Production managementReading time 11 min
The seven wastes, and how to actually spot them in your own plant
Waste is hard to remove because most of it looks like work.
01The seven wastes
| Waste | What it looks like in a shop | How to measure it |
|---|---|---|
| Transport | A semi-finished part crosses the plant three times to reach the next station | Travel distance, number of moves |
| Inventory | WIP stacked in the aisle, finished goods sitting two years | WIP value, age distribution |
| Motion | The operator walks to the far end for a gauge, every single time | Operator walking distance, retrieval time |
| Waiting | Waiting on drawings, material, inspection, a supervisor's signature | Waiting hours as a share of total |
| Overproduction | Fifty extra pieces made to round out a batch | Quantity released ÷ quantity ordered |
| Overprocessing | Polishing a non-functional face to Ra 0.8 | Compare against what the drawing asks for |
| Defects and rework | Reworking a dimension that went out of tolerance | Defect rate, rework hours |
02How to see it
Value stream mapping
Map one representative product's full path from order to shipment, marking process time and wait time at every station, then compute the value-added ratio = process time ÷ total lead time.
- In most shops that have not been through improvement work, that ratio lands between 0.5% and 5%.
- Which is to say: 99% of the time, the part is waiting.
- The room for improvement is almost entirely in the waiting, not in the cutting speed.
Spaghetti diagram
Take the plant floor plan and trace one product's, or one operator's, movement over a day. Lines tangled into a knot mean a layout problem. It is the most intuitive and cheapest diagnostic there is.
Genchi genbutsu
Stand on the floor and watch one complete work cycle. Do not read a report. Toyota's version is to draw a circle and stand in it for thirty minutes, the waiting and the repeated motions never show up on a report, and only appear if you are standing there.
03The matching tools
| Problem | Tool | Effect |
|---|---|---|
| Long changeovers force big batches | SMED | Setup time often drops by more than half |
| Hunting for tools takes forever | 5S | Motion and waiting waste |
| Too much WIP | Pull production / kanban | Inventory and lead time |
| Defects escape to the next station | Jidoka, mistake-proofing | Defects and rework |
| Layout drives transport | Cell production | Transport and waiting |
| The same problem keeps coming back | 5 Why、A3 | Root cause never addressed |
The core idea of SMED is to split changeover into internal work (the machine must be stopped) and external work (can be done while it runs), move everything that can be externalized out first, then simplify what internal work remains. The first step alone typically cuts the time in half.
04Common traps
- Chasing machine utilization. Loading a non-bottleneck to the ceiling only manufactures WIP. See production scheduling.
- Doing 5S for the photographs. The place looks immaculate, the process never changed, and three months later it is back to how it was.
- Rolling it out everywhere at once. Pick one line as a model, get a result, then spread it, the success rate beats a plant-wide launch by a wide margin.
- Treating improvement as cost reduction. The goal is shorter lead time. Cost falling out of it is a result, not the method.