What does warehouse productivity measure?
Warehouse productivity is units of work divided by hours of labor, and both halves of that ratio have at least three common definitions, so the same building in the same week can honestly report three different numbers; measuring productivity well is mostly the discipline of choosing one definition, writing it down, and never changing it silently. Everything below follows from that.
"Units of work" can be shipped units, order lines, orders, cartons or pallets. "Labor hours" can be paid hours, productive hours or direct hours. Each combination is a legitimate measure of something; none of them is the measure.
Which unit of work?
| Unit | What it counts | Best for | Watch out for |
|---|---|---|---|
| Units (eaches) | Every individual item shipped | Buildings with fairly constant units per line; unit-level costing | A case-pick operation looks heroic in units; a promotion of multipacks changes the number with no change on the floor |
| Lines | One SKU on one order, any quantity | Picking; the most stable measure of picker effort in each-pick buildings | Units per line changes mean the same lines carry different work |
| Orders | Completed shipped orders | Packing and shipping; what the business feels | Lines per order changes hide picking productivity entirely |
| Cartons or pallets | Shipping containers | Case and pallet operations; dock work | Says nothing about the picking inside the box |
Which hours?
| Hours | What it includes | Best for | Watch out for |
|---|---|---|---|
| Paid hours | Every hour on the payroll for the agreed labor scope, including paid breaks, meetings, training and waiting time | Cost; budgets; labor cost per order | The most honest number and the lowest; a building that only reports productive-hour rates is hiding its indirect time |
| Productive (on-task) hours | Time clocked to the process while work was available | Engineered standards, staffing models, method comparisons | Depends on how faithfully people clock in and out of tasks; easy to flatter |
| Direct hours | Paid hours of people assigned to the process, excluding supervision and support | The middle ground most weekly reports use | The definition of “direct” drifts unless it is written down |
Same week, three answers: a fictional example
Fictional building, one week: 9,600 units shipped on 6,400 lines and 4,000 orders. Picking staff: 640 paid hours, of which 560 were direct hours on the pick process and 480 were productive hours clocked to picking tasks. All numbers invented to show the method.
| Definition | Calculation | Result |
|---|---|---|
| Units per paid hour | 9,600 ÷ 640 | 15.0 |
| Units per direct hour | 9,600 ÷ 560 | 17.1 |
| Units per productive hour | 9,600 ÷ 480 | 20.0 |
| Lines per paid hour | 6,400 ÷ 640 | 10.0 |
| Lines per productive hour | 6,400 ÷ 480 | 13.3 |
| Orders per paid hour | 4,000 ÷ 640 | 6.25 |
Every number in the table is correct. The building's productivity is 15, 17, 20, 10, 13 or 6 depending on the definition. If last month's report said "20" and this month's says "15", nothing happened except a change of denominator. That is the single most common way a productivity conversation goes wrong.
Which definition flatters the number?
- Productive hours flatter. The smaller the denominator, the higher the rate. A building that wants to look good reports units per productive hour; a building that wants to know what the work cost reports units per paid hour.
- Units flatter in multipack and case operations. When a SKU ships as a case of twelve, units per hour is twelve times lines per hour. Nothing wrong with it, as long as everyone knows.
- Blended rates flatter the slow method. A pallet picker at hundreds of units per hour averaged with each-pickers at tens hides both. Split by pick method.
- Averages of daily rates flatter busy days. A simple average of five daily ratios is not the week's ratio. Divide the week's summed units by the week's summed hours.
The rule is not "use the strictest definition". It is: use the definition that answers the question, name it in the header, and reconcile the hours whenever you switch. A staffing model needs productive-hour rates and a labor budget needs paid-hour rates, and the difference between them, the indirect time, is itself a number worth tracking.
How to reconcile the hours
- Start from paid hours in the agreed labor scope for the period. This is the total nothing else can exceed.
- Subtract hours paid to people not assigned to the process (supervision, support) to get direct hours. Write down who is in and who is out.
- Subtract clocked non-task time (breaks, meetings, waiting, cleanup) to get productive hours. If task clocking does not exist, say so; do not estimate a productive-hour rate from memory.
- Check that direct hours minus productive hours is a plausible indirect share, and track that share on its own line.
- Publish all three denominators once, then report on one, named.
How to build the weekly productivity report
One page, produced the same way every week, beats a dashboard nobody trusts.
| Line | Definition used | Source |
|---|---|---|
| Units per direct paid hour | Shipped units ÷ direct paid hours (agreed scope) | Shipment export; labor export |
| Lines per direct paid hour, by pick method | Lines ÷ direct paid hours, each-pick and case-pick reported separately | Pick transactions; labor by method if available |
| Orders packed per paid hour | Orders ÷ packing paid hours | Shipment export; labor export |
| Labor cost per completed shipped order | Agreed direct labor cost ÷ completed shipped orders | Labor export; shipment export |
| Overtime share | Recorded overtime hours ÷ direct paid hours | Labor export |
| Indirect share | (Paid − productive) ÷ paid, where task clocking exists | Labor and task systems |
| Mix line | Lines per order; units per line; single-line share | Order-line export |
| Comparison | This complete week vs. the prior complete week with the same weekday mix; holidays and shutdowns labeled | — |
The KPI formulas guide works the core measures on a fictional week; the cost per line guide covers the cost ratios; the free Flow Starter builds the one-week page in your browser. When the report says productivity moved, the operations audit checklist is the list of places to look.
Questions people ask
What is a good units per hour for a warehouse?
There is no honest single answer: units per hour depends on units per line, pick method, SKU count, order profile, equipment and the hours definition. The useful question is whether your building's number, on one fixed definition, is moving, and why. Published ranges without a stated sample and definition are not a standard.
Should productivity be measured in units, lines or orders?
Measure what the process handles. Picking handles lines (and units within them); packing handles orders; receiving handles receipt lines and pallets. Report each process in its own unit and keep the building-level report in one named unit.
Paid hours or productive hours?
Both, for different purposes. Paid hours tell you what the work cost. Productive hours tell you the rate at which the work was done while it was being done. A staffing model needs productive-hour rates; a budget needs paid-hour rates. Never move between them without reconciling the hours.
How many weeks do I need before comparing?
Enough complete weeks with the same weekday mix to see a pattern, and at least one that you know was normal. Four weeks shows a trend; a year shows seasonality. One week shows one week.
Get the number on one definition, from your own exports.
The free Flow Snapshot covers one building and one recent week: up to two supported metrics, the exact definitions used, the formulas, and up to three prioritized next checks.
Request a free Flow Snapshot →One per business. No card, no purchase obligation. Data fit and a review slot agreed first.