How to Advertise on Facebook: A Step-by-Step Guide
Learn how to advertise on Facebook step by step: objectives, audiences, budget, pixel setup, learning phase, review times and fixes for common problems.
ADS Beast editorial teamPublished 10 min read
To advertise on Facebook, create a business account and ad account, install the pixel, pick one of the six campaign objectives, define your audience and budget, then build the ad and launch it. Review usually finishes within 24 hours. After launch, leave the campaign alone for about a week so the learning phase can stabilize.
In short:
- Facebook ads run through Ads Manager on top of a business account, an ad account and a payment method.
- Meta offers six campaign objectives: awareness, traffic, engagement, leads, app promotion, and sales.
- Tracking is split between the pixel and the Conversions API, both managed in Events Manager.
- The learning phase needs roughly 50 results per week after your last significant edit.
- Most ads pass review within 24 hours; repeated edits restart the learning clock.
What you need before you run your first Facebook ad
You need three things in place: a business presence, a payment method, and tracking. Everything else in Ads Manager is configuration on top of those.
Start with a Facebook business page for the brand. Ads are published from a page, not from a personal profile, and the page name and image appear in the ad itself. Then create a business account in Meta Business Suite if you do not have one, and add an ad account inside it. Assign yourself admin access to both.
Add a payment method to the ad account. Meta does not publish a fixed dollar minimum for daily budgets. The floor depends on your country, currency, campaign objective and payment method, and Ads Manager will warn you if your number sits below what it accepts for your setup.
Install tracking before you spend anything. The pixel goes on your website, and the Conversions API connects your server to Meta. Both live in Events Manager, and both feed the same optimization system. Without them, Meta has nothing to optimize toward beyond clicks.
If you also buy search traffic, the same event data supports both channels, and the two reinforce each other. Our breakdown of Search engine optimization Google Ads covers how that overlap works in practice.
How to set up a Facebook ad campaign step by step
The campaign structure has three levels: campaign, ad set, ad. The campaign holds the objective, the ad set holds audience, budget, placement and schedule, and the ad holds the creative. Build them in that order.
- Open Ads Manager and create a campaign. Choose the objective that matches the outcome you actually track.
- Name the campaign so you can find it later. Use a convention like objective, audience, and date.
- Set the buying type and whether you want A/B testing on. This is the last point where the objective is cheap to change.
- Build the ad set. Define the audience, placements, budget, and schedule here.
- Set the budget at the ad set level unless you need campaign budget optimization across several ad sets.
- Choose the conversion event you want to optimize for. It must exist in your tracking already.
- Build the ad. Upload creative, write the primary text, headline and description, and set the destination.
- Check the preview on each placement you selected. Square creative and vertical creative behave differently.
- Submit for review. Do not edit while it is pending unless something is genuinely wrong.
The most common early mistake is choosing the objective by habit rather than by what you measure. If you want purchases, use the objective that matches purchases and make sure the purchase event fires correctly. The old conversions and messages objectives are gone, so map your goal onto one of the six that exist today.
Which campaign objective should you choose?
Choose the objective that matches the outcome you can measure after the click. Meta offers six: awareness, traffic, engagement, leads, app promotion, and sales. The objective tells the delivery system what to look for, so a mismatch between objective and tracked event wastes budget.
| Objective | What it optimizes for | Best when |
|---|---|---|
| Awareness | Reach and impressions | You need visibility, not clicks |
| Traffic | Link clicks and landing page views | You want visitors and have on-site tracking |
| Engagement | Reactions, comments, shares, video views | You are building an audience or testing creative |
| Leads | Lead form submissions or on-site lead events | You collect contact details |
| App promotion | Installs and in-app events | You have a mobile app with SDK events |
| Sales | Purchases and other conversion events | You sell online and track purchases |
Two of these deserve a warning. Leads works well with native forms because the form lives inside Facebook, but the quality of those leads depends on what you ask for. Sales depends entirely on your event data being accurate; if the purchase event double-fires or never fires, the algorithm optimizes toward noise.
If your goal is conversations in a messaging app rather than a form or a checkout, the objective you pick still has to map to something you can count. WhatsApp messaging now bills per delivered template message under the WhatsApp Business Platform pricing model, which changed on 1 July 2025, so message-based campaigns need their own cost math before you scale them.
How to define your audience and budget
Define the audience first, then size the budget to produce enough results for the learning phase. Meta needs roughly 50 results per week after your last significant edit, so an audience too small for that volume will keep the campaign unstable no matter how good the creative is.
Start broad and let the delivery system work. Narrow interest stacks were more useful years ago; today the algorithm finds buyers faster when it has room to explore. Use custom audiences from your pixel, your customer list, or engagement on your page when you want to reach people who already know you. Use lookalikes when you have a clean source audience large enough to build from.
