A plant makes aluminum pump housings. The die casting cell was upgraded last year, machining got new fixtures, assembly was rebalanced. Every project reported a gain. On-time delivery is still 91 percent, and the warehouse still holds four days of finished housings. Nobody did anything wrong. Everybody improved a process, and nobody looked at the stream.
What a value stream is
A value stream is everything that happens to one family of products on the way from raw material to the customer: every process, every stock, every schedule and every phone call that decides what is made next. Value stream mapping puts all of it on one sheet of paper.
The map has three bands:
- Material flow. Every process from receiving to shipping, with a small data box, and the inventory waiting in front of it.
- Information flow. Who tells each process what to make next, how often and by which medium.
- Time line. Days of waiting between the processes, seconds of work within them.
No report shows these three together. The map shows how they cause each other.
The example in this series
All eight parts follow the same family of pump housings in a die casting and machining plant with 95 employees.
| Item | Value |
|---|---|
| Demand | 450 housings per day in two variants |
| Available time | 54,000 seconds per day in two shifts |
| Takt time | 120 seconds |
| Processes | Die casting, trim, CNC machining, leak test, assembly |
| Inventory in the stream | 7,425 pieces |
| Lead time | 16.5 days |
| Value-added time | 7.5 minutes |
A housing is worked on for seven and a half minutes. It is in the plant for sixteen and a half working days. For each minute of work it waits 33 working hours.
Why single improvements do not add up
With numbers like these, the logic of process improvement breaks down. Make die casting ten percent faster and the value-added time falls by a few seconds. The lead time does not move, because the lead time is not made in the processes. It is made between them: in the five days of ingots, the three days of castings, the four days of finished goods.
Those days have causes, and the causes sit in the connections. The casting cell runs each variant once a week, so three days of castings wait for the trim press. Four points in the plant receive their own weekly schedule, so each one makes what its list says, not what the next process needs. A daily shortage meeting corrects all four schedules.
What mapping changes
Three things change when a team maps its stream.
The question. It is no longer how fast a machine runs. It is how long a housing needs from ingot to truck, and what decides that.
The numbers. Inventory is counted on the shop floor and cycle times are measured with a stopwatch. The team agrees on one set of facts, which ends a lot of old arguments.
The choice of improvements. After the current state comes a future state: a design of how the stream should flow in about twelve months. Improvements are then chosen because the design needs them. A shorter changeover in machining is no longer a nice project. It is the condition for running every variant every shift.
The sequence of this series
The series follows the method in its natural order: choose the product family, walk the flow and collect data, draw the current state, calculate the lead time ladder, answer eight questions for the future state, draw it, and turn it into a plan with loops and reviews. In the example, that path leads from 16.5 to 5.0 days with the same machines and fewer schedules.