You set the spend, the auction sets the price
Meta does not sell ads at a list price. You set a budget and a goal (leads, sales, messages, traffic), and every time there is a chance to show an ad, Meta runs an auction. According to Meta's help article on the ad auction, the winner is chosen on the bid and on how well the ad is expected to perform, and you are only charged what is needed to win, which may be lower than your maximum.
What that means in practice: two businesses with the same budget can pay very different amounts per lead. The one with the more relevant ad, the clearer offer and the better landing page often pays less, because Meta expects its ad to do well.
So "how much does it cost" has two answers. The budget is whatever you set. The cost per result is something you earn, mostly through the ad and the offer.
The number that matters: budget the learning phase can use
Meta's delivery system needs data before it settles. Its learning phase help article says ad sets exit the learning phase "as soon as they can deliver stably", which usually happens after about 50 results in the week after the last significant edit. If an ad set is unlikely to get about 50 results in a week, Meta marks it Learning limited, and lists a low budget among the usual causes.
That gives you a simple floor for a budget:
- Expected cost per result × 50 = a weekly budget that lets one ad set learn.
- Divide by 7 for the daily budget.
If that number is more than you can afford, you have three options: optimise for a cheaper, more frequent result (for example, a lead form submission rather than a booked call), combine ad sets so the results pool together, or accept slower learning and judge the campaign over a longer period.
A worked example (hypothetical numbers)
A personal trainer in Leeds wants leads from Instagram and Facebook for a 12-week programme.
- She guesses a cost per lead of around £12 from what similar local campaigns have cost her before.
- £12 × 50 = £600 a week for one ad set to learn, or about £86 a day.
- That is more than she wants to spend while testing. So she runs one ad set rather than three, with one offer and a few creatives, and accepts that it may stay in learning for a while.
- After four weeks she has 120 leads at £14 each, £1,680 in total. 12 of them sign up. That is £140 per client.
If a client is worth £600 in profit, that campaign works even though the cost per lead came in higher than she hoped. Every number here is invented for illustration. Use our cost per lead calculator and CPM calculator to run your own.
What pushes the price up or down
Your cost per result moves with a handful of things. Some you control, some you do not.
- The result you choose. A link click is cheaper than a lead, which is cheaper than a purchase. Choose the result that matters to the business and budget for it.
- The creative. A tired ad that people scroll past gets more expensive over time. Fresh, specific creative is the biggest lever most small advertisers have.
- The audience size. A very narrow audience gives the system few people to try. Meta's own advice on learning limited ad sets includes expanding your audience.
- The offer. A free consultation converts at a different rate from a paid course. The offer often matters more than the targeting.
- Competition. You are in the same auction as every other advertiser trying to reach the same people, so costs move through the year as that competition changes. Plan budgets around your own past results rather than general claims.
- Your landing page or form. If people click but do not convert, every result costs more.
What to spend on a first campaign
For a first test, we usually suggest:
- one campaign with one clear objective;
- one ad set (or Advantage+ audience with sensible location and age controls);
- three to five ads with different angles, not just different colours;
- a budget that can reach around 50 results a week if possible, or a decision to judge over a longer period if not;
- a test of at least two to four weeks before major changes, because frequent edits reset learning.
Costs beyond the ad spend
The budget in Ads Manager is not the whole cost.
- Creative: photos, video and copy. Phone-shot video often works, but someone still has to plan and film it.
- Management: your time or an agency's fee. Our guide to agency pricing models explains the common structures.
- Tracking: the Meta pixel and Conversions API set up properly, so Meta can see which clicks became leads or sales.
- Follow-up: if leads wait a day for a reply, many go cold. See our guide to speed to lead.
- VAT and taxes on ad spend, depending on your country and billing setup. Check your invoices.
Signs you are paying too much
- Cost per result keeps rising week after week with the same ads: creative fatigue. Refresh the ads.
- Lots of clicks, few leads: the landing page or form is the problem, not the ad.
- Lots of leads, few customers: the form is too easy, or follow-up is too slow. Our article on why Facebook leads are poor quality covers the fixes.
- Ad sets stuck in Learning limited: budget too thin for the number of ad sets. Combine them.