Exclusions matter more than most advertisers expect. Exclude recent purchasers from prospecting campaigns, and exclude your existing customer list if you are paying for new customers only. Without exclusions you pay to re-acquire people who already bought.
For budget, work backwards. If you know your cost per result from past campaigns, multiply it by the number of results you want in a week. If you do not know it yet, pick a daily number you can sustain for at least a week without flinching, because stopping and restarting resets learning. The right floor is the one that produces roughly 50 results weekly, not the smallest number the platform accepts.
Where to check the ad account before spending is a separate question from where to set the budget. If you want the campaign structure, audience logic and budget controls assembled for you before you touch Ads Manager, AI campaign setup is where that process lives.
Where the Facebook pixel and Conversions API fit
Both live in Events Manager, and they exist to feed conversion data back to Meta. The pixel fires from the browser. The Conversions API fires from your server. When both send the same event, Meta counts it once.
Deduplication works by matching event name and event_id. There is no toggle to switch it on. If the same purchase fires from the browser and from your server with the same event_id, Meta treats it as one event. If the IDs differ, you get two purchases in reporting and the algorithm learns from inflated numbers.
Set up the pixel first, verify that the standard events fire on the right pages, then add the server-side connection. Test with the event testing tools in Events Manager before you spend. A campaign optimizing toward a broken event will spend your budget and teach you nothing.
How long review takes and what to do while you wait
Most ads pass review within 24 hours, though some take longer. If an ad sits in review past that window, check your account status and the ad itself instead of editing it repeatedly.
Review is a policy check, not a performance check. Common reasons an ad stalls are restricted content in the creative, a landing page that does not match the ad, or an account-level issue. Fix the actual problem once, then resubmit.
Editing is the trap. Every significant change restarts the learning phase clock, so a nervous advertiser who tweaks the headline twice a day never lets the campaign exit learning. Batch your changes. Decide on a review point, look at the data, make one round of edits, then wait again.
Why performance drops after you edit an ad
Any significant edit resets the learning phase, and Meta needs about 50 results in a week before delivery stabilizes again. That is the whole explanation for most "my ads died after I touched them" stories.
The second cause is audience size. If your audience cannot produce 50 results per week at your budget, the campaign never stabilizes regardless of edits. Check whether the constraint is the edit or the volume.
The third cause is creative fatigue, which looks similar but behaves differently. Fatigue shows a gradual decline over weeks, not a cliff after an edit. If the drop is instant and follows a change you made, it is the learning phase. If it is slow and steady, refresh the creative.
Common mistakes and how to spot them early
Most wasted Facebook ad budget comes from four mistakes, and each has an early warning sign.
Optimizing for the wrong event. If your reported conversions look implausible next to your actual sales, your event is firing wrong or firing twice. Compare Events Manager numbers against your own order data before scaling.
Editing too often. If your campaign has been running for weeks and still shows unstable delivery, count your edits. Frequent small changes keep it in learning permanently.
Audience too narrow. If reach is tiny and frequency climbs fast, the audience cannot support your budget. Widen it or lower the budget to match the volume available.
Ignoring placement differences. If one placement eats most of the spend and produces nothing, check the breakdown rather than guessing. Creative that works in a feed often fails in Stories, and the reverse is also true.
The same discipline applies across platforms. If you also run video campaigns, the management approach for YouTube Ads Management: Order Campaign Management follows similar rules about learning periods and edit discipline, and it is worth reading if you split budget across channels.
Where automation fits without taking control away
Automation helps most with the mechanical parts: structure, budget pacing, policy checks before submission, and reporting. It helps least with judgment calls about positioning and offer.
Tools that generate campaign structure from your inputs, flag policy problems before you submit, and track CPA, CTR, CPC and CPM across channels remove the repetitive work. What they should not do is spend without your confirmation. If a system changes budgets or launches campaigns on its own, you lose the ability to explain what happened when results move.
This is the same principle behind Programmatic AI: How Automated Ad Buying Works, where the value sits in the decision support rather than in removing the human from the loop. The same logic shows up in creative tooling, which is why pricing for AdCreative AI Pricing: Plans, Packages & Business Costs is worth understanding before you commit to a subscription.
Advertising is spreading beyond the classic platforms, and each new surface has its own rules. If you are curious how paid placements work in assistant interfaces, How to Advertise on ChatGPT: Setup, Targeting, Tracking explains the mechanics, including the fact that ChatGPT ads bill per impression or click with no keyword bidding.
Your next step
Open Events Manager and confirm that your pixel and Conversions API events fire correctly with matching event names and event_ids. Nothing else you do in Ads Manager matters until that data is clean, because every objective depends on it. Once tracking is verified, build one campaign with one objective, one audience and one budget you can hold for a week.