What it means
The metric converts a large, lumpy cost base into a number that can be compared with price and margin. Delivery costs include courier fees, fuel, driver wages, vehicle running costs, packaging and the labour of picking and packing, and looking at the total alone tells you little because it naturally rises as sales grow.
Dividing by units delivered separates growth from inefficiency. It matters because delivery is often the difference between a profitable order and a loss-making one.
In low-margin online retail, delivery can absorb 10% to 15% of order value, so a small movement in cost per unit changes the economics of an entire product range. Free delivery offers only work if this number is understood precisely.
Interpretation depends on splitting fixed and variable costs. Vehicles, depot space and salaried drivers are largely fixed, so cost per unit falls as volume rises and rises sharply when volume drops, which is why the metric can improve without anyone doing anything clever.
Genuine improvement shows up when cost per unit falls at constant volume. The definition needs care to stay comparable over time.
Deciding whether packaging, returns processing, failed delivery attempts and warehouse picking labour belong inside the number is a judgement, but whatever is chosen must stay consistent, and returns in particular are often excluded and then quietly forgotten. The most useful version of the metric is segmented rather than averaged.
Cost per unit by region, by carrier, by product size and by delivery speed reveals which combinations are subsidising others, and it is usually the outliers rather than the average that drive an improvement plan.
In practice
Real-world examples.
Example
A grocery delivery service finds its cost per unit is $4.10 in dense urban postcodes and $9.70 in rural ones. It introduces a rural surcharge and a wider delivery window outside the cities rather than raising prices everywhere.
Example
A furniture retailer reduces delivery cost per unit from $38 to $31 by switching from single-item dispatch to weekly consolidated regional runs. Customers wait an average of two days longer, and the retailer saves roughly $420,000 a year.
Example
A subscription box company discovers that a small packaging redesign drops each parcel into a cheaper carrier weight band. The change costs $0.15 per box in materials and saves $0.85 in carriage, a net gain of $0.70 on every unit shipped.
Think of it
“Delivery cost per unit is what each item costs to ship-your shipping expense per piece.
Formula
Calculation
Delivery Cost Per Unit = Total Delivery Cost / Units Delivered
A homeware business spends $270,000 in a quarter on courier fees, fuel, packaging and driver wages, and delivers 45,000 units. Delivery cost per unit is $270,000 / 45,000 = $6.00.
The following quarter it consolidates delivery routes and negotiates a better carrier rate. Total delivery cost rises to $324,000, but volume rises faster to 60,000 units, so cost per unit falls to $324,000 / 60,000 = $5.40. At the old rate, 60,000 units would have cost 60,000 x $6.00 = $360,000, so the saving is $360,000 - $324,000 = $36,000 for the quarter. With an average order value of $48, delivery has dropped from $6.00 / $48 = 12.5% of revenue to $5.40 / $48 = 11.25%.Case study
Seen in the real world.
Marrow and Vale is an invented online homeware retailer used here as an illustrative example. Its delivery cost per unit had been sitting at $6.00, calculated from $270,000 of quarterly delivery cost across 45,000 units, and the board wanted it under $5.50 before launching a free delivery threshold.
The operations team split the average instead of attacking it as a whole. Two findings stood out: single-item dispatches to one region cost almost twice the average, and oversized packaging pushed roughly a fifth of parcels into a higher carrier weight band. Neither was visible in the headline number.
They consolidated that region into three weekly runs, introduced two smaller box sizes and renegotiated the carrier contract on the back of higher committed volume. The next quarter delivered 60,000 units for $324,000, a cost per unit of $5.40, and delivery fell from 12.5% to 11.25% of the $48 average order value. The illustrative lesson is that averages hide the problem, and segmentation is what makes this metric actionable.
Watch out
Common mistakes.
- Watching total delivery cost rather than cost per unit, so ordinary growth is mistaken for a cost problem.
- Changing what is included in the calculation between periods, which makes the trend meaningless.
- Excluding failed delivery attempts and returns, which can understate the true cost by a wide margin.
Questions
People also ask.
What should be included in total delivery cost?
Typically carrier and fuel charges, driver and vehicle costs, packaging, and the picking and packing labour, with the same items included every period.
Why does cost per unit fall as volume rises?
Because a large share of delivery cost is fixed, so spreading depot, vehicle and salaried labour costs across more units lowers the average.
Should returns be counted in the metric?
Ideally yes as a separate reverse logistics figure, because return shipping and processing can add substantially to the true cost of serving a customer.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%