How much do Meta ads cost? Real numbers, not "it depends"
Updated July 2026
Meta ads have no fixed price, and no reliable price bands either. They are sold in a live auction, so the cost per result depends on who else is bidding, the season, your creative, and how well the system can find your buyers. Anyone quoting you an average CPM is guessing. A high CPM that converts well is cheap; a low CPM that never converts is expensive.
And nobody actually buys on CPM. You buy on cost per action and return. The one cost that decides whether Meta is worth it is your break-even CPA: what you can pay to acquire a customer and still profit, which is your order value times your margin. A €60 order at 40% margin breaks even at a €24 CPA. Everything on this page is about finding, and beating, that number.
The real number is your break-even, not Meta's price
Most "how much do Meta ads cost" answers quote a CPM and stop there. That is the wrong number. A CPM tells you what attention costs; it says nothing about whether your business can afford it. Two shops paying the exact same €10 CPM can have wildly different outcomes, because one keeps 60% of each sale and the other keeps 12%.
Your real cost ceiling is set by your margin, not by Meta. The formula is simple: your break-even ROAS is 1 divided by your gross margin. A 50% margin means you break even at 2x return. A 17% margin means you need almost 6x just to stand still. That single relationship explains why identical ad costs are comfortable for one brand and fatal for another.
It also explains a pattern you see over and over: high-margin businesses thrive on Meta, thin-margin businesses struggle on the exact same auction. Software, games, and luxury do consistently well, and often find attention relatively cheaply, because they can afford to outbid for the same buyer and still profit. If your margin is slim, your problem is rarely the ad itself. It is the maths underneath it.
What real brands need to break even
To make that concrete, here are the break-even numbers behind brands you know, taken from our ad plans library. They are estimates modelled from public pricing and margins, but they show how far the break-even point moves with the business model.
| Brand | Why | Break-even |
|---|---|---|
| Suitsupply | High-AOV menswear, fat margin | 1.6× |
| Rituals | Beauty, high repeat, high margin | ~1.8× |
| Daily Paper | Streetwear, healthy margin | 1.8× |
| Zalando | Fashion marketplace, blended margin | 3.4× |
| Coolblue | Electronics, very thin margin | ~6× |
Read as break-even ROAS: revenue per euro of ad spend just to cover product cost. Below the number is a loss; the target for profit sits above it. Estimates, not the brands' own figures. See the full library for how each plan is built.
The takeaway is the whole point of this page: Coolblue and Suitsupply pay the same auction, but Coolblue needs nearly four times the return to survive. When someone asks whether Meta ads are "expensive," the honest answer is: expensive relative to what your margin can carry.
Work out your own number
Enter your average order value and gross margin. This is the same math the campaign planner runs, and it gives you the three numbers that actually govern your cost.
Break-even calculator
Your cost ceiling, in ten seconds.
Break-even CPA is order value times margin. Break-even ROAS is one divided by margin. Test budget is 50 conversions times a target CPA set 20% below break-even, the spend it takes to get a real read. Run your full plan →
Aim to run below your break-even CPA, not at it. The gap between break-even and your actual CPA is your profit.
What actually moves your cost
Once you know your break-even, these are the levers that decide whether your real cost lands above or below it. In rough order of impact:
- Your margin. Covered above, and it dwarfs everything else. It sets the ceiling before a single ad runs.
- Creative. The single biggest in-account lever. A hook people stop for lowers CPM and CPC at the same time. Creative fatigue (frequency climbing past 6 to 7, click rate falling) is the most common reason costs quietly rise.
- Signal quality, and the iOS 14 effect. When Apple's App Tracking Transparency cut off a chunk of the data flowing from the apps back to Ads Manager, the system got fewer signals to find your buyers. Costs rose and conversion rates fell across the board, because the bidder was working from a blurrier picture. The fix is feeding the signal back yourself: optimise to Purchase with clean server-side tracking (the Conversions API). Weak signal, or a broken pixel, quietly raises your real cost per result.
- Budget versus the learning phase. Meta needs about 50 conversions a week to exit learning. Under that floor, cost stays high because the system never gets enough data to optimise.
- Audience and competition. Broad targeting on a direct-response objective usually beats narrow interest stacks. Seasonality (Q4, promotions in your category) raises the auction price for everyone.
