Where ERP and MES stop: the shop-floor gap

An ERP holds the commercial and financial truth of the plant. It knows what was planned, what was ordered, what was booked and what it cost. Its unit of work is a transaction, and its timescale is the accounting period.

An MES holds the machine truth. It knows what ran, for how long, at what rate, and against which order. Its unit of work is a production order, and its timescale is the shift.

Both are correct. Neither was designed to be the place where a dispatcher compares three carrier quotes for the same lane, a supervisor authorises a write-off, a technician closes a maintenance job at the machine, or a picker is stopped from taking stock an earlier delivery still needs.

Transport is priced by hand. Carriers, quotes and compliance paperwork are managed in email and spreadsheets, and the cost per route only becomes visible when the invoice arrives, at which point nothing can be changed.

Material loss has no cost attached. Scrap and rework are recorded somewhere, but nobody can put a figure on a single event, so nobody can rank which loss is worth fixing first.

Maintenance is reactive. Equipment health is nobody's dashboard, so the first signal that something needed attention is the line stopping.

Stock is allocated by whoever reaches the shelf first. Two deliveries need the same part in the same week, and the earlier one finds out at the dock that its material is gone.

The work in that gap has three characteristics. It happens at the point of action rather than at a desk, it needs a decision from a named person, and it has to be recorded in a way that survives an audit. Software that fills the gap therefore has to be usable at a scanner or on a phone, has to enforce who may do what, and has to version its records rather than overwrite them.

It also has to complement rather than replace. Material master data, cost centres and project structures already exist in the ERP, and duplicating them creates two truths. The right integration treats those as first-class inputs and carries data back through an open API, so the ERP stays the system of record for what it is good at.

If your ERP already prices transport lanes and produces a defensible carrier comparison, you do not need a separate transport layer. If your MES already captures defects at the work cell with a cost attached and an authorisation chain, you do not need a separate material-loss system.

In our experience most plants have one or two of these covered and the rest running on spreadsheets. The useful question is not which system is better, but which of the four gaps above is actually costing you money this quarter. That is also why we suggest starting with one area rather than a platform rollout.