A warehouse KPI is a metric that tells you whether the operation is delivering what it promised: that inventory is real, that the order ships complete, and that it ships on time. Everything else is context.
If you work with a logistics operator, these seven tell you whether things are going well — and they are worth requesting in writing, on a fixed schedule, not only when something goes wrong.
The 7 that actually say something
1. Inventory accuracy
How it is calculated: locations counted with no discrepancy ÷ locations counted × 100.
It is the base KPI, because nearly every other one depends on it. If inventory is not real, the store sells what is not there and the order dies at picking. It is measured with cycle counting, not with the annual inventory: a number calculated once a year cannot correct anything.
Watch the trap of measuring it by piece instead of by location: counting 10,000 pieces and missing 20 gives a 99.8% that looks great and hides that 20 orders are about to go out wrong.
2. Perfect order
How it is calculated: orders delivered complete, correct, on time and undamaged ÷ total orders × 100.
It is the only one that measures the full customer experience. Its virtue is that it is multiplicative: if each stage runs at 98%, the perfect order does not land at 98% — it lands considerably lower. That is why it is usually the most uncomfortable number in the report, and the most honest.
3. Picking accuracy
How it is calculated: lines picked correctly ÷ lines picked × 100.
It isolates warehouse human error: wrong product, wrong quantity, wrong size. It explains a good share of the returns that were not the customer's fault. We wrote separately about what a picking error costs.
4. Order cycle time
How it is calculated: time between the order reaching the warehouse and the parcel being handed to the carrier.
Do not confuse it with delivery time: that one belongs to the carrier. This measures only your part, and it is the part you can actually demand. If your operator takes 26 hours to release the parcel, no carrier saves the promise.
5. Cutoff compliance
How it is calculated: orders received before the cutoff that shipped the same day ÷ orders received before the cutoff × 100.
The cutoff is the deadline for an order to ship today. This metric is what actually holds up the delivery promise you make to the buyer, and almost nobody asks for it. It belongs in the SLA — the service level agreement, meaning what the operator commits to in writing.
6. Receiving time
How it is calculated: hours between the shipment arriving and the inventory becoming available to sell.
The most ignored one, and the one that holds up the most money: goods physically in the warehouse but not yet sellable are idle capital. It depends heavily on information arriving before the boxes — which is why the packing list and how receiving at a 3PL works both matter.
7. Returns restocked
How it is calculated: returned units made available for sale again ÷ returned units × 100, and in how many days.
A return that arrives and sits for weeks on an unreviewed pallet is inventory you paid for twice. We cover it in depth in returns: what happens to goods that come back.
Quick table
| KPI | What it answers | Where it breaks if it's bad |
|---|---|---|
| Inventory accuracy | Is the data real? | Overselling and incomplete orders |
| Perfect order | Was the customer served well? | The whole process |
| Picking accuracy | Did they pull the right thing? | Picking zone |
| Cycle time | Did it ship fast? | Capacity or priorities |
| Cutoff compliance | Did we keep the promise? | Cutoff, staffing or coordination |
| Receiving time | Can I sell it yet? | Dock and data entry |
| Returns restocked | Do I get the product back? | Reverse logistics |
The ones that get reported and say nothing
Orders processed per day. Measures volume, not performance. A high number with 3% error is worse than a smaller one done right.
Warehouse occupancy percentage. It matters to whoever rents the building, not to you. A warehouse at 95% occupancy usually picks slower, not better.
Productivity per person. It depends on product mix, order size and season. Comparing months with different mixes compares nothing.
Any metric without a denominator. "We had 12 incidents" means nothing without saying out of how many orders.
How to ask for them
Three conditions for the report to be useful:
- A fixed schedule. Monthly at minimum. A report that shows up only when something went wrong is a defense, not a measurement.
- The formula in writing. Two operators can both report "98% accuracy" while measuring different things. If how it is calculated is not written down, the number is not comparable — not even against itself.
- The denominator visible. Always. It is the difference between data and a statement.
And one warning worth more than the seven metrics combined: if the operator measures, reports and grades itself, the number carries a natural bias. That is not bad faith, it is design. So cross-check at least one against your own source: your support tickets, your returns, or your store's delivery dates.
Want an operation that lets itself be measured by these numbers? See our warehousing service, or let's talk about 3PL in Mexico if you need the full operation.
Frequently asked questions
What is the most important warehouse KPI?
Inventory accuracy, because nearly all the others depend on it. If the data is not real, perfect order and cycle time are measuring on a wrong basis.
What inventory accuracy percentage is good?
There is no standard figure we can quote you with a source, and be wary of anyone who gives you one without explaining how they measure it. What you can demand is that the formula be written down, that it be measured by location and not only by piece, and that the trend be visible month over month.
Is my logistics operator required to give me these metrics?
Only if it went into the contract. That is why it is worth requesting the report and its cadence during negotiation, while you still have leverage.