Shopify Profit After Shipping

Of all the costs that sit between an order and its profit, delivery is the one that varies most and gets averaged most. Those two facts together are why shipping is the most common place a store discovers it has been losing money on a whole segment of orders.

Why delivery drives the spread

Cost of goods is a property of what was sold, and it moves with the basket. Payment fees are a percentage, so they move with order value. Packaging is close to constant.

Delivery is different. It depends on weight, on parcel dimensions, on how far the parcel is going, on whether the destination is classed as remote, on whether the order shipped in one box or two — and on none of those things being visible at checkout, where a flat rate was charged.

The consequence is that two orders taking identical money can differ by ten or fifteen points of margin. In the example above, thirteen points separate a light metro delivery from a heavy remote one, and every other line is identical.

An average destroys exactly that. A store-level delivery cost reports both orders at the same margin, which is arithmetically true of neither.

Charged versus paid

There are two ways to handle shipping in a profit calculation, and only one of them stays honest.

Keep them separate: shipping charged as revenue, shipping paid as a cost. Correct, and it requires discipline, because the revenue line is easy to remember and the cost line arrives on an invoice weeks later.

Net them into one delivery line: shipping charged minus shipping paid, a single figure per order. Less faithful to the accounts and more useful operationally, because it directly answers whether delivering that order made or lost money.

For management reporting, netting is usually the better choice. The failure mode to avoid at all costs is the hybrid — counting shipping revenue and forgetting the label — which flatters the number twice.

Free shipping, and where it bites

A free-delivery threshold turns delivery from a revenue line into a cost line for every order that crosses it.

The threshold is normally set from average order value, which considers only one side. The question it should answer is whether the additional margin from the larger basket exceeds the full delivered cost of the order — invoiced cost, surcharges included, not the rate card.

Run that on your own orders and the finding is usually specific rather than general: free shipping works fine across most of the catalogue and is loss-making on one weight band or one region. That is a solvable problem. Withdrawing free shipping entirely, which is the instinctive response to a bad delivery month, throws away the conversion benefit everywhere to fix it in one place.

Reading the distribution

Sort orders by profit after shipping, ascending, and look at the bottom decile. The patterns repeat across almost every store:

A weight band where invoiced delivery consistently exceeds what was charged. Fix by weight-based pricing or by excluding those items from free-shipping eligibility.

A cluster of postcodes carrying remote-area surcharges on orders priced as metro deliveries. Fix by regional rates, or by routing those postcodes to a carrier whose extended-area pricing is better.

Split shipments. Two labels against one shipping charge, usually caused by inventory allocation rather than by anything the customer did.

Oversized parcels billed on dimensional weight. The cheapest fix in the entire list, because a smaller carton applies to every future order of that product.

The adjustment problem

Delivery cost has a timing property that distorts recent figures.

Carriers re-measure parcels after collection and rebill the difference. Fuel and remote-area surcharges appear as separate invoice lines. Failed deliveries generate a return leg billed under a different service code. All of it arrives days or weeks after the parcel moved.

So profit after shipping for last week is systematically optimistic, and profit after shipping for a quarter that has fully settled is the number to make decisions on. When adjustments do arrive, attribute them back to the original order rather than to the month they landed in — otherwise the products and regions that cause adjustments look identical to the ones that do not, and the analysis above becomes impossible.

Multi-parcel orders

One case deserves separate mention because it distorts per-order figures more than anything else on the list.

When an order ships in two parcels — because inventory sat in two locations, or because the items would not fit one box — it pays two base rates and two sets of surcharges against a single shipping charge. Delivery cost can double while revenue does not move at all.

The distortion runs both ways in reporting. If only one label is matched back to the order, delivery cost is understated by half and the order looks fine. If both are matched, the order appears as an outlier that nobody can explain without knowing it shipped twice.

Two things fix most of it. Attribute every label to the order rather than the first one found, so the figure is at least correct. Then check how often split shipments happen and why — if inventory allocation is the cause rather than parcel size, the second label is avoidable cost rather than a property of the order.

What good looks like

Not a low delivery cost — that is a function of what you sell and where.

What good looks like is a narrow spread. A store where the cheapest and most expensive deciles differ by a few points has delivery priced in line with what it costs. A store where they differ by fifteen or twenty is subsidising one group of customers with another, usually without having decided to, and usually in a direction that gets worse as it grows into heavier products or wider geography.