Go-live week at the new distribution centre
Contoso Retail's new DC went live last week with warehouse WH-01, three zones (bulk, pick, pack) and full warehouse management processes. Sales orders are released to the warehouse in a nightly batch. On day one roughly 40% of released lines returned no work was created; the affected lines all ship full pallets from the bulk zone. Where work is created, pickers are sent to the back of the bulk zone even though an equivalent pallet sits near the front. The pack zone is the bottleneck: single-order picking has 14 pickers each walking about 9 km a day. Inbound runs 40 loads a day with appointments booked in Outlook, and detention charges reached $14,000 last quarter. Six carriers are used, freight is agreed at a rate per mile plus a per-stop fee, and rates are keyed manually onto each load.
1. Diagnose the no-work-created message for the bulk pallet lines. Give an ordered checklist and explain how each candidate cause would present.
Read the worked answer
Work top-down from the master data to the directive; the symptom is identical across most causes, which is why you need the order.
- Warehouse and item eligibility. Is WH-01 flagged for warehouse management processes, and does the item's storage dimension group include Site, Warehouse and Location with the item released to this legal entity? Ineligible items are handled as non-WHS stock and produce no work at all - this presents as a whole item failing, not a whole zone.
- Wave template. Does the template cover the shipment/load for these orders, is it active, and is it assigned as the default for the warehouse? A template that does not cover full-pallet loads or one that is deactivated during cutover leaves released orders sitting with no work.
- Work template. Is there a work template for the outbound/sales-order-issue work order type for this warehouse? Missing work template presents as a clean refusal with no work lines.
- Location directive. Does a directive exist for the same work order type and transaction type, and does its query actually return bulk-zone locations? Then check the two classic mismatches: (a) the quantity restriction - a directive built for full pallets with a minimum quantity, where available quantity is fragmented; (b) the unit of measure - a directive expressed in Pallets against order lines in Eaches with no conversion will find nothing.
- Availability. Is the on-hand in a location whose inventory status is excluded from picking, blocked, or already reserved by another order? This presents differently: the message typically names the item and warehouse rather than being a pure directive failure.
Because all 40% are full-pallet bulk lines, my first two suspects are the wave template's coverage of full-pallet loads and the directive's quantity/unit restrictions - start there, then widen.
2. Fix the walking and the service level. Which warehouse management configurations actually reduce travel, and what does each require to work?
Read the worked answer
Consolidate demand before optimising the route. Nine km a day per picker on single-order picking is a wave design problem before it is a routing problem.
- Wave and load consolidation. Automate wave creation per load or per carrier cut-off so many orders are picked together. Build loads with a load template so outbound work is created at the load level. Fewer, larger waves means pickers stop at fewer locations.
- Cluster picking for the pick zone: one device and picker serve several orders simultaneously into separate containers. This requires a cluster profile (cluster size, sorting and verification rules), containerization switched on, and a mobile device menu item configured for cluster picking. It is the single biggest lever on travel in a pack-heavy DC, but it fails quietly if the cluster profile exceeds what the containers can segregate.
- Location directive strategy set to Consolidate in bulk and pick zones so the system finishes partially consumed locations before opening new ones - the counterpart of the problem where pickers are sent to the back of the bulk zone while an identical pallet sits at the front. If the directive sequences on location name rather than using a consolidate or FIFO/FEFO strategy, it will keep sending people to the far end.
- Sequence the work. Use directive sequence numbers to express priority (full pallet bulk before piece pick) and keep the directive query tight, so pickers are not routed into a zone for one unit.
Measure before and after: lines per hour, travel per order and on-time pack-out. Without those three numbers the redesign has no verdict.
3. Automate the carrier rates and the dock. What does the transportation setup need, and how does it reduce the $14,000 of detention?
Read the worked answer
Rate automation
- A shipping carrier record per carrier, with a rate engine assigned. Without an engine the load has no rate and someone keys it by hand - which is exactly the current state.
- A rate master in the engine, built for the contract you actually have: a mileage-based structure plus a break or stop-based component for the per-stop fee, with effective dates that match the contract. Effective dates that do not cover today's shipments are the usual reason a configured engine still returns nothing.
- Load building via load template so shipments consolidate into loads and the rate is applied once per load rather than per order line.
- A transit time engine so promised delivery dates and the carrier's actual service days are the same number.
Dock and detention
Replace the Outlook calendar with dock appointment scheduling: dock doors with capacity rules, appointment windows per carrier and per inbound load, and a booking view the carrier can see. Detention is almost always a mismatch between the appointment the vendor booked and the window the DC can actually receive into - consolidating inbound loads through load planning and tying appointments to those loads removes the queue.
Close the loop with freight reconciliation: match carrier invoices against the rated loads with a tolerance, so a rate that is wrong, missing or double-applied is caught at invoice rather than in a quarterly argument. Detention, manual rating and invoice variance should all appear on one weekly scorecard - that is what proves the config is working.