- Which market you compete in. Cost follows demand for attention, not a fixed regional rate. The US has the most advertisers fighting over the same feeds, so it carries the highest CPMs. Smaller markets like the Netherlands look cheaper but are crowded with local competitors bidding against each other. Genuinely cheap markets exist, but usually because there are fewer buyers worth reaching. Automatic placements let the algorithm buy the cheapest attention that still converts.
Notice what is not on that list: a magic budget number. There is no spend level that makes bad economics good. The budget question only matters relative to the learning phase, which is the next question people ask.
From experienceThe most expensive mistake I saw up close was over-segmentation. A global food-delivery marketplace we worked with had split their account by age, by gender, by location, even by time of day. Dozens of ad sets, each running on a sliver of budget. It felt like control. It was the opposite. The system never got enough signal to leave the learning phase, and they struggled to spend the budget at all, even though it was a lot of money.
So we ran a clean test. Two comparable countries, same brand awareness, similar engagement. One we left exactly as it was. The other we consolidated to about three ad sets, six strong creatives each, and let the targeting go broad. We ran them at the same time.
The consolidated country started slow, CPMs high while the algorithm worked out who it was talking to. Then it settled. Running that broadly surfaced whole customer types they did not know were buying from them, the real audience was far bigger than they assumed. CPMs came down, click rate went up, and the consolidated setup won. They rolled consolidation out across every market.
The cost lesson: fragmenting a budget to feel in control usually raises your cost per result, because you starve the algorithm of the signal it needs to get cheap. Give it room, feed it strong creative, and it finds cheaper attention than you can hand-pick.
Common questions
How much do Facebook ads cost per day?
There is no daily minimum that makes ads work, only a minimum that lets the algorithm learn. To exit the learning phase in a week Meta needs roughly 50 conversions of your chosen event. So the honest daily floor is your target cost per result times 50, divided by 7. At a €30 target CPA that is about €214 a day. You can spend less, but the campaign never gathers enough data to optimise, so it stays expensive.
Is €10 a day enough for Facebook ads?
For a purchase objective, usually not. €10 a day cannot generate the 50 conversions a week the algorithm needs, so it stalls in the learning phase and costs stay high. Where €10 a day can work: a cheaper event like a lead or an add to cart, or pure retargeting of a warm audience that is already primed to buy. For optimising to sales on cold traffic, it is below the floor.
Why are my Meta ads suddenly so expensive?
A sudden rise is almost always one of four things: creative fatigue (frequency climbing, click rate falling), budget below the learning-phase floor, weak conversion signal so the bidder optimises on a blurry picture, or a thin margin that makes any cost feel expensive. The auction price is the symptom. Find which of the four is the cause before you touch the budget. Our diagnostic tool ranks them for your situation.
What is a good ROAS for Meta ads?
A good ROAS is any number above your break-even ROAS, which is one divided by your gross margin. On a 17% margin you need roughly 6x just to break even, so 4x is a loss. On a 60% margin you break even at 1.7x, so 2x is already profit. There is no universal good ROAS. Use the calculator above to find yours.
What is the average CPM on Facebook or Meta?
There is no useful average, and that is not a dodge. CPM is set live in an auction, so it swings with who else is bidding, the season, your creative, and your market. A published "average CPM" is someone else's account on a different day. More to the point, CPM is not what you are buying. A high CPM with a high conversion rate is cheap; a low CPM that never converts is expensive. Judge cost on cost per action against your break-even, not on CPM.
How much should I budget to test properly?
Enough to reach about 50 conversions, which is what it takes to get a read that is not noise. That is roughly your target CPA times 50. At a €25 target CPA, plan for around €1,250 before drawing conclusions. Spending a little for a few days and quitting tells you nothing, because the algorithm was still learning the whole time.
I led growth at Meta and Snap, working with brands like Booking.com, Disney, and Takeaway.com. paid.social is the free tool suite I built out of that experience. More about this →
Method: Meta ad cost is set live in an auction, so this page deliberately does not quote CPM or CPC "averages," which do not carry between accounts. Brand break-even figures are estimates modelled from public pricing and margins, detailed in the ad plans library, not the brands' own data. The learning-phase figure (about 50 conversions to stabilise) is Meta's published guidance. Break-even and budget math is deterministic, the same logic the planner runs.