ERP for manufacturing: MRP, OEE and the shop floor explained

Factories suffer from scattered information more than almost any other business. Sales promises delivery dates without seeing whether the machines are free, planning can’t see what’s in stock, and finance works out what a product cost weeks after it was made. ERP for manufacturing closes those gaps. This guide explains, in plain terms, what a manufacturing ERP needs to do — from product recipes to machine efficiency — and how to start without a risky all-at-once project.

What a manufacturing ERP must do

Sales order →Material planning →Production order →Factory floor →Finished goods →Real cost

1. Know exactly what each product is made of. This list of parts and materials is called a bill of materials (BOM). A good ERP keeps versions of it, and handles parts that are themselves made of other parts. It also records the steps and machines each product goes through (the routing).
2. Plan materials. MRP — material requirements planning — turns orders and forecasts into dated suggestions: buy this, make that, by this date.
3. Plan machine time. MRP on its own assumes machines can always do the work. MRP II and advanced planning tools (APS) check that the machines and people actually have the time.
4. Run and record production. Production orders, materials used, how much was made and how much was scrapped.
5. Measure. Machine efficiency (OEE), why machines stopped, and quality.
6. Work out real cost. What each production order actually cost compared with what it should have cost, passed straight to accounting.

MRP in one worked example

A factory must deliver 400 air-conditioners by the end of week 4. The numbers here are made up to show the method. Each unit needs:

• one compressor (bought in): 380 in stock, takes two weeks to arrive;
• one metal housing (made in-house in one week): 120 in stock. Each housing uses 2 kg of steel sheet; there are 200 kg of sheet in stock, and new sheet takes one week to arrive.

MRP suggests:

Notice the steel. It is worked out only for the 280 housings that still need to be made (280 × 2 kg = 560 kg, minus the 200 kg in stock), not for all 400 units — because 120 housings were already on the shelf. Working down through every layer of parts like this is exactly what spreadsheets can’t do reliably once you have hundreds of items.

OEE: measuring the factory floor

OEE — overall equipment effectiveness — shows how much of the time you planned to run a machine was actually spent making good parts at the right speed. It is three percentages multiplied together:

OEE = availability × performance × quality

• Availability: of the planned time, how much was the machine actually running?
• Performance: while it was running, how fast was it compared with its best speed?
• Quality: of everything it made, what share was good?

Example for one 480-minute shift with 30 minutes of planned breaks, 45 minutes lost to breakdowns and mould changes, a best speed of one part every 30 seconds, 750 parts made and 15 rejected:

The 85% “world-class” figure is linked to Seiichi Nakajima, who founded the TPM approach to maintenance. But it was meant for factories that make the same parts over and over. Use it as a rough guide, and compare each machine with its own past results.

ERP and MES: who does what

In practice the two must share information both ways: the ERP sends production orders down to the factory floor, and the MES sends back what was really used, how long it took and how good it was.

Quality and tracing every batch

In regulated industries such as food and pharmaceuticals, you must be able to trace a batch of raw material forwards to every customer who received it, and backwards to every supplier it came from. This is not optional — it is what lets you recall a product quickly.

Two more tools help:

• Statistical process control (SPC): watching measurements such as temperature or weight, so you notice when a process starts to drift before faulty products are made.
• Quarantine rules: batches that haven’t passed quality checks are blocked from being shipped.

See food & pharma.

How to start

1. Tidy your basic data first: product recipes, routings, and realistic delivery times taken from your actual purchase history, not suppliers’ promises.
2. Count all your stock before you start. Don’t copy stock figures over from the old system.
3. Try it on one product family or one line. For a few weeks, compare MRP’s suggestions with what your planner would have decided.
4. Measure OEE on your bottleneck machine — the slowest machine, which limits how much the whole line can make — using a short, specific list of reasons for stopping.
5. Expand once the data and definitions have settled down.

Fidar for manufacturers

Today Fidar offers CRM, reporting and detailed access control. These are useful for managing industrial sales and building management dashboards. Manufacturing & MRP and MES & Quality are coming soon, following the roadmap after accounting, purchasing and inventory. Their design includes live OEE, direct connection to machines, SPC and full batch tracing. If you would like those features shaped around a real factory, tell us about your plant on the Manufacturing module page.

Data checklist before you go live

Start with your most important items — the most valuable ones, or the ones that take longest to arrive. Getting those right has the biggest effect on the plan.

Make-to-stock, make-to-order and process manufacturing

• Make-to-stock (you build ahead and sell from the shelf): MRP runs from a sales forecast and a production plan. Safety stock and reorder points matter most.
• Make-to-order (you build only what customers order): each order creates its own needs, and every purchase should stay linked to the order it is for.
• Process manufacturing (food, chemicals, pharmaceuticals): recipes instead of parts lists, items counted in two units (for example pieces and kilograms), by-products, and quality checked batch by batch. Efficiency is measured by output rate rather than number of pieces.

Many factories do all three, so check that the ERP supports each way of working rather than forcing one model on every product.

What to measure in the first six months

Summary

ERP for manufacturing links the order book to the factory floor. Accurate product recipes feed material planning, checking machine time keeps plans realistic, and data from the floor feeds efficiency figures and real cost. Start with tidy data and one line, and measure before you expand. For the bigger picture, see what is cloud ERP.