Picture a mid-size metal fabrication shop. Two years ago the company installed an MRP (Material Requirements Planning) system. During setup, the implementation team entered a set of numbers: how many units to order at a time, how many days it takes for raw material to arrive from suppliers, and how much scrap to expect during production. The system has been running ever since. But lately the warehouse is overflowing with some parts while others run out before the next delivery arrives. The production manager cannot figure out why. The software is not broken. The numbers inside it are.
An MRP system calculates production orders and purchase orders automatically. To do this it relies on three core parameters. The first is lot size: how many units are ordered or produced in a single run. The second is lead time: how many days it takes for a material to arrive from a supplier or move through a production step. The third is scrap rate: the expected percentage of material lost during production. These three numbers are entered by hand during setup. The software treats them as correct and builds every plan on top of them.
The problem is straightforward. Everything in a factory changes, but these numbers often do not. A supplier may have shortened its delivery window. Or the opposite — after recent disruptions, lead times have stretched. A machine overhaul may have cut scrap rates significantly. Lot sizes that made sense two years ago may no longer match current order volumes. The software keeps using the old numbers. The result: the system orders too much of one material and too little of another. The warehouse fills with excess stock while production lines wait for parts that should already be there.
Think of it like a recipe that calls for four eggs. If you are cooking for two people, four eggs is too many. If you never adjust the recipe, you waste ingredients every single time. MRP parameters work the same way. Every number that no longer matches reality produces a small error in each planning cycle. One cycle, the error is invisible. Across hundreds of part numbers, dozens of suppliers, and months of production runs, the errors stack up into real money and real delays.
The fix does not require anything complicated. A review every six months is enough to catch most drift. All it takes is a spreadsheet and a short meeting with the production lead, the purchasing officer, and the warehouse supervisor. For each major material, ask three questions: How many days did deliveries actually take over the last six months? What was the real scrap rate in production? Has our average order quantity changed? Compare the answers to what is currently in the system. Where there is a gap, update the parameter.
Measuring the impact of a parameter update is also straightforward. Before the MRP system opens production orders, it generates a suggestion list. Look at that list: how many units did the system recommend ordering, and how many did you actually open? A large and consistent gap between the two is a clear signal that the parameters are out of step with reality. Some MRP packages include a comparison report for exactly this purpose. But a printed suggestion list reviewed against actual orders works just as well. The important thing is to look regularly and write down what you find.
One practical note on lead times: do not update the parameter based on a single delivery. If a supplier happened to deliver in two weeks last month, that does not mean two weeks is the reliable average. Pull the last six months of delivery records, calculate the average, and add a small buffer on top. Apply the same logic to scrap rates. One production run is not enough data. Use several months of actual production records to calculate a realistic figure. The closer the numbers in the system are to what actually happens on the floor, the more accurate every plan the system produces will be.
For any company that has invested in MRP software, the most important decision after go-live is not to forget the parameters. The installation is done, the system is running, everyone breathes easier — and the setup screen is never opened again. This is a very common pattern. No matter how capable the software is, if the numbers inside it have grown stale, production planning will drift off course. A half-day review every six months prevents that drift. Companies that make this a regular habit find that warehouse costs come down and production stoppages become less frequent. The value of MRP software does not come from the installation. It comes from running the system on numbers that still reflect how the factory actually works.
This article was originally written in Turkish by Gökhan MERCANOĞLU on May 3, 2004 and has been automatically translated into English and other languages using machine translation.