Facebook Ads Cost in 2026: Ranges and the Smallest Test

Nobody can tell you what your Facebook ads will cost, and any agency page that quotes you a number is selling a retainer. What you can have is a starting point: the smallest budget that still produces an answer.

In 2026, most advertisers pay somewhere between fifty cents and a few dollars per click, around ten dollars per thousand people reached, and in the low twenties per lead. There is no fixed price, because Facebook is an auction: every slot in the feed goes to whoever the system scores highest at that moment. There is no rate card. Your price is whatever that auction produced in the hour you ran.

What that means for you is simpler than any agency page makes it sound. You cannot know what your ads will cost until you test, but the ranges give you a floor: about five dollars a day for one week. Then you judge it by what a customer is worth to you.

This piece covers what each result actually cost advertisers in 2026, why the same ad costs two businesses different prices, what the holidays do to the bill, and the smallest test worth running before Black Friday arrives.

What advertisers actually paid in 2026

The answer is a range, because many advertisers bid against each other every hour. Across the 2026 benchmarks, the same numbers keep turning up:

Cost per click: fifty cents to $3.50 is the span most advertisers land in, with reported averages near $1.90. Cost per thousand impressions (CPM): $8 to $15 for most audiences, though US e-commerce audiences can run higher. Cost per lead: the teens up to $50, with the low twenties typical. Cost per purchase: twenty dollars to sixty or more, and this one swings hardest by what you sell and how much one order is worth.

Treat those as reference points only. An average blends a law firm buying leads with a candle shop selling small orders. Your real number depends on your audience, your ad, your objective and the week you run it, which is exactly why the test below matters more than the table.

One thing to ignore while you are here: the minimum. Meta sets a floor for daily budgets that varies by country, objective and what you optimize for. Running on the minimum is possible. Learning anything from it usually is not, because the system needs a handful of results before it can tell who to show your ad to.

Why there is no single price

Facebook sells attention through a real-time auction. Every time a slot opens in someone's feed, Meta scores the competing ads on three things: how much you are willing to pay, how likely that person is to take the action you asked for, and how good your ad is by the reactions it earns.

This is why two businesses on the same budget can pay wildly different prices for the same audience. One wins clicks cheaply; the other pays five times as much for the identical people. The second one is usually losing on the parts you control for free: the ad itself, and whether it matches what the audience wants.

The levers that move your price, in order of how much you control them: your creative, because ads people actually engage with cost less to deliver; your objective, because an ad optimized for traffic brings clickers, and clickers rarely buy; your audience, because crowded, expensive audiences bid each other up; and the calendar, because November and December cost more than February, no matter what you do.

What the holidays do to the bill

You are reading this in October, planning holiday marketing, so the timing matters more than usual.

The auction is crowded every Q4: the number of ad slots in people's feeds is roughly fixed, and thousands of advertisers pour holiday budgets into bidding for them at once. CPM trackers have recorded Black Friday and Cyber Monday Meta costs at more than double the everyday rate, with December historically the most expensive month of the year to buy impressions.

The practical moves are unglamorous. Launch early, in late October or early November, to bank cheaper impressions before the spike. Budget for December at roughly double your October cost per result. And recompute your break-even before the expensive weeks: an ad that was profitable in October stops being profitable when the click price doubles and your prices did not move with it.

If your test lands in the middle of the spike, be fair to it: a result that looks weak on Cyber Monday might look fine on February 3rd.

The smallest test worth running

This checklist is built for an owner who has never run an ad and does not want to fund an agency's retainer to find out.

1. Start from one customer's worth. What is the profit on one order or one client? That number is your ceiling, everything else stays under it. The full napkin math is in our guide to how much to spend on marketing (https://nouna.io/blog/how-much-to-spend-on-marketing).

2. Set the budget at $5 a day for 7 days. That is the smallest spend we would run, because it reliably produces enough results for the system to learn and for you to judge. Less than that and you are only buying delivery.

3. Run one ad, one audience, one objective. Choose the objective that matches the money: sales if you sell online, leads if you book clients. One ad means one clear winner or one clear loser, not a guessing game across six versions you cannot afford to fund.

4. Pick the narrowest audience that makes sense. Everyone is expensive and vague. 'Women 30 to 45 within 15 kilometres who clicked my page before' beats 'people interested in shopping'.

5. Judge one number at the end: cost per result against your break-even. If one customer is worth $400 in profit and one in ten enquiries becomes a customer, you can pay up to $40 for an enquiry and break even. Under that, keep going. Over it, turn it off.

6. Turn it off or scale it, in that order. If the test paid: add a little more at a time. If it did not: switch it off and change one thing, the ad or the audience, never both. The most you lost is one week's budget.

7. Set a spending limit at the account level before day one. A hard cap means the worst case is known before you start.

That is the whole test. One week, one number at the end, about the price of a decent dinner out. Anyone who tells you that you need a four-figure monthly budget to 'learn the platform' is describing their own invoice.

What the ad itself costs, separate from the spend

The budget buys the clicks. Somebody still has to make the ad.

The usual routes: you write and design it yourself in an evening; a freelancer charges $800 to $2,000 a month per channel; an agency wraps the creative into a retainer that starts in the thousands per month (what each route costs in 2026 is in our guide to AI marketing agencies, https://nouna.io/blog/ai-marketing-agency-what-it-is-what-it-costs); or a subscription department makes the ads inside a plan.

This is where the smallest test gets cheaper in practice, because the creative is the lever that moves your click price more than the budget does. On Nouna's Growth plan (https://nouna.io/pricing), the ads specialist writes four new ads a month, ready to run, with a daily check on your ad spend. Nothing goes live and no dollar moves until you approve it: ads always wait for your explicit yes. If the test on this page is your first one, that is the version of it where the ads are already made and you are only choosing whether to spend. See what the department makes first, free, no card (https://nouna.io/start).