The easiest way to lose money in eCommerce is to look at the wrong numbers and feel reassured. These are the metrics that decide whether growth is creating value — or burning it.
The wrong numbers will reassure you all the way to insolvency.
Revenue is not enough. ROAS on its own is not enough. Even a healthy gross margin can flatter a weak model if paid media, shipping, returns and discounts are quietly eating the business.
That is why we anchor founders around trading margin: the money left after the direct product cost, fulfilment drag and variable marketing spend have all had their say. ROAS, AOV, repeat purchase and time-to-second-order then explain whether trading margin is structural — or temporary.
2The five metrics
The numbers that tell the truth.
Average order value matters because low baskets leave too little room for paid acquisition. Repeat purchase matters because that is where a weak first-order model can be rescued. Time to second order matters because a repeat six months away is much less helpful than one inside thirty days.
01Anchor metric
Trading Margin
Revenue − COGS − Fulfilment − Variable marketing
The money left after the direct product cost, fulfilment drag and variable marketing spend have all had their say. A healthy gross margin can flatter a weak model — trading margin cannot be flattered the same way.
25–35%
Target range
Watch-out
Paid media, shipping, returns and discounts quietly eat the business. Anchor founders here.
02Efficiency
ROAS / MER
Ad revenue ÷ Ad spend • Total revenue ÷ Total marketing
Too many dashboards use the wrong revenue base, the wrong attribution view or a VAT-inclusive number that flatters performance. Pair ROAS with MER — total revenue divided by total marketing spend.
3–5×
Blended MER target
Watch-out
ROAS isolated by platform double-counts conversions; reach for MER.
03Acquisition headroom
Average Order Value
Net revenue ÷ Number of orders
AOV matters because low baskets leave too little room for paid acquisition. Bundles, free-shipping thresholds and post-purchase upsells are the levers — but a discount-driven AOV is self-defeating.
+15–30%
Lift from bundling
Watch-out
AOV propped up with discounts erodes margin faster than it lifts cash.
04Retention
Repeat Purchase Rate
% of customers who buy again within 90 days
Repeat purchase matters because that is where a weak first-order model can be rescued. A single repeat-rate number hides cohort decay — always look at it by acquisition month.
30%+
At 90 days
Watch-out
Reported as a single number, it hides the cohorts that never come back.
05Cash velocity
Time to Second Order
Median days between order 1 and order 2
A repeat six months away is much less helpful than one inside thirty days. Time-to-second-order is the truest read on product–market fit because it reflects whether customers actually want the next purchase.
30–60d
Healthy band
Watch-out
Power-user skew. Use median, not mean.
Get clearer forecasts and smarter stock decisions
Book a discovery call if you want clearer forecasting, better stock decisions and more confidence in the economics of growth.
Ask these four questions of any dashboard. If it cannot answer them, it is decoration.
01
Are we making money on the first order?
02
If not, how quickly does the second order arrive?
03
What proportion of a new cohort comes back?
04
How much of our customer value is real, and how much is just hope?
“If a business does not know those answers, it is not scaling. It is guessing at scale.”
That is why Ascendant pushes a design-for-profit mindset early. If you expect around a third of revenue to disappear into paid media, another meaningful share into product cost, more into shipping and returns, and more again into overhead, then a weak starting margin profile leaves almost nothing left to fund growth.
You cannot rely on enthusiasm and headline revenue to rescue that.
This is illustrative — the shape is the lesson, not the exact percentages. Where revenue actually goes before it becomes trading margin.
Illustrative · Indexed to 100% of revenue
Where revenue actually goes before it becomes trading margin.
Illustrative · Indexed to 100% of revenue
Where revenue actually goes before it becomes trading margin.
Revenue100%
COGS−25%
Shipping & fulfilment−15%
Paid media−20%
Returns & discounts−10%
Trading margin=30%
100%
Revenue
−25%
COGS
−15%
Shipping & fulfilment
−20%
Paid media
−10%
Returns & discounts
=30%
Trading margin
PillarDeductionResult
≈ 30%
Of every £1 of net revenue, only about 30p reaches trading margin. That is the pool that has to fund overhead, growth and profit — which is why a weak starting margin profile leaves almost nothing left to fund growth.
5At a glance
Formula reference
Print this. Tape it to the wall above the dashboard. Hold every monthly review against it.
Metric
Formula
Healthy range
Common pitfall
Trading margin
Rev − COGS − Fulfilment − Var. marketing
25–35%
Overhead allocated in; flatters channel mix.
Gross margin
Revenue − COGS
60–75%
Hides true unit economics. Useful, not sufficient.
ROAS
Ad revenue ÷ Ad spend (per platform)
3–5×
Double-counts conversions across channels.
Total revenue ÷ Total marketing spend
3–5×
Lags shifts in mix between paid and organic.
Net revenue ÷ Number of orders
Brand-specific
Discount-driven AOV erodes margin faster than it lifts cash.
Repeat purchase rate
% buying again within 90 days
30%+ at 90d
Single # hides cohort decay; report by acquisition month.
T2Time to 2nd order
Median days between order 1 and order 2
30–60 days
Power-user skew. Use median, not mean.
Flagship Guide — Free Download
The eCommerce Playbook
Our master guide: 50 pages of everything we know about scaling online brands profitably — channel economics, ROAS, margin design, stock, cash and exit value.
We have hundreds of online brands supported across Shopify, Amazon and wholesale. Whether it’s for help with cash, margin, stock or messy numbers, please book in.
No. ROAS isolated by platform routinely double-counts conversions, ignores organic and email-driven revenue, and is sensitive to attribution model choice. Two campaigns with the same on-platform ROAS can have very different blended impact.
Pair ROAS with MER (total revenue ÷ total marketing spend). MER is harder to game and surfaces the question that actually matters — is marketing in aggregate creating margin?
02How does AOV affect what we can spend on acquisition?+×
Directly. If your trading margin per order is £18, you cannot spend £25 acquiring the customer and expect to survive on the first order. Lifting AOV widens the budget you can spend on Meta or Google without going under-water.
The lever pairs with repeat. A modest AOV combined with a strong 90-day repeat rate gives you patience; a low AOV with a weak repeat rate forces ruthless CAC discipline.
03What is a healthy repeat-purchase rate?+×
For consumables and apparel, 30%+ within 90 days is a good marker. For higher-consideration goods, the window stretches but the principle holds: customers who do not return inside 90–180 days rarely return at all.
Always report by acquisition cohort. A single blended number can mask the fact that recent cohorts have collapsed while older ones prop up the average.
04How do we calculate trading margin properly?+×
Start from net revenue (after returns and discount leakage). Deduct COGS at fully-loaded landed cost. Deduct fulfilment (3PL, packaging, last-mile carriage). Deduct variable marketing — every pound of Meta, Google, marketplace and influencer spend that scales with revenue.
What is left is trading margin. Do not allocate overhead inside it. Overhead belongs in the next step — operating margin — because mixing the two destroys channel and SKU comparability.
05Why does time-to-second-order matter more than total repeat?+×
Cash. A second order at 30 days repays acquisition while you are still spending into the next cohort. A second order at six months means you have funded six months of working capital with hope.
It is also the most honest signal of product fit. People who genuinely love the product come back fast — long gaps suggest the first purchase was driven by discount or one-time intent.
