LESSON 11 OF 13
Three mistakes that quietly kill eCommerce profitability
Budgets built on feel, blind discounts, and overhead that compounds.
Knowing what not to do matters as much as knowing what to do. These three take brands out of the game before they reach their potential.
Mistake 1 — Making up the ad budget
Setting the budget on what feels comfortable to risk rather than on what the business maths requires. Here is what it produces in practice: you hand a marketing team a budget that is mathematically incompatible with the goal — a chunk of money with a hope attached — and then hold them accountable for the result.
The underlying maths required a 22 ROAS to break even. The team delivers a 4 and is delighted, because a 4 is genuinely good work. You are haemorrhaging money. Everyone is angry and nobody understands why: from your seat marketing is underperforming, from theirs they did a great job. The infrastructure was set up to fail.
What usually happens next makes it worse — fire the agency, or cut the budget, which raises the ROAS the business needs and makes the problem harder. Then the cycle repeats with the next agency, because the underlying maths was never fixed.
The fix
Calculate break-even ROAS from the ad budget before campaigns run, not after. Work backwards: fixed expenses and variable margin give you the revenue needed to produce enough contribution profit to cover the fixed costs, and the ad budget remainder against that revenue gives you the ROAS.
Mistake 2 — Running discounts blind
This is one of the most common ways eCommerce brands destroy their own capacity for profit. You run a sale, the number is more or less made up — 10%, 15%, 20%. Traffic rises, conversion rate rises, the revenue line looks great. Six weeks later the P&L arrives and you made less money than before the discount.
The mechanism is simple. Your variable costs do not change when you discount. Only the price you are driving against changes, so the contribution margin between the two compresses. If you keep spending the same marketing dollars, they now have to perform significantly better just to stand still.
| Discount | Variable margin | Break-even ROAS |
|---|---|---|
| None | 50% | 2 |
| 10% | 56% | 2.3 |
| 20% | 63% | — |
| 30% | 71% | 3.45 |
Look at the 10% row closely. If actual ROAS comes in at 2.15, that reads like a near-10% performance improvement and feels like a win. The business is losing money on every order. At a 30% discount, variable margin of 71% is well past the roughly 60% danger zone, and the business needs a 3.45 just to break even on each order.
The assumption behind every discount is that conversion rate will improve enough to compensate. Sometimes it does. The point is to know the threshold before you run it, so you have a guardrail: above this number it is working, below it, it is not. If you need a 4.2 ROAS for a discount to be positive and two weeks in you are at 3.9, turn it off.
The two move together. If the maths says ROAS has to rise 50% — from 3 to 4.5 — to break even, your site conversion rate has to make the same magnitude of move, from 2 to 3, to avoid generating less contribution profit than before the discount.
Why this matters most in Q4
Clients regularly arrive in January having discovered that everything they discounted through Black Friday, Cyber Monday and the holidays made them less money than not discounting at all. Knowing the guardrail going in is what lets you correct while it is happening, rather than reading about it in Q1.
Mistake 3 — Adding fixed expenses without doing the maths
Payroll, offices, large legal bills — the costs that exist whether or not you make money. This is not an argument against hiring; BottleKeeper had to hire by its fourth year. It is an argument for knowing what the business has to do to pay for the decision.
| Fixed expenses | Break-even ROAS | Performance increase needed |
|---|---|---|
| Baseline $100,000 | 4.0 | — |
| +10% → $110,000 | 4.2 | 5% |
| +20% | 4.4 | 10% |
| +30% | 4.6 | 15% |
Marketing has to simply produce that much more, or you have to increase ad budget to pull the required ROAS back down — which means putting more capital at risk. Either is fine as a choice. Neither is fine as a surprise.
This happened at BottleKeeper in 2019. After the best year the company had had — 2018, post-Shark Tank, printing money by their standards — it felt like the business could support more payroll and a better office. Adding those things materially changed how the company operated: from lean, scrappy and crafty to heavy. The people were excellent. It still meant giving back all of 2018's profit in 2019, letting those people go, and losing the office.
Every dollar of fixed expense you add takes that dollar away from net profit, and raises the bar your marketing has to clear. That bar compounds fast.
One diagnostic worth knowing: if you are running at a very high return on ad spend — an 8 or a 9 — and still struggling to break even, look at fixed expenses. It often means you are underfunding the ad budget, so even at a high return you are not generating enough contribution profit to cover the overhead. Either bring the fixed costs down, or increase the budget and let required ROAS fall to where the two balance.