Cost per order, done properly
Ask two 3PLs for a quote and you get two documents that cannot be compared. One prices per pick, one bundles the first pick, one charges storage by pallet and the other by bin. All of it is legitimate, and none of it answers the only question you have: what will it cost to get one order to one customer? Here is how to build that number yourself.
Why quotes resist comparison
Fulfillment pricing is unstandardized, and that is not primarily a conspiracy. Warehouses genuinely differ in how they incur cost: a bin-shelved apparel operation and a pallet-heavy beverage operation have different economics, so they bill differently. The result is that every quote is internally coherent and mutually incomparable.
The brands that get this right stop trying to compare rate sheets and instead build one number, from their own order history, and price every provider against it.
The eight components of a real cost per order
Landed cost per order is not one line on an invoice. It is the sum of eight things, some of which never appear on a fulfillment quote at all.
| Component | What drives it | Usually quoted? |
|---|---|---|
| Pick and pack | Units per order, item handling, whether the first pick is bundled | Yes |
| Packaging materials | Carton mix, void fill, branded inserts, dunnage | Sometimes |
| Parcel transportation | Billable weight, zone, service level, accessorials | Estimated |
| Storage | Unit of measure, velocity, seasonal inventory peaks | Yes |
| Receiving | Inbound volume, carton or pallet counts, labeling condition | Yes |
| Returns processing | Return rate, inspection depth, restock or dispose decision | Rarely |
| Exception handling | Reships, address corrections, lost parcels, split shipments | No |
| Your own labor | Hours your team spends managing the provider | Never |
The bottom three rows are where brands get surprised, and they are ranked in ascending order of how badly. Exception handling and internal labor are real costs that no quote contains and no rate comparison surfaces.
The formula: landed cost per order = (fulfillment fees + materials + transportation + allocated storage + allocated receiving + returns cost + exception cost) ÷ orders shipped. Run it monthly on actuals. Compare it to itself over time before you compare it to anyone else.
The three costs that hide
Exception handling
An order that ships correctly the first time costs you the quoted amount. An order that goes wrong costs you the original fulfillment, a second parcel, possibly a replacement unit, a support conversation, and some fraction of a customer relationship. If 2% of your orders become exceptions, they can consume a disproportionate share of your fulfillment budget and nearly all of your team’s attention.
This is why we treat exception rate as an operating metric rather than a service complaint. Our Exception Reports surface problems before you feel them, so issues get managed proactively instead of reactively, and models watch every carrier scan to flag shipments trending toward trouble before your customer writes the review. Reducing exceptions is cheaper than absorbing them, which is a statement about arithmetic rather than about attitude.
Returns
Return rates average roughly 19% to 20% of online orders across ecommerce, with apparel running 20% to 40%. Processing a single return costs roughly $25 to $30 all in once you count return shipping, labor, inspection, and restocking. A 20% return rate is not a customer service statistic. It is a line item that can rival your outbound transportation.
If your cost per order excludes returns, it is not a cost per order. Our breakdown of return economics works through the full arithmetic.
Your own team’s hours
The least measured cost in fulfillment is the time your people spend chasing your fulfillment provider. Reconciling counts, asking where an order is, re-explaining a kitting rule, waiting for a report. It never appears on an invoice, and for smaller teams it is frequently the largest unmeasured expense of a bad provider relationship.
The test is simple: how many hours a week does your team spend on logistics that is not strategic? If the answer is meaningful, a cheaper rate card is not saving you money.
What a low quote can be buying you
Cheap is not automatically bad, and expensive is not automatically good. But a quote materially below market is priced that way for a reason, and the reason is worth identifying before you sign.
- Unbundled essentials. Materials, receiving, or account management priced separately, so the comparison you made was not the deal you got.
- Storage that reprices. A low pick rate paired with storage terms that get expensive exactly when your inventory peaks for peak season.
- Thin verification. Barcode-verified picking, piece-level receiving, and cycle counting cost money to run. Operations that skip them are cheaper per pick and more expensive per month.
- No remedy. What happens when they get it wrong? If the answer has no mechanism behind it, you are self-insuring and paying someone else for the privilege.
On that last point, here is our version, stated as a mechanism rather than a promise. Your orders are picked against a barcode scan: the scan checks the SKU against the order before it leaves the aisle, and a verified Argo pick or pack error earns a service credit against your Argo fees. Terms apply. Separately, when product is lost, stolen, or damaged in our possession, we reimburse at 100% of replacement cost, domestic, with no per-package cap, through a claims process we run for you.
We hold that standard because nobody in this industry asks for it, which is roughly the definition of Excellence as we use the word internally. It is also a genuine cost we chose to carry, and it is part of why we do not compete by being the lowest bid.
Making the number actionable
A cost per order you calculate once is trivia. A cost per order you track monthly is a management tool. Three ways to use it:
Segment it. Cost per order by SKU, by channel, and by zone will not be uniform, and the outliers are where the money is. A single bulky SKU can carry a cost per order several times your average and quietly set your blended margin.
Watch the trend, not the level. Whether your cost per order is $9.40 or $11.20 matters less than whether it moved 6% last quarter and why. Rate increases, product mix shifts, and creeping exception rates all show up here first.
Bring it to your provider. This is the part most brands skip. A provider who can discuss your cost per order against their own operating data is a partner; one who can only discuss their rate card is a vendor. We plan on keeping and earning your partnership every quarter, reviewed openly against the value we’re adding to your operation, and this number is the honest way to have that conversation. Your Partner Success Rep brings the same live data you can see yourself.
That visibility is deliberate. Every shipment, delivery time, and charge is itemized and live in your Partner Portal, exportable whenever you want it. You cannot manage a cost per order you have to file a request to see.
The bottom line
Rate cards are designed to be compared and cannot be. Cost per order is hard to build and answers the question. Build it from your own actuals, include the three costs that hide, segment it, and track it monthly.
Then judge every provider, including us, against that number.
Related reading
What a return actually costs you
Roughly 20% of online orders come back, at $25 to $30 a return. The full arithmetic, and where to intervene.
Read →What a 3PL transition actually looks like
Seven gates, five dependencies, and why nobody honest quotes you a switching timeline up front.
Read →Request a fulfillment review
Bring us your order history and current invoices and we’ll build the landed cost per order with you, segmented by SKU and zone, including the exception and returns costs your quote left out. Thirty minutes, your numbers, an honest read. If you’re already in good shape, that’s a fine outcome to hear.
Request a fulfillment review