The Brain Behind the Warehouse

The WMS connects three records. The item master describes the product: its dimensions, weight, packaging, and any lot or serial tracking. The location master describes the places it can occupy. The inventory record says how much of that item is in a particular location, who owns it, and whether it can be used.

A SKU, stock keeping unit, identifies a distinct item. A medium blue shirt and a large blue shirt have different SKUs. A lot identifies a batch; a serial number identifies an individual unit. Those distinctions let a business recall one production batch without pulling every unit of the product.

Quantities need units. Twelve cases are not twelve shirts. If each case holds ten shirts, the WMS needs that conversion to receive cases and pick individual units correctly. A wrong case quantity can make every scan look valid while every count is wrong.

A directed picking task might read: go to A-12-03, scan the location, scan the shirt, put two units in tote T-17. The system chooses the task using the order, the stock, and the warehouse's rules.

The scans check identifiers against the task. Confirmation records the quantity moved from the shelf to the tote, with a time and user. The shirts are still in the building; their recorded location has changed.

A barcode doesn't count the contents of a box. Someone can scan the right label and enter the wrong quantity, or move stock without recording it. Voice systems, sensors, and automation can also confirm work. What matters is that physical movement gets a matching transaction, and exceptions can be traced.

A supplier may send an advance ship notice, or ASN, before the truck arrives. It describes what is coming. At the dock, the receiver checks the delivery against that notice and the purchase order, records the quantities actually received, and logs shortages, extras, or damage.

A pallet or tote can have a license plate number, or LPN: an identifier attached to a container and its recorded contents. Scanning it retrieves that record. It doesn't prove the contents match, so the receiving process still needs rules for inspection and discrepancy handling.

Dock-to-stock time measures how long receiving takes to make goods available for use. A pallet waiting for inspection can be physically present without being available to an order. With cross-docking, goods needed for an outbound shipment go toward shipping without a trip through storage.

Putaway moves received stock into storage. The WMS chooses a destination using location rules: suitable temperature, enough capacity, permitted item types, and other constraints.

Slotting decides where products should live to make picking easier. Frequently picked items benefit from accessible locations and shorter routes. Weight, dimensions, congestion, and items often ordered together also matter. A single popularity ranking won't settle all of those tradeoffs.

Full pallets may sit in reserve storage while individual units sit in a smaller forward pick location. Replenishment moves stock from reserve to that pick location, either when it falls below a threshold or when upcoming work needs more. An empty pick location can stop a task even when there is plenty of stock elsewhere.

On hand is the quantity recorded in the warehouse. Allocated stock is reserved for demand. A hold blocks stock from use, perhaps pending inspection or because it is damaged. Reservations and holds are separate controls, even when a screen groups them together.

The recorded quantity also needs a physical check. Cycle counting audits selected locations throughout the year. A discrepancy goes through review, recount, or adjustment according to the operation's rules. Repeated misses in one location are a reason to investigate the process there.

For dated stock, FIFO means first in, first out. FEFO means first expiring, first out: use the stock with the earliest eligible expiration date. A customer may also require enough shelf life remaining at delivery, so the oldest lot isn't always suitable.

An order arrives from the business's sales or order system with items, quantities, a destination, and a deadline. Allocation reserves eligible inventory for it. The WMS may reserve a quantity first and choose the exact stock later, or assign specific locations, lots, or containers.

Release authorizes work to proceed. The operation must fit that work around people, equipment, pack stations, and carrier cutoffs. A wave groups orders for processing; a waveless operation feeds work more continuously. The sequence varies by product. In Oracle WMS, for example, running a wave can create the allocations.

In discrete picking, a worker picks one order at a time. Batch picking combines demand from several orders, then sorts the items afterward. Cluster picking keeps those orders separate during the trip, often in different totes on one cart. Zone picking gives workers parts of the floor, with the order moving between zones or being consolidated later.

Those choices solve different problems. Batching saves repeated visits to a popular item but adds sorting work. Clusters avoid that later sort but limit how many orders fit on the cart. Zones reduce travel within each area but need a reliable handoff.

Workers can follow scanners, spoken instructions, or lights at the bins. In goods-to-person systems, conveyors, cranes, shuttles, or robots bring stock to a station. The WMS still tracks inventory and demand; equipment software controls the movement. Less walking can leave packing or station capacity as the next constraint.

At packing, the items are checked against the order and assigned to containers. Cartonization uses item and container data to propose how to pack them, subject to weight, fit, and packing rules. It may happen before picking rather than at the pack bench.

Box size affects the bill. Parcel carriers can charge by dimensional weight, calculated from package volume, when it exceeds actual weight. Carrier selection and label printing may belong to the WMS, a shipping application, or a transportation management system (TMS).

