Blog · Board and management reporting · Supply chain
A fulfillment dashboard connects open orders, picking quality and shipment timeliness by warehouse. Follow a worked US distributor example, reconcile the backlog and see why cancellations can improve a headline without improving service.
A fulfillment dashboard shows whether each warehouse is keeping up with orders, picking them correctly and shipping against its commitments. Reconcile the open-order balance first, then compare quality and timeliness using the appropriate event counts.
The backlog is a stock at a cutoff. Shipments and picks are activity during a period. A review can display them together, but it should not divide them by whichever total happens to be nearby.
For the example below, a fictional US distributor, Central States Supplies, reviews Dallas and Columbus for September 2026. Each order is handled by one warehouse, and an order leaves backlog only when shipped in full or canceled. There are no transfers, partial shipments or reopened orders in this example.
Pick accuracy is measured on all first-pass pick tasks inspected during September. On-time shipment is measured on orders shipped in September against their original promised ship date. That completed-order measure does not include still-open overdue orders; display those separately in a real review so delays cannot disappear simply because the orders remain unshipped.
Actual customer delivery is a later event. The comparison of OTIF, fill rate and on-time delivery explains those neighboring service measures. Here the shipment register supports a warehouse shipping measure, not a customer delivery claim.
Use an opening and closing open-order register, a new-order and cancellation log, a pick-inspection export and a completed-shipment register. Include the warehouse, order ID, order-line ID where relevant, event date, original promised ship date and final shipment timestamp.
Keep the identity of a logistics unit separate from the identity of an order. GS1 US explains that an SSCC identifies a logistics unit, such as cases or pallets packaged for transportation. Several containers can therefore belong to one customer order; counting container records as orders would change the measure.
Preserve the inspected pick-task ID as well. A second inspection after a correction should not overwrite the first-pass result. Before combining files, document the business unit each row represents and check that its identifier is unique for that type of record.
The following counts are the complete summarized inputs and results. Opening orders are open as of August 31; closing orders are open as of September 30.
| Warehouse | Opening orders | New orders | Canceled orders | Fully shipped orders | Closing open orders |
|---|---|---|---|---|---|
| Dallas | 20 | 60 | 5 | 55 | 20 |
| Columbus | 10 | 40 | 0 | 35 | 15 |
| Total | 30 | 100 | 5 | 90 | 35 |
Closing open orders = Opening orders + New orders - Canceled orders - Fully shipped orders
Dallas closes at 20 + 60 - 5 - 55 = 20 orders. Columbus closes at 10 + 40 - 0 - 35 = 15 orders. The company balance is 30 + 100 - 5 - 90 = 35 orders.
That company result must match the distinct order IDs in the closing open-order file. If it does not, look for reopens, moves between warehouses, late cancellations or an extraction cutoff difference. The reconciliation definition describes the principle: explain the movement between balances rather than accepting a plausible-looking total.
The Census Bureau's manufacturing backlog definition similarly connects opening unfilled orders, new orders net of cancellations and sales. This warehouse example applies an operational count version, removing completed shipments instead of sales dollars.
For a USD view, assume every open order in this example has a remaining net selling value of $100. Dallas holds 20 × 100 = $2,000, and Columbus holds 15 × 100 = $1,500. Total backlog value is 35 × 100 = $3,500. In an actual mixed-order business, sum each order's remaining value rather than applying a company average.
The next table contains the pick and shipment counts for the same month. All pick tasks are inspected in this simplified example. Each on-time shipment is a subset of the fully shipped orders in the backlog table.
| Warehouse | Inspected first-pass picks | Correct first-pass picks | Pick accuracy | Fully shipped orders | Orders shipped on time | On-time shipment |
|---|---|---|---|---|---|---|
| Dallas | 150 | 147 | 98.00% | 55 | 50 | 90.91% |
| Columbus | 100 | 98 | 98.00% | 35 | 30 | 85.71% |
| Total | 250 | 245 | 98.00% | 90 | 80 | 88.89% |
Pick accuracy is correct first-pass picks / inspected first-pass picks. The company result is 245 / 250 = 98.00%. On-time shipment is orders shipped on time / fully shipped orders: 80 / 90 = 88.89%. Percentages are rounded to two decimal places after calculating from unrounded counts.
Late shipped orders provide a useful check. Dallas has 55 - 50 = 5, and Columbus has 35 - 30 = 5. Together they give 10 late shipments, matching 90 - 80 = 10. Both sites have the same late count, but Columbus has the lower rate because its shipment denominator is smaller.
The picking result does not explain that difference by itself. Ask the Columbus warehouse manager to inspect the late order IDs and their promise dates. Carrier handoff, inventory availability, order release and labor coverage are possible explanations to investigate, not causes established by the percentages.
When the problem is items that were not supplied, use the narrower lost-order-lines worked example. A whole fulfillment review needs the backlog and event checks as well as that line-level analysis.
Now change only Dallas cancellations from 5 to 10. Opening orders, new orders, picks and completed shipments stay as shown above. The additional cancellations concern unshipped orders.
| Measure | Original | After more cancellations | Calculation after change |
|---|---|---|---|
| Dallas canceled orders | 5 | 10 | 5 + 5 |
| Dallas closing open orders | 20 | 15 | 20 + 60 - 10 - 55 |
| Company closing open orders | 35 | 30 | 30 + 100 - 10 - 90 |
| Company backlog value (USD) | 3,500 | 3,000 | 30 × 100 |
| Company pick accuracy | 98.00% | 98.00% | 245 / 250 |
| Company on-time shipment | 88.89% | 88.89% | 80 / 90 |
Backlog falls, but the warehouse ships no additional orders and its quality result does not change. The commercial owner should review why customers canceled before anyone describes the smaller backlog as a service improvement. Keep cancellations visible as a movement in the bridge.
Bring a sample of your own order, picking and shipment files to Covirage. Covirage maps the columns, and its tools compute and check the figures; the external AI model explains the results with every figure traced to its rows. Create a free Covirage account and bring the exports behind your warehouse review.
See Covirage for supply-chain teams for the industry context.
Start with ending backlog, overdue open orders, first-pass pick accuracy and on-time shipment. Give each metric its count, denominator and cutoff. Add delivery performance only when the files include actual customer delivery events and the relevant promised delivery dates.
For an order-count view, add new orders to opening open orders, subtract cancellations and subtract orders completed during the period. Match the result to the closing open-order register. Partial shipments require order-line quantities or a clear rule for when an order becomes complete.
On-time shipment compares the shipping event with a promised shipping date. On-time delivery compares customer receipt with a promised delivery date. Transit delays can separate the results, so a warehouse shipment record cannot establish whether the customer received the order on time.