Why no local ERP has a real WMS in its core
Inventory accounting answers how much you have; a WMS answers where it sits and who touches it next — and those two questions demand different data models.
- Published
- Author
- Konis Software
Every ERP in this market has a warehousing module: goods receipts, delivery notes, transfers, stocktakes, lots and expiry dates, FIFO or weighted average, inventory-to-GL reconciliation. It is a mature layer — without it there is no defensible COGS. Trouble starts when a company with four thousand pallet positions asks where lot 240118 stands, and the shortest route to pick seventeen lines across four sales orders. Inventory accounting has no answer; it was never built to have one.
Two questions that look like one
Inventory accounting answers how much. A WMS answers where, and who is next. The first is bookkeeping: it must be correct at the end of the day, in the stock card and the general ledger. The second is operational: it must be correct in the second a picker stands in front of a rack, terminal in hand. Two different contracts about accuracy.
That is why the models diverge. Inventory accounting keys stock on item – warehouse – lot – ownership. A WMS keys it on item – warehouse – zone – location – licence plate – lot – status. Not extra columns: the second key introduces entities the first lacks — a position with capacity and rules, and a carrier that moves as one unit.
| Question from the floor | Inventory accounting | WMS |
|---|---|---|
| How much of item 10-4471? | Exactly | Exactly |
| Where is it standing? | Warehouse level | Location and carrier |
| Which pallet ships first? | FEFO by lot | FEFO by lot, plus access and equipment |
| Where do 22 inbound pallets go? | Out of scope | Directed put-away by zone and capacity |
| How far will the picker walk? | Out of scope | Pick path and task sequence |
| Who took from A-04-2-3 at 11:42? | No concept of location | Task, operator, timestamp, carrier |
| Will the pallet fit? | No concept of capacity | Weight, volume, position type |
A bin is not a field, it is a dimension of stock
This is where most retrofits break. Location gets treated as a missing attribute: add a column to the stock card and move on. But location is not an attribute of goods, it is a dimension of stock. The same item in one warehouse exists at once in eight positions, three lots and five pallets, with different statuses. A reservation is made against a quantity in a place, and the sum of those rows must equal one number on the stock card.
So introducing bins after the fact is not a feature, it is a migration. The primary key of stock changes, every transaction that moves it changes, every report that reads it changes, and stocktaking with them. The incumbents lack it not because nobody built the screen, but because a twenty-year-old inventory ledger cannot be opened to that depth without rewriting its core.
Directed put-away: the system picks the slot
Without a WMS the forklift driver picks the slot from memory and habit. That is cheap with three aisles and one driver who has worked there nine years. At scale you pay twice: on put-away, and again on every subsequent pick from a badly chosen slot. Directed put-away weighs storage regime and zone, ABC class by velocity, capacity by weight and volume, and lot-mixing policy.
- 1
Advance notice and receipt
The shipment is announced (ASN) or the receipt is opened against a purchase order. GTIN, lot, expiry and quantity are scanned at the dock; discrepancies are resolved there, not three days later once the goods sit in the rack.
- 2
Carrier creation
Received goods are bound to a carrier — a pallet with its own SSCC. From that point the system tracks a specific pallet, not an abstract quantity.
- 3
Slot proposal
The rule proposes a position: zone by regime (cold chain, excise, bonded stock under a customs declaration), ABC class, free capacity, lot-mixing. The proposal is explainable — the operator sees why this slot.
- 4
Confirmation and posting
The operator scans the position. If it differs from the proposal, policy decides whether to accept it, but the deviation is recorded with operator and timestamp. Stock posts on confirmation, never on the proposal.
