Since starting Romi Digital in 2024 we've worked with nearly 50 e-commerce companies and spoken to hundreds of founders. The single most common pattern we see isn't bad creative or a weak website. It's that the fundamentals of the business never made sense, and nobody checked before the first purchase order was placed.
The usual sequence looks like this:
- Get an idea for a product
- Build a prototype
- Buy initial stock and decide a retail price
- Begin marketing Every step in that sequence is about the product. But manufacturing a good product has never been easier, third-party fulfillment has commoditised logistics, and everyone is on the same store platform. For most e-commerce companies, the only real differentiator left is marketing and brand.
Which means marketing isn't step four. It's the constraint that should have determined steps three and two.
The calculation
You're launching a vegan protein powder. It retails at $65 with free shipping.
Founders usually stop at the manufacturing cost, so let's not. Here's what it actually costs to put one tub in a customer's hands:
| Line | Cost |
|---|---|
| Manufacturing (powder, tub, label) | $18.00 |
| Inbound freight and duty | $2.00 |
| Pick, pack and fulfillment fee | $4.00 |
| Outbound shipping | $9.00 |
| Payment processing | $2.21 |
| Discount allowance (10% welcome code) | $6.50 |
| Returns and damages allowance | $1.95 |
| Total landed cost per order | $43.66 |
That leaves $21.34. This is your contribution margin — the money available to acquire the customer. Not your net profit. Net profit is what's left after you've also paid for creative production, software, your agency, your salary and everything else that doesn't scale with unit volume.
So: $21.34 to acquire a customer.
In 2025, the median Meta CPA across Triple Whale's dataset of 30,000+ brands was $38.19. In Health & Wellness specifically — your category — it was $38.55, the sharpest year-over-year increase of any vertical at +12.64%.
You lose roughly $17 on every new customer you acquire.
Put differently: your break-even ROAS is 1 ÷ 0.328, or 3.05x. The median Meta ROAS across that same dataset was 1.86x. You need to be substantially better than the median advertiser on day one, in a category where acquisition costs are rising faster than anywhere else, before you make a single dollar.
This is the calculation. It takes ten minutes and it should happen before you commit to a formulation, a price or a purchase order.
"But I'll have repeat customers"
You probably will. Protein powder is a consumable in one of the highest-repeat categories in e-commerce. That's a genuine argument, and dismissing it entirely would be wrong.
Here's the problem with relying on it: on launch day, you have no data. You have an assumption. And the assumption is load-bearing for the entire business.
Run the second order through the model. A returning customer costs you almost nothing to reach — email and SMS are effectively free at the margin. So order two contributes another $21.34 against the $38.55 you already spent. Two orders per customer, and you're finally $4 ahead. That $4 is what pays for your creative, your software stack, your agency and yourself.
That is what "we'll make it up on repeat purchase" actually means when you write it down. Not comfortable profitability at order two — barely break-even, before overhead.
You can and should underwrite a repeat rate before you launch. Use published category benchmarks, look at comparable brands' subscription offers, talk to a manufacturer about how their other clients perform. Then take that number and cut it, because your brand has no reputation and no reason to be trusted yet. If the business only works at the optimistic end of that range, you don't have a business, you have a bet.
AOV is the number you can estimate before you launch
You can't know your repeat rate in advance. You can get remarkably close on Average Order Value.
If your AOV is under $30, you have picked an extremely difficult business to start. The 2025 median AOV across Triple Whale brands on paid channels was $74.12. A $30 AOV means competing in the same ad auction, at the same prices, with 40% of the margin dollars the median brand has to work with.
You can estimate AOV before you have a single order, with honest reasoning:
- What does your hero product cost, and what proportion of orders will be hero-product-only? For a new brand with no catalogue awareness, assume most of them.
- How many units per order? Be pessimistic. Assume 1.0 to 1.2 unless you have a specific reason not to.
- What's your bundle attach rate, realistically, when nobody knows your brand yet?
- Where's your free shipping threshold, and does it actually pull orders up or just erode margin on orders that would have converted anyway? One important caveat, because the naive version of this advice is wrong: raising prices does not hand you a free lunch. Higher-priced products carry longer consideration cycles and lower conversion rates, so CAC tends to rise with AOV. What a higher AOV buys you is more contribution margin dollars per order — more room for error — not a cheaper customer.
The one that actually kills companies: payback period
Suppose the model works. Two orders per customer, marginally profitable, everything checks out on the spreadsheet.
The spreadsheet has no timestamp. Reality does.
You pay the $38.55 CAC today. The first order returns $21.34 today. The second order might arrive in month four. So every new customer you acquire opens a cash hole that closes 120 days later.
Now scale. At $10,000 a month in ad spend with a 120-day payback, you need roughly four months of spend — $40,000 — permanently tied up in customers who haven't paid you back yet. Double your spend to grow and that requirement doubles too. Growth consumes cash faster than it generates it.
This is why brands that are profitable on a per-customer basis still fail. They didn't run out of demand. They ran out of working capital.
Two numbers matter here, and most founders track neither: how long until a customer repays their acquisition cost, and how much cash that gap ties up at your current spend.
Do it in the other order
The fix is to invert the sequence you started with. Instead of product first and marketing last:
- Start with your category's CAC benchmark. Not the number you hope for — the published median for your vertical.
- Decide how much better than median you can credibly be. At launch, with no creative library, no data and no brand: not much. Assume median.
- Solve backwards for the contribution margin you need. If you want to be profitable on the first order, contribution margin per order must exceed that CAC. If you're willing to be profitable by order two, it needs to exceed roughly half of it — and you need the capital to fund the gap.
- Only now choose your price and your COGS target. If paid media is your primary acquisition channel, you're generally looking for 70%+ gross margin on product. That's not a law, but if you're materially below it, paid social is going to be very difficult.
- Then build the product. If you're already launched and the numbers don't work, the levers are the same ones, in rough order of how quickly they move: raise prices, restructure bundles to lift units per order, put a subscription offer in front of every new customer, cut the discount that's costing you 10% of revenue, renegotiate fulfillment, and get serious about email and SMS so order two doesn't need paid media at all.
None of it is exotic. The failure isn't that founders can't do this arithmetic — it's that they do it after the stock has landed, when the only variable left to change is the ad account.
Do it first. Ten minutes, before you commit a cent.