Smart Factory Logistics vs WMS vs Supply Chain Planning: Where Each One Ends

Most factories have optimised their machines and left the space between them untouched. Cycle times are measured to the second, changeovers are timed and improved, OEE is on a screen in the corridor. Meanwhile the material moving between those machines is coordinated by a person with a radio, a clipboard and twenty years of institutional memory. When that person is on holiday, the factory notices.

What smart factory logistics is — and what it is not

Smart factory logistics is the digitalisation of that layer: the internal movement of raw material, work-in-progress and finished goods between receiving, storage, the line and dispatch. It is distinct from warehouse management, which governs what happens inside the storage area, and distinct from supply chain planning, which governs what arrives at the gate. Internal logistics is the connective tissue, and it is usually the least instrumented part of the whole operation.

The clean way to draw the boundary is by the question each system answers. Warehouse management (WMS) answers where is it and in what quantity inside a storage area — bin locations, FIFO/FEFO, picking, cycle counts. Supply chain planning (SCP) answers what should arrive at the gate and when — supplier lead times, purchase proposals, delivery promises. MES answers what the line is doing right now — the order, the operation, the consumption. Smart factory logistics answers the question none of them own: how material gets from the dock to the bin to the line to dispatch, by whom, and by when. In ISA-95 terms it sits at Level 3 alongside MES, in the material-handling and tracking activities the standard lists and most MES products only half implement.

Line-side starvation: the number most plants don't have

The failure mode it addresses is line-side starvation — a machine that is capable, staffed and scheduled, standing idle because the material is somewhere else. Every factory has a number for this and most do not know it, because the downtime gets coded as "waiting for material" and then aggregated into a category nobody drills into. When you start recording it against specific movements and specific requests, the pattern usually turns out to be concentrated in a handful of routes and shift transitions.

Transport requests replace the radio

Digital transport requests replace the radio. When a line needs material, the request is created in the system — automatically from the production plan and consumption rate where possible, manually where not — and it enters a queue with a priority, a source, a destination and a deadline. Nothing about that is technically difficult. What it changes is that the work becomes visible, measurable and assignable rather than dependent on who shouted loudest.

Dispatching as an optimisation problem

Dispatching then becomes an optimisation problem rather than a memory exercise. The system assigns each transport task to the vehicle that can serve it most efficiently, considering current position, current load, and the other open tasks nearby. The forklift that just dropped a pallet at line three gets the pickup at line four rather than driving back empty to a staging area. Fleet utilisation improves without adding a single vehicle.

RTLS: when live position beats the last scan

Real-time location tracking sharpens this considerably, though it is not a prerequisite for starting. When the system knows where each vehicle actually is — via UWB, BLE or gate-based RFID reads — assignment stops relying on the last scan and starts relying on the current position. Manufacturers running dense internal logistics with many vehicles tend to reach this stage quickly; smaller operations often get most of the benefit from digital requests and dispatching alone.

The standards that keep it interoperable

Three standards decide whether an internal-logistics system talks to the rest of the plant or becomes another island. VDA 5050, the German automotive industry's interface for driverless transport vehicles, lets AGVs and AMRs from different vendors take orders from one fleet controller — the practical condition for adding a second vehicle brand later without replacing the dispatching layer. ISA-95 defines the material lot, sublot and location models that MES and logistics share, so a transport order and a production order refer to the same batch. GS1 labels (SSCC for pallets, GTIN with batch and lot application identifiers) make every scan readable by the customer's warehouse as well as your own. None of them are exciting; all of them are cheaper to adopt at the start than to retrofit.

Four metrics that show whether it is working

The metrics that matter are unglamorous and immediately actionable. Average response time from request to delivery. Percentage of transport tasks completed within the deadline. Empty-run ratio across the fleet. Line stoppage minutes attributed to material availability. None require new hardware to define, and all of them are invisible in a radio-based operation.

Traceability comes for free

Traceability comes along for free, and in regulated industries it justifies the project by itself. Every movement carries a record: what was moved, from where, to where, by whom, at what time. When a quality issue surfaces three weeks later and the question is which batch went to which line on which shift, the answer is a query rather than an investigation.

From responsive to predictive: connecting the production schedule

The connection to production planning is what makes this smart rather than merely digital. When the logistics layer can read the production schedule, material can be staged ahead of demand instead of chased after it. The line does not request the material; the material is already there because the system knew the order was coming. That is the difference between a responsive internal logistics operation and a predictive one, and it is reachable without a single autonomous vehicle.

The sequencing question that follows is the same as for maintenance or scheduling: not everything at once. Digital transport requests and dispatching come first, because they need no hardware beyond the handhelds a WMS already uses and they create the data every later step depends on. RTLS is second, and only where vehicle density or traceability requirements justify the infrastructure. Driverless vehicles are last: an AGV fleet dropped onto a plant that still coordinates by radio inherits every problem the radio had, at a much higher price. A plant that automates in that order gets its empty-run ratio and its material-related line stoppages onto a screen — as reported numbers instead of arguments — before the first autonomous vehicle is ordered.