Average shipping costs: how do you calculate per order?

A rate table can tell you what an individual route starts at. It cannot tell you what your own shop spends on the average order without your order mix. A store sending mainly domestic parcels gets a different result from one sending mainly cross-border parcels, even if both read the same table. The calculation becomes useful when you keep the prices, shipment counts and customer delivery payments in separate columns.

Which average shipping cost are you trying to calculate? #

Use cost per parcel to understand the delivery units you buy. Use cost per dispatched order to understand what fulfilling a customer purchase costs. Use net shipping subsidy per order to understand how much remains after the delivery money collected from customers. Write the chosen measure beside the result. A single number called average shipping can hide a change in parcel count or customer charges that matters more than the label price.

Also decide whether you are forecasting or measuring. A forecast combines dated quotes with expected shipment counts. An observed average combines actual expenses with the orders or parcels those expenses relate to. The worked example below is a forecast, using checked AirPost carrier-only index minima and explicitly hypothetical counts. It is neither an average paid by AirPost users nor a benchmark for all European shops. For the wider price snapshot, see the existing European parcel costs guide.

What information belongs in your shipping worksheet? #

Attach a quote date for forecasts and an invoice or transaction reference for measured costs. Keep packaging purchases and stock replenishment in separate expense categories, then include an appropriate share if you are calculating the full fulfilment budget. Avoid mixing a box-only transport average with another month’s transport-plus-preparation average. The result can be arithmetically correct while still comparing two different things.

How do you calculate a weighted average from route prices? #

Multiply each route’s price by its parcel count, add the resulting subtotals and divide by the combined parcel count. The formula is weighted carrier average = sum of route price × parcels on that route ÷ total parcels. Sum the amounts before rounding the average to cents. For your own records, repeat this for each service and parcel band that has a different price rather than assigning one country-wide value to unlike boxes.

This example assumes 20 parcels within the Netherlands, 6 from the Netherlands to Germany and 4 from the Netherlands to Belgium. Each weighs 2 kg. The source is exclusively AirPost’s carrier-only rate index, retrieved on 10 October 2026, which excludes the platform fee. Each observation is the cheapest standard-service minimum for its route column; confirm eligibility and the desired delivery product before using it in a real order budget.

The carrier-only spend is €178.00 across 30 parcels, giving €5.93 per parcel after rounding. Averaging the three route prices equally would give €7.14 instead. That second calculation assumes the same number of shipments on each route, which this example does not have. Neither average is a new carrier quote: it is a summary of the chosen inputs.

What changes when you add platform and preparation fees? #

The index excludes AirPost’s platform fee, so add €0.85 once per parcel. For an illustrative small-parcel budget, the recommended preparation fee is €2.25. A sender sets their own fee; replace the recommendation with the amount you agree. This model assumes every packed box fits the small-parcel category and an eligible service matches its indexed minimum. The platform and preparation inputs together add €3.10 per parcel.

Across 30 parcels, that adds €93.00 to the carrier-only subtotal of €178.00, making €271.00. Divide by 30 to get €9.03 for the modelled label, platform and preparation average. Packaging you buy and transporting stock to the sender remain outside this figure. Add their actual allocated amounts when you need a complete operating cost. Calling the narrower number all-in would leave those expenses hidden.

Why can cost per order be higher than cost per parcel? #

In the first scenario, assume all 30 parcels correspond to 30 separate customer orders. Cost per order and cost per parcel then match at €9.03 on the stated model. In a separate scenario, keep the same 30 parcels and prices but link them to 24 orders: 18 orders use one parcel each, and 6 use two. The parcel average is unchanged, while expense per order becomes €11.29.

That difference comes from the denominator. You bought the same amount of transport and preparation, but those purchases served fewer customer orders. Count physical boxes rather than treating every order number as one shipment. If a bundle requires two parcels, both belong in the order’s expense. If a replacement creates another movement later, link it to the original order as well so you can see the cost of serving that purchase.

How do customer delivery payments affect the calculation? #

Net shipping subsidy per order = shipping-related expense minus retained customer shipping payments, divided by the relevant orders. Customer shipping payments are income for this calculation, not a reduction in the carrier’s price. Keep the original expense visible so you can compare transport costs independently of your checkout policy. Use shipping payments rather than total merchandise revenue, or the result will stop answering a shipping question.

Return to the 30-order, 30-parcel scenario. As an arithmetic example, suppose you charge each customer exactly their route’s carrier-only index amount. You would collect €178.00 for delivery against the modelled expense of €271.00. The shop would absorb €93.00 across the batch, or €3.10 per order, before the separate packaging and stock-transport costs. Offering delivery free on every order would leave the whole modelled €271.00 to cover elsewhere.

Shopify’s shipping-strategy guidance advises considering the average cost when setting a flat delivery rate. Your weighted calculation supplies that input, but it does not determine the right merchandise margin or free-delivery threshold by itself. If discounts and refunds affect your checkout offer, the free-delivery threshold after discounts guide handles that separate decision.

Where do returns, replacements and refunds belong? #

Include the movements your chosen measure is meant to cover. For the cost of serving customer orders, retain outgoing expenses and add paid replacements or returns linked to those orders. Keep return counts separately visible so an increase can be explained. A refund of a customer delivery charge changes the retained shipping-income column; an actual refund of an expense changes the expense column. Do not assume one automatically produces the other.

Timing also needs a consistent rule. A return may arrive after the month of the original dispatch. You can report costs paid in the current month or group later costs with their originating orders, but state which approach you chose. For day-to-day budgeting, mark unresolved charges as pending and revisit them when the records arrive. That avoids presenting an incomplete early estimate as a final measured average.

How can the average rise when no route price changes? #

Change only the illustrative mix to 10 Dutch domestic parcels, 12 Netherlands-to-Germany parcels and 8 Netherlands-to-Belgium parcels. The total remains 30, and every route keeps the same indexed price. Carrier-only spend becomes €216.50, or €7.22 per parcel. With the same platform and recommended preparation inputs, the modelled average becomes €10.32. More orders now use the dearer cross-border routes.

When comparing periods, show parcel count, order count and route mix beside the average. Also check whether parcel sizes, service choices or customer charges changed. To isolate a tariff change, reprice the same fixed mix with the new prices. To forecast next month, use the mix you actually expect. A single rising average cannot identify which of these factors caused the change.

How do you use the result to assess local fulfilment? #

Through AirPost’s local fulfilment model, a person near your customers holds an agreed stock batch, packs orders and hands them over at a parcel point. Evaluate that arrangement by repricing the relevant orders from the actual local dispatch point and including the sender’s agreed fee. Add the stock batch’s transport cost from your own quote, allocated across the orders it can serve. Compare the same products and delivery requirements before drawing a conclusion.

Use the domestic dispatch guide to plan stock placement. Then check the forecast against actual costs and order counts, keeping the same expense definitions before and after.