What an Ecommerce Order Actually Earns You

One order, all the way down

Everything below is a worked assumption, not a benchmark. Use the structure and substitute your own contracts; the point of the exercise is that the structure has more rungs than most sellers count. The order: one item listed at $65, shipping charged to the customer at $5, so $70 arrives at the payment gateway. Product cost is $26, which is 40% of the item price. Payment processing is 2.9% of the amount transacted plus $0.30, giving $2.33. The marketplace or platform takes 8% of the total collected, $5.60. Warehouse pick, pack and the box cost $2.50. The carrier charges $8.00 for a parcel of this size and destination.

Subtract them in order and $70.00 becomes $44.00, then $41.67, then $36.07, then $33.57, then $25.57. That last figure is the contribution of a delivered order: what is available to pay rent, salaries, software, advertising and profit. Expressed as a percentage it depends entirely on which denominator you pick — 39.3% of the item price, 36.5% of the total collected — which is the first place two people comparing notes will disagree. Pick one, write it down, and never mix the two in the same sentence.

Why small baskets are disproportionately expensive

Card processing is quoted as a percentage plus a fixed amount, and it is the fixed amount that shapes your catalogue. If the schedule is 2.9% + $0.30, the effective rate on a basket of value V is 2.9% + 30/V expressed in percentage points. On a $10 basket that is 5.90% — more than double the headline. On $30 it is 3.90%. On $70 it is 3.33%. On $200 it is 3.05%, and it never gets below 2.9% no matter how large the order. The fixed fee is 51% of the total fee on a $10 order and 5% of it on a $200 order.

Two per-order costs behave the same way and compound the effect: pick-and-pack, and the carrier bill. On the $70 order they are $2.50 and $8.00 respectively; on a $10 order they would be roughly the same in absolute terms, which is why a $10 order is usually unprofitable at any margin. Add the fixed processing fee, the pick-and-pack and the shipping subsidy together and you have the real reason free-shipping thresholds, minimum order values and multi-buy offers exist. They are not marketing gimmicks; they are attempts to spread three per-order costs over more revenue.

Commission, handling and the shipping you quietly subsidise

Marketplace and platform commissions are usually a percentage, but read the base carefully: many are charged on the total the customer paid, shipping included, so charging the customer for delivery increases the commission you pay. In the worked order, 8% applies to $70 rather than $65, which costs $0.40 more than most sellers expect. Some platforms also add a fixed per-item fee, some charge a separate payment-processing fee on top, and some apply a different rate by category. Three different lines can hide behind the word commission.

Shipping deserves its own line rather than being netted against what the customer paid, because the two numbers move independently. Here the customer paid $5 and the carrier invoiced $8, so the order carries a $3 subsidy. Charge $5 on a heavier or more distant parcel and the subsidy grows silently; offer free delivery above a threshold and every order above it carries the whole $8. The discipline is to record shipping revenue and shipping cost as separate lines and to watch the net, because a change in your customer mix — more rural addresses, more islands, more bulky products — moves it without anybody changing a price.

A return costs more than the sale it cancels

The instinct is that a returned order is a zero: you give the money back, you get the product back, nothing happened. The arithmetic disagrees on six lines. The revenue goes back to the customer in full, $70.00. The commission is refunded by most platforms, so that comes back. But the outbound carrier charge of $8.00 has already been spent, the pick-and-pack of $2.50 has already been spent, and the payment processor — on the assumption used here — keeps its $2.33. Getting the parcel back costs $7.00 to the carrier, inspecting, repackaging and restocking it costs $3.00, and the unit does not come back worth what it left at: assume a 20% write-down on a $26 cost, which is $5.20.

Add them: $5.20 + $2.33 + $2.50 + $8.00 + $7.00 + $3.00 = $28.03 of cost against no revenue at all. The returned order contributes minus $28.03, while a delivered one contributes plus $25.57. The ratio of the two is 1.10, so every return you take destroys the contribution of 1.10 orders that stayed sold. That is the sentence to remember: a return is not a cancelled sale, it is a sale plus a small one.

The break-even return rate

Across a cohort of orders at return rate r, the average contribution per order is (1 − r) × 25.57 − r × 28.03. Setting that to zero gives the break-even return rate: r* = kept contribution ÷ (kept contribution + return loss) = 25.57 ÷ 53.60 = 47.7%. Below that rate the cohort still contributes something; above it, selling more destroys value. Each percentage point of return rate costs $0.54 of contribution per order — the sum of the two figures divided by 100 — so a rise from 8% to 14% takes $3.22 off the average order, more than the payment fee and the pick-and-pack combined.

