Before the end of the month when finance closes the books, the warehouse supervisor dreads hearing one sentence most: “There are still 3,200 items in the system, but only 2,980 on the physical inventory.” The discrepancy of 220 items may be scattered across three black holes: unposted inbound orders, materials issued but not yet deducted from inventory, and inventory count discrepancies awaiting approval . When inbound and outbound records don’t match, it seems like a warehouse issue, but the root cause is often that business workflows and system flows are out of sync —goods arrive at the warehouse but haven’t been entered into the system; materials have been picked but the outbound order remains in the draft box.
Why can’t Excel or paper documents handle inventory accounting?
Many companies use ERP systems to manage financial inventory, yet on the warehouse floor they rely on paper inbound slips, manual ledgers, and WeChat group reports for inventory counts . This approach might work when SKUs are few and turnover is slow, but once faced with situations such as suppliers delivering in batches, production over-issuing materials and later making up the difference, urgent sales orders requiring expedited shipping, outsourced material shipments and returns, or simultaneous returns and exchanges, the gap between book and physical inventory grows exponentially.
Typical scenarios: goods arrive at the warehouse but aren’t recorded —a supplier delivers to the dock, warehouse staff unload and put items on shelves first, then enter them into the system after quality inspection passes; meanwhile, sales queries show stock available, but actual shipment isn’t possible. Outbound shipments aren’t deducted —the production line urgently needs materials, so warehouse staff ship first and update the order later; if the correction isn’t completed that day, the inventory figure appears artificially high. Unit conversion errors —procurement records quantities by “box,” while production picks by “piece,” and BOM usage doesn’t align with packaging specifications in the system. Inventory counts become mere formalities —at year-end, only large-value items are counted, leaving edge-of-line storage areas and inspection zones permanently “outside the scope.”
How to break down operations: inbound, outbound, and inventory counting form three main threads
Inbound operations
Inbound processing isn’t a single step—it’s a chain of arrival confirmation → quality inspection determination → shelving and placement → financial posting . Classified by source: procurement inbound, production completion inbound, outsourced return inbound, return inbound, and transfer inbound. Each source has different triggering conditions and responsible roles : procurement inbound typically requires matching PO lines; production inbound must link to work orders and time sheets; return inbound needs to tie back to the original sales outbound order.
Design highlights: no entry until goods arrive, no inclusion in available inventory until quality inspection passes . The system should distinguish five inventory statuses—“in transit,” “awaiting inspection,” “available,” “frozen,” and “defective”—rather than just having a single “inventory quantity” field.
Outbound operations
Outbound processes also vary by scenario: sales shipments, production material requisitions, outsourced material dispatches, transfers, and scrap disposals. Core rule: reserve available inventory before picking items (reservation/allocation mechanism) to avoid overselling or double-picking the same batch. Production material requisitions should automatically suggest required quantities based on the BOM, allowing over‑picking but requiring approval and traceable documentation.
Inventory counting operations
Inventory counts fall into two categories: cycle counts (high-frequency checks for fast-moving SKUs) and full inventories (year-end). Crucially, during inventory counts, all changes must be frozen : either lock locations to prevent entries or exits, or adopt a “snapshot + discrepancy adjustment” model. Any inventory differences require approval: classify reasons for surpluses or shortages (data entry errors, loss, damage, unit conversion), then generate adjustment orders rather than directly altering inventory figures without proper documentation.
How to design: location, batch, and transaction models
Locations and containers
Inventory tracking goes beyond just recording “material + quantity”; it must also note where it is located : warehouse → storage area → location → container (pallet/box). Picking routes, FIFO principles, and batch traceability all depend on granular location data. Edge-of-line storage, inspection zones, and defective-item areas must all be integrated into the same location system; otherwise, the problem of “on the books but nowhere to be found” will persist.
Batches and serial numbers
Materials with shelf lives or quality‑traceability requirements must be managed using lot numbers or serial numbers . Bind supplier batches upon receipt, and allocate according to FIFO or designated batches upon dispatch. Clearly define during design: which materials follow batch tracking, which use serial numbers, and which simply record quantities.
Inventory transactions must remain tamper‑proof
All inventory changes must be logged through transaction records (inventory_transaction) : type of transaction, material, quantity, opening and closing balances, associated documents, operator, timestamp. Direct updates to inventory tables are prohibited; any adjustments must generate a transaction record to provide audit and reconciliation evidence.
How to develop and implement: data collection, interfaces, and reconciliation
On-site data collection
Barcode/RFID scanning is mainstream: scan upon arrival → scan during quality inspection → scan when shelving and placing items → scan when picking and issuing items . Mobile PDAs or mini-programs are ideal for warehouse staff inspections; integration with electronic scales and label printers enables automatic printing of location tags. Principle: complete binding of “material + quantity + location + batch” with a single scan.
Interfaces with ERP and finance
Division of labor between WMS and ERP: WMS manages physical goods and locations , while ERP handles financial inventory and costs . Timing of postings must be agreed upon: ERP triggers receipt when inbound inventory becomes “available”; ERP triggers write‑off and cost allocation when outbound shipments are “confirmed for dispatch.” Avoid letting WMS and ERP operate independently—nightly reconciliations are merely remedial measures, not solutions.
Mechanism for reconciling book and physical inventory
- Daily cleanup : all inbound/outbound/transfers/adjustment orders for the day must be closed; pending ones go into an exception list.
- Age‑based and stagnant‑stock alerts : batches that haven’t moved for N days automatically turn red, reducing inventory left forgotten in corners.
- Cycle counts : set frequency based on ABC classification, and trigger a review process if discrepancy rates exceed thresholds (is it a data entry error or a workflow flaw?).
Acceptance criteria
- Any SKU can be traced within 10 seconds to check its available quantity, items awaiting inspection, distribution across locations, and the last five transactions . During sales order placement,
- promised quantities can match the actual amount available for picking (with allowances for safety stock deductions). Monthly inventory discrepancy rates must stay within company targets (e.g., below 0.5%), and every discrepancy must have approval and a clear reason code.
Common pitfalls
Only implementing WMS without changing workflows : on the ground, staff still ship first and update later, rendering even the most perfect system useless. Ignoring edge-of-line storage : production materials reach the line side but then “leave the system,” causing mismatches when completing inbound procedures. Confused master data on units : multiple measurement units for the same material lack conversion relationships. Not locking inventory during counts : halfway through the count, new discrepancies appear, leaving differences forever unexplained.
To fundamentally solve the mismatch between book and physical inventory, each step—from “arrival, inspection, shelving, allocation, picking, shipping, counting, and adjustment”—must be transformed into a verifiable system event , rather than relying on human memory or post‑hoc corrections.
Shandong XYN Information Technology Co., Ltd. (XYN Tech) delivers customized software related to warehousing and supply chain for industries such as manufacturing, trade, and logistics, covering WMS design, on-site data collection, ERP integration, and inventory reconciliation mechanisms. For more details, please see About Us.