Wave, batch and zone are not synonyms
In tenders the words get used interchangeably. They are chosen by order profile: fifty orders a day with ten lines each and a thousand orders with two lines are different problems on identical racking.
| Strategy | When it fits | What the system must provide |
|---|---|---|
| Discrete (order by order) | Few orders, many lines, project deliveries | Locations and walk sequence |
| Batch | Many orders with few lines (e-commerce, pharmacy) | Grouping by item, sortation back to orders |
| Zone | Cold room, small parts, excise goods | Zone hand-off and consolidation |
| Wave | Waves tied to routes, carriers and cut-off times | Wave planning, priorities, labour capacity |
| Wave + zone + batch | Distribution centre | All of it, plus replenishment that never lags |
Replenishment is a precondition, not an add-on
Zone picking without pick-face replenishment fails on day one. The pick face empties faster than anyone notices; the picker arrives at an empty slot, calls the shift lead, a forklift goes to the reserve rack, the wave stalls. Replenishment must be a task the system raises itself, from min/max thresholds and from what the next wave will demand. This is why a WMS cannot be bought in slices: cross-docking without an ASN, waves without replenishment, put-away without capacity — each is half a mechanism.
Cross-docking: goods that should never see a rack
Of twenty-two inbound pallets, six are already sold and leave on tomorrow's route. The classic flow puts them away and retrieves them next morning: two forklift movements, two scans, one position occupied for no reason. Cross-docking is simple as an idea and demanding in its preconditions — an ASN that matches open sales orders, an on-order state that can be reserved, a shipping dock modelled as a real location with a queue, and accounting willing to accept that goods crossed the warehouse without ever holding a position. That last part is the hard one: books and floor must agree without a single put-away record.
Cycle counting: how to stop closing the warehouse
The December stocktake is memorable: the warehouse stops, people count against lists showing the expected quantity, and — reliably — find exactly what the list says. The shortages surface in March, when nobody can say when they happened. Cycle counting inverts it: a few positions every day, no shutdown.
- ABC classification by velocity and value — the A class is counted far more often than the C class.
- The task is raised per location, not per item — you count a place, not a code.
- Counts are blind: the terminal hides the expected quantity, because a displayed quantity is an answer most people copy.
- A variance does not open a correction, it opens a second count, ideally by a different operator.
- Only a confirmed variance becomes a document with a reason, a posting entry and an approval trail.
SSCC: the number a carrier has and a quantity does not
The SSCC is an eighteen-digit GS1 identifier of a logistic unit, printed in GS1-128 symbology under application identifier (00). Its point is clearest next to the GTIN: the GTIN says what is inside, the SSCC says which pallet this is. Without it the packaging hierarchy — pallet, carton, each — is not data but an understanding between the people on shift.
- Loading closes by scanning carriers instead of ticking forty lines off a delivery note.
- A complaint attaches to a carrier and its history: who packed it, from which locations, in which wave.
- Retail chains demanding a GS1-128 pallet label are stating a condition of entry, not asking for a service.
- Electronic delivery notes lean on the same identifier; a system without carriers has nothing to send.
Why it switches on per warehouse
A typical distributor runs five warehouses, not one: a distribution centre that needs a WMS, a spare-parts store with two racks, a packaging store, consignment stock at a customer, a returns quarantine. If the WMS is a global switch, the packaging store inherits zones, put-away tasks and cycle counts it has no use for — so nobody turns it on at all. WMS-managed must be a property of the warehouse, not the installation.
Technically, inventory accounting stays canonical at item – warehouse – lot, and the WMS adds a layer beneath it — location and carrier — whose sum must equal the stock above. Enabling it means defining zones and positions, one opening count by position, and a switch. Nothing above changes: the same receipt, the same posting, the same costing, the same reports.
A WMS that requires a migration is not a layer, it is a second system. The sum across locations must be the stock card — not a report that is periodically reconciled against it.
Which is why in NG One this is one system, not two. Inventory accounting runs in full — all documents, lots and serials with expiry and FEFO, the available/reserved/on-order/in-transit/blocked states, backdated recosting, consignment. The WMS described here sits in the same core, underneath it: bin locations and zones with capacities, directed put-away and picking with pick paths, wave and batch, replenishment, cross-docking, handheld tasks, cycle counting, packing with SSCC, warehouse KPIs. Location and carrier are in the stock model from the first line — a layer that switches on per warehouse, not a column added the week a customer first asks for a GS1-128 pallet label. So the distribution centre gets waves and directed put-away while the packaging store stays a receipt and a delivery note — one installation, one stock card.