That 47.7% looks reassuringly far away until you add the cost of getting the order in the first place. Advertising is not refunded when a customer sends the parcel back. Assume $12.00 of acquisition cost per order — again, an assumption; use your own spend divided by your own orders. A delivered order now contributes $13.57 and a returned one costs $40.03. The break-even return rate collapses to 13.57 ÷ 53.60 = 25.3%. And well before that point the damage is visible: at a 10% return rate the average order contributes $8.21 instead of $13.57, so 39.5% of the margin has gone to returns alone.

The formula generalises: r* = C ÷ (C + L), where C is the contribution of an order that stays sold and L is the cost of one that comes back. Both terms are yours to move. C rises with basket size, with a better carrier contract and with anything that lifts average order value past the per-order costs. L falls when the inbound leg is cheaper, when inspection is faster, when the product is easier to resell at full value — and, most of all, when fewer parcels come back at all. Better sizing information, honest photography and accurate descriptions attack L by attacking r.

Building your own version without fooling yourself

Take each number from a document, not from memory. Payment fees come from your processor's published schedule and from the settlement report, which will show you the cases where a different rate applied — international cards, currency conversion, chargebacks. Commission comes from the platform's fee page and from an actual payout statement, because the payout is where the surprises are. Carrier cost comes from a real invoice, including surcharges for fuel, remote areas and oversize, which are the lines nobody quotes when they say what shipping costs them.

Two habits protect the model. First, count returns in the same period as the sales that caused them, not in the month the parcel arrives; a growing business that counts returns as they land will systematically understate its return rate. Second, model per-order costs per order and per-euro costs per euro, and never blend them into a single percentage — the moment you write payment costs me 3% you have lost the fixed fee, and with it the whole reason your small orders are unprofitable.

Frequently asked questions

Why do my small orders lose money even at a good margin?

Because three of your costs are charged per order rather than per dollar: the fixed part of the payment fee, the pick-and-pack, and the parcel itself. In the example here those come to $0.30 + $2.50 + $8.00 = $10.80 on any order, large or small. A $10 order simply cannot carry them at any gross margin, which is why the effective payment rate alone rises from 2.9% to 5.9% as the basket shrinks from very large to $10. Fixing it means raising average order value — thresholds, bundles, multi-packs — not raising the margin percentage.

Should I offer free shipping?

Treat it as a price cut of exactly the shipping revenue you give up, then apply the contribution-margin arithmetic. In the worked order, dropping the $5 shipping charge takes contribution from $25.57 to $20.57, a fall of 19.6%, so free shipping pays for itself only if it lifts orders by more than 24.3% at the same basket size, or lifts the basket enough to compensate. A threshold changes the question: above the threshold you carry the whole carrier bill, so set the threshold where the extra contribution from a larger basket at least covers the shipping you have stopped charging.

Do payment processors refund their fee when I refund a customer?

It varies by provider, by country and by contract, and it has changed more than once — so read your current terms rather than trusting a figure from a forum. Some return the percentage but keep the fixed fee, some keep everything, some return nothing and charge an additional refund fee. The model here assumes the whole $2.33 is kept, which is the conservative case. Whatever your provider does, put it in the model explicitly: on a 10% return rate the difference between keeping and refunding the fee is worth about $0.23 per order across the book.

Is a high return rate always bad?

Not automatically, but it must be paid for somewhere. A generous returns policy can raise conversion and lifetime value enough to fund the losses it causes; the test is whether the extra contribution from the orders you would not otherwise have won exceeds r × (C + L) across the cohort. With $25.57 of kept contribution and $28.03 of loss per return, each extra point of return rate costs $0.54 per order, so a policy change that lifts returns by five points has to lift contribution by $2.68 per order to be neutral. Compute it before, not after.

Where should advertising spend go in this calculation?

Below the contribution line if it is a budget you set, above it if it is spent per order won. Most performance advertising behaves like the second: you pay to acquire this order and you do not get it back if the order is returned. Modelling it as a per-order cost is what turns the break-even return rate from 47.7% into 25.3% in the example here. Brand spend, agency retainers and salaries belong in the fixed pool, where they are covered by total contribution rather than charged to any single order.