How to use this checklist
This is a warehouse operations audit checklist for productivity, cost and service, not a safety or compliance inspection: forty checks across nine processes, each with the number to pull, the record it comes from, and what a good answer looks like measured against your own building rather than an industry table. Work through it with your exports open. Where a record does not exist, write not measured and move on; the gaps are findings too.
Three rules keep the checklist honest. First, define every measure before you calculate it, and write the definition down; the KPI glossary lists the common variants. Second, calculate period ratios from summed totals, not from averages of daily percentages. Third, compare complete periods with the same weekday mix, and label holidays, shutdowns and material workload changes. The KPI formulas guide works a fictional example of each core measure.
"What good looks like" on this page is deliberately not a benchmark. Published warehouse benchmarks rarely state their sample, their definitions or their date. A building compared with its own agreed baseline, on its own definitions, produces a finding you can act on; a building compared with an unsourced range produces an argument.
1. Data and definitions
| Check | Number to pull / source | What good looks like |
|---|
| 1.1 One row means one thing | Row count vs. distinct order IDs in the shipment export | Counts agree, or the difference is explained (multi-carton, split shipments) and documented |
|---|
| 1.2 Promise dates exist | Share of completed orders with a promised ship date/time | Close to 100%; the excluded share is reported, never silently treated as on time |
|---|
| 1.3 Time zones agree | Ship timestamp zone vs. promise timestamp zone | Same zone, or a documented conversion |
|---|
| 1.4 Labor scope is written down | Definition of direct labor: wages, agency, employer costs, overtime premium, which departments | One definition, used in every report; changes dated |
|---|
| 1.5 Units are units | Numerator used in productivity: units, lines, orders or cartons | One numerator per report, named in the header |
|---|
| 1.6 Duplicates are reconciled | Exact duplicate rows and repeated IDs (free CSV check) | Zero exact duplicates; repeated IDs explained |
|---|
2. Receiving and putaway
| Check | Number to pull / source | What good looks like |
|---|
| 2.1 Dock-to-stock time | Receipt timestamp to available-to-pick timestamp, by receipt; median and 90th percentile | Stable week to week; the 90th percentile close to the median; no receipts stranded past a defined limit |
|---|
| 2.2 Receipt lines per paid hour | Receipt lines confirmed ÷ receiving paid hours | Stable for a stable inbound mix; step changes explained by mix or staffing |
|---|
| 2.3 Receipt-to-ASN match | Receipts with quantity variance ÷ receipts | Variances traced to supplier, not to the count process |
|---|
| 2.4 Putaway confirmation lag | Putaway confirmed timestamp minus receipt timestamp | Confirmed the same shift; long lags visible as a backlog, not hidden |
|---|
| 2.5 Dock backlog at shift end | Unreceived trailers or pallets at the cut-off, by day | Flat or falling; spikes tied to known inbound peaks |
|---|
3. Inventory accuracy and replenishment
| Check | Number to pull / source | What good looks like |
|---|
| 3.1 Location accuracy | Locations counted correct ÷ locations counted, by zone | High and stable; the worst zone identified and counted more often |
|---|
| 3.2 Count coverage | Locations counted in the period ÷ active locations | Every active location counted on a defined cycle; fast movers more often |
|---|
| 3.3 Pick exceptions from inventory | Short picks and location-empty exceptions ÷ picks | Falling after counts improve; exceptions traced to a cause |
|---|
| 3.4 Replenishment timeliness | Pick-face stockouts per shift; replenishment tasks completed before the pick wave | Stockouts at zero on the fast movers; replenishment ahead of the wave |
|---|
| 3.5 Shrink and adjustments | Net inventory adjustments by reason code and value | Small, explained, and not growing |
|---|
4. Picking
| Check | Number to pull / source | What good looks like |
|---|
| 4.1 Units per direct paid hour (picking) | Units picked ÷ picking direct paid hours, by shift and week | One definition; variance between shifts explained by mix or staffing, not left unexplained |
|---|
| 4.2 Lines per hour by pick method | Lines ÷ hours, split by each-pick, case-pick, pallet-pick | Each method measured on its own; no blended rate hiding a slow method |
|---|
| 4.3 Pick accuracy | Picks corrected at pack or returned as wrong item ÷ picks | Tracked, and the top three SKUs or locations with errors named |
|---|
| 4.4 Travel share | Pick transactions with location: frequency by zone vs. distance from pack | Fast movers in the nearest zones; the top 20% of SKUs by picks sit in the golden zone |
|---|
| 4.5 Batch and wave fill | Orders per batch; wave release time vs. cut-off | Batches full; waves released early enough to ship on the day's cut-off |
|---|
