and the link people click.

Think of it top to bottom: the campaign says what you want, the ad set says who and how much, the ad says what they see. Most beginner confusion comes from setting something at the wrong level, so keep that map in your head.

How do you choose the right objective?

The objective tells Meta who to show your ads to, so it matters more than almost any other setting. Pick the objective that matches the real result you want, not the one that looks busy.

If you sell something online, choose the sales objective and optimize for purchases. If you collect leads, choose the leads objective. If you genuinely just want people on a page, traffic works, but be honest, most beginners pick traffic because it feels cheaper and then get clicks that never buy.

Optimize for the action closest to the money you can. Meta gets better at finding buyers when you ask it for buyers. Asking for clicks and hoping they convert is how budgets get wasted.

Should you use Advantage+ or a manual campaign?

Meta now pushes its automated setups hard. Advantage+ Shopping or Sales campaigns hand most of the targeting and placement decisions to the algorithm, you mainly bring the budget and the creative. Manual campaigns let you control the audience, placements, and structure yourself.

For a beginner with a real product and a working pixel, Advantage+ is often the simpler starting point because it removes choices you are not ready to make well.

Reach for a manual campaign when you have a specific audience you need to control or a clear hypothesis you want to test. A practical move is to run an Advantage+ campaign as your main engine and a small manual campaign on the side to test your own ideas.

CBO or ABO: where should the budget live?

You can set your budget in two places, and people overthink this. Campaign budget optimization (Meta now calls it Advantage campaign budget, or CBO) puts one budget at the campaign level and lets Meta move money to whichever ad set is performing. Ad set budget optimization (ABO) puts a fixed budget on each ad set so you control the split.

CBO is usually the better default. It keeps spend flowing to the winners automatically and helps each ad set gather enough events to learn. It also stops you from starving a good ad set or feeding a bad one out of habit.

ABO makes sense when you are testing and want to guarantee every ad set gets a fair, equal amount of spend, for example when you are comparing two audiences and refuse to let one get throttled early. Once you know what wins, move it into a CBO campaign and let the algorithm scale it.

How many ad sets and ads should you run?

This is where most beginners sabotage themselves. They think more ad sets means more chances to win. In reality, each ad set needs roughly 50 optimization events a week to exit the learning phase, and splitting your budget across many ad sets means none of them ever get there.

Consolidate. Fewer ad sets with more budget each will outperform a pile of tiny ones almost every time. The algorithm learns faster when the signal is concentrated, not scattered.

Here is a clean structure you can copy:

  • One campaign, sales objective, with Advantage campaign budget turned on at, say, 50 dollars a day.
  • One or two ad sets: one broad audience and, if you want, one tighter audience to compare. No more for now.
  • Three to five ads per ad set, different creatives and angles, so Meta has variety to choose from without spreading budget across structure.

That gives the algorithm enough room to find people and enough budget to actually learn. When something wins, raise the budget slowly, around 20 percent every few days, so you do not throw the campaign back into learning.

Key takeaways

  • A campaign has three levels: objective at the campaign, audience and budget at the ad set, creative at the ad.
  • Consolidate into one or two ad sets with several ads each so the algorithm gets enough signal to learn.
  • Use campaign budget optimization and give it enough budget to clear the learning phase, then scale winners slowly.

Frequently asked questions

What is the learning phase and why does it matter?
The learning phase is the period when Meta is still figuring out who to show your ads to, so performance is unstable and costs swing around. Each ad set needs about 50 optimization events in a week to exit it. If your budget is split too thin or you keep editing things, you never leave learning. Concentrating budget and leaving the campaign alone are the fastest ways out.
How much budget do I need to start?
Enough for your chosen optimization event to reach roughly 50 conversions a week. If a purchase costs you 10 dollars, that is about 70 to 80 dollars a day to give the ad set a real chance to learn. If your product is pricier, optimize for an earlier event like add to cart while you build up data. Work backward from your cost per result, not from a figure you saw online.
Should I duplicate winning ad sets to scale?
Usually no, especially early. Duplicating resets learning and the copies often compete against each other in the same auction. A cleaner approach is to raise the budget on the winner gradually, around 20 percent every two or three days, so you scale without shocking the algorithm. Save duplication for later, once you understand how your account behaves and have a reason to test a fresh setup.
How often should I edit a running campaign?
As little as possible while it is learning. Significant edits to budget, audience, or optimization event can restart the learning phase and waste the data you already paid for. Give a new campaign at least a few days and a meaningful number of conversions before you judge it. Check it daily to make sure nothing is broken, but resist the urge to tweak something every time you look.