Do this math before you spend a dollar on Meta ads.

How much can you afford to pay per customer? If the answer isn't a number, you're not ready to run Facebook ads, because every result Meta shows you will be unreadable. A $40 cost per purchase is a win for one store and a slow bleed for another, and the only thing that separates them is a ceiling the owner set from their own margins before launch. This is the math, in the order I do it.

What you can afford to pay per customer starts at the order

Not at the budget. The ceiling is called max CPA: the most you can pay to acquire a customer and not lose money on them. It comes from two numbers you already have. Say your average order is $80, and after the product, shipping, packaging and payment fees you keep $30 of it. Then $30 is your max CPA. Pay $25 for a customer and you made $5 on the first order. Pay $35 and you lost $5, and no amount of "but the return looks fine" changes that.

Break-even ROAS is the same ceiling in the units Ads Manager speaks. Divide the order value by what you keep: $80 over $30 is about 2.7. Below 2.7 you're losing money on each order, above it you're making it. I walked through where breakeven ROAS comes from in how much should you spend on ads; the point here is that it isn't a second number. It's the same ceiling with a different label.

One thing to get right from the start: report the full order value to Meta and handle profit outside the platform. Owners sometimes feed Meta a reduced number so it "optimizes for profit." It doesn't. Underreporting disrupts the optimization, and the break-even ROAS already does the profit work for you.

Cash flow decides where under the ceiling you aim

Max CPA is the line you can't cross. Your target CPA is where you choose to sit under it, and that choice comes from cash flow, not from Meta. If cash is tight, your target is break even on day one: every customer pays for themselves on the first order, because you can't wait for the second. If you have room, you can pay more up front, sometimes right up to the ceiling, and collect the margin on the second and third order.

That second option only works if the second order is real. Paying $30 for a customer on the promise of a reorder is a plan when you have email and SMS flows bringing people back on a schedule, and a hope when you don't. So decide which business you are before you set the target, and then draw three zones around it. Say the target is $30: under about $26 is green, scale; $26 to $30 is yellow, hold; over $30 is red, cut or fix. Zones stop you from reacting to one bad day, because an ad set at $28 isn't failing. It's holding.

Chain the target down to the click

Now the target becomes useful, because you can chain it down the funnel and read Meta's columns days before a purchase shows up. The chain is plain multiplication, working backward one step at a time.

• Max cost per lead = target CPA × the share of leads who become customers
• Max cost per booked call = target CPA × close rate × show rate
• Max cost per click = max cost per lead × the landing page's conversion rate

For a store there's no lead step, so it's the target times the page's purchase rate: a $30 target on a page that converts 2% of visitors means you can't pay more than $0.60 a click. For a service business, say a $300 target, a 25% close rate and a 60% show rate. That's $45 for a booked call, and if a quarter of leads go on to book, about $11 a lead. Every one of those is a number you can hold up against a Meta column on day two.

This is why the math comes before the budget rather than after the first month. With the chain built, a $1.10 click on that 2% page is already telling you the campaign can't work at this order value, while the purchase column is still empty. Show rate matters here more than owners expect: if it slips from 70% to 50%, you need 40% more bookings to land the same number of customers, and your max cost per call drops with it. The chain catches that too.

When the chain breaks, the ads aren't the problem

Sometimes you build the chain and it tells you something uncomfortable: the clicks Meta can buy in your market cost more than the chain allows. The instinct is to fight the auction. In my experience the real constraint is usually funnel economics, not cost per acquisition.

A low-ticket offer I studied is the clean example. A $7 front-end product, decent ads by every top-of-funnel measure, and a cost per customer stuck around $60, because the funnel was one page, no order bump, no upsell, leaning on a $1,000 back end that only a small share ever bought. Nothing in the ad account could fix that. The fix was a two-step order form, a bump in the $27 to $47 range and an upsell around $297, so the spend came back faster from the people who were already buying. Improving the back end beat shaving the CPA, by a distance.

So if your max CPA comes out small and the clicks come out expensive, the math isn't telling you to bid harder. It's telling you to raise what a customer is worth before you pay to find one: a bump, a bundle, an upsell, a reorder flow. I've written about the four culprits behind unprofitable ads, and this is the first one, the one you can settle at a kitchen table before a dollar goes out. Do the math. If it fails, fix the funnel. If it holds, you have a target, three zones and a max click cost, and you can launch your Meta ads at whatever daily budget you're comfortable with, because the budget just became the least important number on the screen.

I build this chain in the first week with anyone on the media buying service, from their real order value, margin and close rate, and sometimes the chain says fix the funnel before the ads. See how the economics sit with tracking, creative and testing in the complete paid media playbook, or book a call and I'll run the numbers with you live.

Common questions

What is a good cost per customer for Facebook ads?

There's no industry number worth using. Yours is your average order value times the margin you keep, and that's the ceiling, not the target. The target sits under it, set by how long your cash can wait for a second order.

What's the difference between max CPA and break-even ROAS?

Same ceiling, two units. Max CPA is the most you can pay per customer in dollars. Break-even ROAS is order value divided by what you keep, the return at which an ad exactly covers its own cost. Use whichever one your ad manager shows.

Should I report a lower order value to Meta so it optimizes for profit?

No. Underreporting disrupts the optimization and Meta starts chasing the wrong thing. Send the full order value, know your break-even ROAS, and handle profit outside the platform.