| 4.6 Single-line share | Single-line orders ÷ orders | Known and stable; a change here explains a change in cost per order before anything else does |
|---|
5. Packing and shipping
| Check | Number to pull / source | What good looks like |
|---|
| 5.1 Orders packed per paid hour | Orders packed ÷ packing direct paid hours | Stable for a stable mix; packing station starvation visible as low output with full staffing |
|---|
| 5.2 On-time shipment rate | Eligible completed orders shipped by promise ÷ eligible completed orders | High; misses clustered and explained (carrier cut-off, late waves, backlog) |
|---|
| 5.3 Carrier cut-off misses | Orders shipped after the carrier pickup on their promise day | Rare; each miss dated and traced |
|---|
| 5.4 Carton and DIM fit | Billed dimensional weight vs. actual; box size distribution | Few boxes billed far above actual weight; box mix matches order mix |
|---|
| 5.5 Ship confirmation lag | Pack complete timestamp to carrier manifest timestamp | Minutes, not hours; no orders packed but unmanifested at day end |
|---|
| 5.6 End-of-day staging | Orders packed but not shipped at the cut-off | Zero on promise-day orders |
|---|
6. Labor and staffing
| Check | Number to pull / source | What good looks like |
|---|
| 6.1 Labor cost per completed shipped order | Agreed direct labor cost ÷ completed shipped orders, by week | Explained by volume, rate or mix when it moves; not left as a mystery |
|---|
| 6.2 Overtime share | Recorded overtime hours ÷ direct paid hours | Follows volume, not the calendar; not structural on a fixed weekday |
|---|
| 6.3 Direct vs. indirect split | Indirect paid hours ÷ total paid hours, by department | Known, stable, and reviewed when volume changes |
|---|
| 6.4 Absence rate | Unplanned absent hours ÷ scheduled hours | Known by shift; staffing plans use it |
|---|
| 6.5 Temp and agency share | Agency hours ÷ direct paid hours | Rises and falls with volume; not a permanent share at a premium rate |
|---|
| 6.6 Hours vs. plan | Paid hours vs. planned hours from the volume forecast | Within a stated tolerance; the plan is updated when forecasts change |
|---|
7. Orders, backlog and service
| Check | Number to pull / source | What good looks like |
|---|
| 7.1 Backlog age | Age of open orders at the daily cut-off, by bucket | Nothing beyond the promise horizon; buckets shrink after peak days |
|---|
| 7.2 Cancellations | Orders cancelled after release ÷ orders released, by reason | Low; inventory-caused cancellations traced to accuracy checks |
|---|
| 7.3 Order accuracy | Orders with a reported error ÷ orders shipped | Tracked from returns and tickets, by error type |
|---|
| 7.4 Lines per order and units per line | Averages by week | Watched: a mix shift here changes every productivity number |
|---|
8. Returns
| Check | Number to pull / source | What good looks like |
|---|
| 8.1 Returns processing time | Return receipt to disposition timestamp | Days, stable, with a backlog count at the cut-off |
|---|
| 8.2 Return reasons | Returns by reason code | Warehouse-caused reasons (wrong item, damaged in pack) separated and falling |
|---|
| 8.3 Restock rate | Returned units restocked to sellable ÷ returned units | Known; the non-restock share explained |
|---|
9. Systems, reporting and peak readiness
| Check | Number to pull / source | What good looks like |
|---|
| 9.1 Export availability | Days to produce a shipment, labor, inventory or receipt export | Same day; no report requires a project |
|---|
| 9.2 Weekly report exists | One page with the four core measures and their definitions | Produced every week, same definitions, same weekday mix in the comparison |
|---|
| 9.3 Capacity vs. peak forecast | Peak daily orders vs. process capacity at planned staffing (free calculator) | Capacity above forecast with a stated margin; the binding process named |
|---|
| 9.4 Hiring and ramp plan | New-hire productivity ramp assumptions in the staffing model | Written down and checked against the last peak |
|---|
| 9.5 Definitions sheet | A document listing every KPI, formula and variant in use | Exists, dated, and matches the reports |
|---|
What to do with the findings
Sort the checks you could not complete from the checks that failed. The first list is a data project; the second is an operations project. Then rank the failed checks by the money or service involved and the effort to fix, and give each one an owner, a measure and a review date. If the list is longer than five items, it is a plan, not a priority list; cut it.
If you would rather have the numbers built for you from the same exports, the warehouse operations audit covers every check above that your data supports, at a fixed $4,500 for one building, and the free Flow Snapshot covers the four core measures from one week of two exports.
Want the core measures built from your own exports?
The free Flow Snapshot covers one building, one recent week and up to 10,000 combined shipment and labor rows: one page with up to two supported metrics, the formulas, and up to three prioritized next checks from this list.
Request a free Flow Snapshot →One per business. Data fit and a review slot agreed first. No card or purchase obligation.