Modeling Profit at a 0.8 ROAS: How Subscription Brands Outspend Everyone on Meta

A homewares-adjacent client came to us last year selling a consumable. Refillable, runs out every few weeks, the sort of product you buy once and then keep buying forever if you like it. Their media buyer had just turned off the best-spending campaign in the account because it was sitting at a 0.9 ROAS, and on a last-click basis that looked like setting money on fire.

Here's the thing. For a brand where half the buyers subscribe, a 0.9 ROAS on first order isn't a problem. It might be the whole strategy. And the buyer who paused it had no model in front of them that said so, which is the actual issue.

This is a post about building that model. Not the theory of LTV, the actual numbers you plug in so that a sub-1 ROAS reads as a plan rather than a panic. I'll build one worked example up piece by piece so you can copy the logic onto your own brand.

Start with what the first order really costs you

Forget LTV for a second. Most subscription brands never properly write down their first-order economics, and you can't model the back end until the front end is honest.

Take our consumable. Say the first order is ~$66 and the recurring order is ~$54 (the discount is the hook to subscribe). Cost of goods on a refill is ~$18, and there's ~$9 of pick-pack-ship per order because last-mile shipping is brutal on small consumables. So before a single ad dollar, your contribution margin on that first order is roughly $66 minus $27, call it ~$39.

That ~$39 is your real budget for acquiring the customer if you want to break even on day one. Spend more than that to get the sale and you are, by definition, underwater on the first order. Which sounds alarming until you remember the whole point of the model is that you're meant to be underwater on the first order.

To put that in perspective: at ~$39 of first-order margin against, say, a ~$48 cost to acquire, you're losing ~$9 on every new customer the day they buy. A last-click ROAS would show that as roughly 0.8 ($66 revenue over ~$84 of spend-plus-cost, near enough). And a media buyer staring at a 0.8 with no further context will kill it every time.

Why 0.8 can be the right number to spend at

The reason a brand can rationally spend down to a 0.8, or lower, is that the first order isn't the product. The relationship is.

If a meaningful share of those buyers subscribe and stick around, each customer is worth several orders, not one. So the question stops being "did this order pay for itself" and becomes "how long until the customer pays back what I spent, and how much do they throw off after that". That's payback period and lifetime value, and they're the only two numbers that make a sub-1 ROAS safe.

I'll say the quiet part plainly: a 0.8 ROAS is only sane if you can fund the gap. You're choosing to lose money on the first order in exchange for a bigger number later, and "later" means cash leaves the building before it comes back. If you can't carry that, the model doesn't care how good your LTV is. More on that further down, because the cash trap is where most brands actually come unstuck, not the maths.

The churn number you have to plan around

Here's where founders get optimistic and the model quietly breaks. Consumer subscription churn is nothing like software. Nobody cancels their work tools every month. People cancel consumables constantly, because taste changes, the cupboard's full, the card expires, or they just lose interest.

Plan for ~10-15% monthly churn. That's not pessimism, that's the normal range for consumer subscription, and at the top of it you can lose effectively your whole cohort inside a year. If you model on 5% because it flatters the spreadsheet, every downstream number is fiction.

Let me run our example at 12% monthly churn, which is middle of the range.

A customer who pays ~$27 of margin per recurring order (that's $54 minus the same ~$27 of cost) and churns at 12% a month sticks around, on average, about 8 months. Not 8 orders necessarily, depending on cadence, but say the refill ships monthly so it's roughly 8 recurring orders plus the first one.

So lifetime contribution is roughly: ~$39 on the first order, then ~8 recurring orders at ~$27, which is ~$216. Total LTV in contribution terms, call it ~$255 per acquired customer who subscribes.

Now the 0.8 looks different. You spent ~$48 to acquire someone worth ~$255 over their life. That's not reckless. That's a ~5x return on the acquisition cost, it just takes months to land rather than showing up on day one.

Work out your payback window, then breathe

The single most useful number to put in front of a nervous media buyer isn't ROAS or LTV. It's the payback window: how many months until a customer has returned what you paid to get them.

In our example you're ~$9 down after the first order. Each month after that, a retained customer returns ~$27 of margin. So you claw back the ~$9 gap inside the first recurring order, and you're in genuine profit from roughly month two onward.

That's a fast payback, and it's fast because the recurring margin is healthy relative to the acquisition cost. If your numbers are tighter, say you're spending ~$70 to acquire and only making ~$20 a month back, your payback stretches toward four or five months, and that's a very different cash conversation even if the eventual LTV is identical.

I believe payback window is the number that should sit at the top of a subscription brand's dashboard, above ROAS. It tells you how long your money is tied up, which is the thing that actually kills subscription brands. A 3-month payback at a 0.8 ROAS is a comfortable business. A 9-month payback at the same 0.8 will have you raising money or stalling out, no matter how pretty the LTV looks.

The blended-CAC trap nobody warns you about

This one's subtle and it's cost brands a fortune, so stay with me.

Almost every subscription brand runs a single landing page where you can either buy once or subscribe. The ads drive to that page, some people take the one-time option and some subscribe. At the end of the month you take total ad spend, divide by total new customers, and call that your CAC.

That blended CAC is lying to you. It quietly averages two completely different customers together.

Say you spend ~$100 and get two new customers off it: one one-time buyer, one subscriber. The blended maths says ~$50 CAC each, clean and tidy. But it's entirely possible you actually paid ~$70 to land that subscriber and ~$30 for the one-timer, because subscribers are the harder, more considered sale. You can't see it, because both came through the same page, so all you get is the average.

Why does that matter? Because your whole model rests on the subscriber being worth ~$255. If their true CAC is ~$70 rather than the ~$50 you assumed, your payback window and your real return both move against you, and you'd never know from the blended number. Meanwhile the one-time buyer you've been ignoring might be quietly profitable on day one.

The fix isn't a fancy tool. It's to play with the offer and watch what moves. Push the subscribe discount, or A/B the page with and without the one-time option, and see how your volume and blended CAC shift. The direction of the change tells you roughly what each customer type really costs. You're never going to get it to the decimal, but knowing it's ~$70 and not ~$50 changes how hard you can lean on the spend.

How to brief the media buyer so they don't panic

All of this falls apart if the person in the account is still being judged on last-click ROAS. They'll keep switching off your best campaigns, exactly like the buyer at the top of this post.

So the brief has to change. Hand them three things: the ROAS floor you're comfortable spending down to (the 0.8, with the reason written next to it), the payback window you're protecting, and the rule that a campaign hitting the floor with healthy new-customer volume does not get touched. Then judge them on cost to acquire a new subscriber and blended new-customer numbers, not on the platform's first-order return.

The mental shift is the hard part. A 0.8 isn't a failing 0.8 here. It's a deliberate one, funded on purpose, paying back in two months. Without that context written down, every good month gets sabotaged by someone doing what looks like their job.

If you want a second read on whether your own sub-1 ROAS is a strategy or a slow leak, that's exactly the kind of thing a Signal/Noise Audit pulls apart. We'd build the payback model against your real first-order margin and churn and tell you straight which campaigns you can safely push and which ones are quietly burning cash. No pitch, just the numbers laid out so the next person in the account stops panicking at a 0.8.