It always starts with a good month. ROAS beats the target it has missed for 3 months, revenue crosses a milestone, and the obvious move is to triple the budget.
Here is what the dashboard says. You spend £1 on ads and it reports £3.80 back. Say you keep 28% of each sale after product, fulfilment, returns and payment fees:
Pay back the £1 of ads and 6p is left. Staff, rent and software still have to come out of that.
Now triple the budget. The new spend reaches colder audiences and ROAS slips. Say it falls to 3.0, which is 84p kept per £1:
Every £1 of ads now costs 16p. 6 months later the revenue is real and the bank account is not. The unit economics never supported the scale, and the ROAS in the dashboard never showed it. In our client work, this pattern can cost a brand 6 figures in a year. The campaigns ran as they should. Nobody had checked the foundation.
Why a good ROAS feels like proof
“I spend £1 and get £3.80 back” is immediate, visible and easy to believe. It does not say whether the £3.80 covers product, fulfilment, returns and payment processing, whether 3.8x holds at 3 times the budget, or whether this month’s customers ever come back.
These brands read the number the platform shows and act on it sensibly. The number leaves out their costs.
3 calculations before any scale decision
1. Contribution margin per order. Take your average order value and subtract:
- COGS: the landed cost of the product, including inbound freight
- Fulfilment: pick, pack and ship, including return handling, which most brands undercount
- Payment processing: typically 2 to 3% of revenue
- Returns, as a blended rate: if 18% of orders come back, apply that to the margin
For a healthy DTC brand running paid media, the result sits at 40 to 55% of AOV. Below 35%, scaling paid acquisition rarely turns a profit at any realistic ROAS. Below 30%, the model likely does not work without a change to pricing or COGS.
2. Blended new-customer CAC. Take total ad spend for the month: every channel, every agency, every fee. Divide by new customers, not total orders. Returning buyers convert cheaply and make the average look better than it is, so this figure is almost always higher than the cost per purchase in the dashboard.
3. The comparison. Say your contribution margin per order is £24 and your blended new-customer CAC is £30:
If contribution is above CAC, every new customer covers their acquisition cost and scaling is supportable. If CAC is higher, as here, scaling speeds up the loss.
3 things to fix first
Product margin. Below 35% contribution margin, fix pricing or COGS before you raise ad spend. In our client work, this is the usual root cause, and media efficiency alone cannot solve it.
Funnel conversion. At 1.0% conversion you pay for 100 clicks to get 1 buyer. Raise it to 2.5%, a realistic target with proper CRO work, and you pay for 40 clicks to get 1 buyer, so CPA falls by 60% with no change to the budget.
Repeat purchase rate. Document your real 90-day and 12-month repeat rates. If 40% of customers buy again within 12 months, the 12-month LTV changes the economics a lot. If only 15% do, the first order has to carry the full acquisition cost, and that limits the CAC you can afford.
With contribution margin above 40%, a converting funnel and documented repeat rates, scaling is an investment with a predictable return. With any one of them broken, every extra pound of spend adds to the loss.
The question to answer before you raise the budget
If you acquire 500 new customers next month at your current blended new-customer CAC, and each earns the average contribution margin from your last 90 days of Shopify data, are those 500 customers profitable on acquisition?
If yes, scale.
If the answer needs a future LTV improvement, a COGS reduction or ROAS holding at higher spend, test that assumption at your current scale before you bet 6 figures on it.
The £100k mistake is scaling spend before the model is proven.
Know your contribution per order first. Then the budget has something to stand on.
Related reading:
- Breakeven ROAS: the calculation that prevents this mistake: the number you need before you scale
- CAC vs LTV: the 2 numbers that determine whether scaling is safe: unit economics explained clearly
- How to allocate your ecommerce marketing budget: budget sizing that starts from unit economics, not gut feel
We work through this calculation before every new engagement. Our growth strategy service builds the plan around it. If you want to know whether your economics support your growth plan, let’s look at the numbers together.
Quick answers
What is the most common £100k mistake in ecommerce marketing?
Scaling paid media spend on a product or funnel where the unit economics are negative or barely positive. The brand sees ROAS above 3x, assumes it is profitable, and pushes spend from £20k a month to £60k a month. 6 months later, the bank account is down 6 figures despite growing revenue. The revenue was real; the profit never existed.
How do you know if your unit economics support scaling?
Calculate your contribution margin (revenue minus COGS, fulfilment, returns, payment fees) and your blended new-customer CAC (total ad spend divided by new customers, not total orders). If contribution margin per order exceeds blended CAC, the unit economics are positive and scaling is supportable. If they don't, every pound of additional spend loses money.
What should you fix before scaling ad spend?
3 things: product margin (contribution margin should be at least 40 to 55% for DTC before scaling paid media), funnel conversion rate (raising it from 1.0% to 2.5% cuts your CPA by 60% without touching the budget), and documented repeat purchase rates that validate the 12-month LTV your acquisition cost is being measured against.
Is it ever right to scale at negative unit economics?
Only with a clear, modelled path to positive unit economics, typically through subscription conversion, high documented LTV, or confirmed COGS reduction from volume. You need sufficient capital to sustain losses to the inflection point. Most brands that attempt this run out of capital before they get there.