With a GS1 logistics label, the container's identifier is the SSCC, or Serial Shipping Container Code, commonly carried in a GS1-128 barcode. The ASN associates that identifier with the contents. The number on the label and the number in the message must agree.

A truck can leave while the business system still shows an open order. Check whether the WMS created the ship confirmation, whether it was delivered, and whether the receiving system accepted it. Inventory posting and invoicing depend on that business's setup; printing a label alone doesn't settle either.

A returned shirt needs an identity, a condition assessment, and a decision. It might be restocked, repaired, returned to the supplier, or scrapped. A disposition code records that decision. The physical return and the customer's credit are related events, but one does not prove the other happened.

Stock awaiting a decision can be held in the software and separated physically. A software hold prevents normal allocation when the rules are enforced. It cannot stop someone taking the item off a shelf. Quarantine needs both a recorded restriction and a floor process that respects it.

A bill of materials, or BOM, lists the components required for a product. A work order specifies what to build and how many. The warehouse supplies those parts and receives finished goods. The production system manages the build.

Kitting gathers parts for a build. Line replenishment keeps components available at the production station; a kanban is a signal to supply more, such as an empty bin or an electronic request. Partly completed production is work in process, or WIP.

Component consumption can be recorded directly or inferred. With backflushing, a reported production event deducts the expected components. In the example, reporting 200 finished units could consume 1,200 fasteners. If the line actually used 1,230, the extra 30 need recording too. A correct BOM cannot account for unreported scrap.

A line can stop for want of one component while every other component is abundant. Some plants need parts delivered just in sequence, in the order production will use them. Total warehouse stock is a poor measure of whether the next build can begin.

A WMS shares work with other software. These names describe responsibilities; they don't guarantee a separate product for each one.

Electronic data interchange, or EDI, exchanges standardized business documents. X12's 940 requests warehouse shipping work; the 945 reports the shipment. The buyer's 856 serves a different recipient and purpose. Systems also exchange data through application programming interfaces (APIs) and files.

An integration needs to handle delays, rejections, and duplicates. If a shipment was posted but its acknowledgment was lost, resending the message must not deduct the inventory again. Message IDs, processing logs, and reconciliation are what let teams distinguish a missing event from an event processed twice.

A perfect order arrives on time, complete, undamaged, and with accurate documentation. For a direct measurement, count orders meeting all four conditions and divide by the total orders. Four separate success rates don't show whether the failures hit the same orders.

APQC's measure multiplies the four rates and calls the result an approximation. Four rates of 97 percent produce an estimate of 88.5 percent. As a probability calculation, that assumes independent failures. The demo lets you change that assumption.

For 1,000 orders, a 97 percent rate means 30 failures on each measure. If the same 30 orders fail all four, 970 are perfect. If four separate groups of 30 fail, only 880 are. Both operations report 97 percent on every individual measure.

Other measures answer narrower questions: receiving time, pick accuracy, units per labor hour, or cost per order. Compare like work over a consistent period. A full-pallet pick and a single-item pick are different jobs, even if a dashboard calls both one pick.

A labor management system compares recorded work with planned or engineered task times. Those comparisons need to account for travel, load size, equipment, and delays outside the worker's control. Otherwise, a slow replenishment looks like a slow picker.

Suppose a change saves 20 seconds on each of 10,000 daily tasks. That is about 56 labor hours. At an assumed fully loaded cost of $25 an hour, it represents roughly $1,389 of daily capacity. These are example inputs, not an industry payback estimate.

That capacity becomes a cash saving only if the business can reduce paid hours or avoid adding them. It may instead absorb growth or ease a backlog. Software, equipment, maintenance, and new work elsewhere in the process still belong in the calculation.

A WMS can be a standalone product or part of an ERP. ERP warehouse modules can support directed work and real-time inventory; the label on the product doesn't tell you how capable it is. Evaluate the actual operation: owners, lots, automation, shipping requirements, peak volume, and integrations.

Configuration sets rules the product supports. Customization adds behavior through code. Either can cause failures and need testing after an upgrade. Supported extensions may be reasonable; a heavily modified process carries a larger maintenance burden.

Implementation means mapping the work, cleaning and migrating data, connecting systems, testing, training, and cutting over. Before that cutover, run real examples through receiving to shipment. Include a short delivery, a held lot, a partial pick, a rejected message, and a retry. A happy-path demonstration leaves those questions unanswered.

If a pick location runs empty ten minutes before the carrier cutoff, who gets the replenishment task, can they complete it, and what happens to the order if they can't? A trial on the floor can establish that. A feature list can't.