Value Stream Mapping
Show material and information flow, then distinguish process time from elapsed lead time.
Map one product family from request to delivery
Map one product family from request to delivery
A value stream map (VSM) shows the steps and information that move a product or service toward the customer. Define the boundary and select work following a similar route. Walk the current process and record what happens, including queues and handoffs.
Show who requests work, how instructions reach production, where material moves and where it waits. An operations-only map misses the information controlling those operations.
Read process boxes, inventory triangles and flow arrows
Read process boxes, inventory triangles and flow arrows
This original sketch uses common conventions: a rectangle for a process, a triangle marked I for inventory, attached data boxes for measurements and arrows for flow. The legend distinguishes material movement from information. This is a teaching sketch, not a complete symbol standard.
Measure elapsed lead time separately from process time; Interactive; Practice
Measure elapsed lead time separately from process time
Lead time is elapsed time through the stated boundary, including waits. Process time is time actively spent doing the work within that boundary. Not all processing adds customer value.
Transfer size changes when jobs finish
10 jobs; A = 4 min/job, B = 6 min/job. Compare transfer-all with transfer-one.
Transfer all 10 jobs
First item 46 minFull order 100 minProcessing / job 10 min
Transfer 1 at a time
First item 10 minFull order 64 minProcessing / job 10 min
A is upper blue, B lower gold; bars are jobs, dashed lines completed-batch transfers. Minutes from order release 0; both plots share a scale.
Job 10 with transfer-one: 36 min waiting before A + 0 held for batch + 18 waiting at B = 54 min waiting. Add 10 min processing: 64 min elapsed from order release 0.
Ideal schedule: all equal jobs arrive at 0; serial single-job A then B, first come first served. A never blocks; waiting space is unlimited. Completed batches transfer immediately; the last may be smaller. Constant times; no setup, transport, failure, rework or variation. Finite-order illustration, not a factory forecast or a universal best batch.
Default 10-job example: exact transfer-one schedule
Fixed default example: 10 jobs; A = 4 min/job, B = 6 min/job; transfer one job at a time. This table does not update with live settings.
| Job | A start–end | Transfer | B start–end | Wait before A | Held for batch | Wait at B | Total wait |
|---|---|---|---|---|---|---|---|
| 1 | 0–4 | 4 | 4–10 | 0 | 0 | 0 | 0 |
| 2 | 4–8 | 8 | 10–16 | 4 | 0 | 2 | 6 |
| 3 | 8–12 | 12 | 16–22 | 8 | 0 | 4 | 12 |
| 4 | 12–16 | 16 | 22–28 | 12 | 0 | 6 | 18 |
| 5 | 16–20 | 20 | 28–34 | 16 | 0 | 8 | 24 |
| 6 | 20–24 | 24 | 34–40 | 20 | 0 | 10 | 30 |
| 7 | 24–28 | 28 | 40–46 | 24 | 0 | 12 | 36 |
| 8 | 28–32 | 32 | 46–52 | 28 | 0 | 14 | 42 |
| 9 | 32–36 | 36 | 52–58 | 32 | 0 | 16 | 48 |
| 10 | 36–40 | 40 | 58–64 | 36 | 0 | 18 | 54 |
Bounded model reference: scheduling equations and limits
These equations describe the bounded ideal model; they do not report the current live settings.
For job i: A start = (i − 1)tA; A end = i × tA; transfer = min(ceil(i / batch) × batch, N) × tA. B start is the later of transfer and previous B finish; B end = B start + tB. Waiting before A = A start; held for batch = transfer − A end; waiting at B = B start − transfer. The three waits plus tA + tB equal that job’s completion time from order release 0.
Bounds: N 1–20 jobs, transfer 1–N, station times 1–20 whole min/job. This original scheduling example is separate from the map’s fixed 120-minute queue. It does not diagnose a factory or identify a universal optimal batch. Real setup, transport, limited space and variation can change the decision; no demand, quality or capacity forecast is made.
The transfer-batch experiment above schedules a whole order. The earlier map sketch uses a fixed queue for one item:
For one item in this simplified sequential example: cutting = 4 min, queue = 120 min, assembly = 6 min. Process time is 10 min; lead time is 130 min. No other delays or overlapping steps are included. Removing half the queue saves more elapsed time here than cutting one minute of processing.
A queue suggests a question, not a proven cause
A queue may reflect a capacity constraint, batching, missing information or an unreliable next step. Observe demand, timing and variation before calling one operation the bottleneck. A working day and a calendar day are not interchangeable time units.
Draw a future state and test a specific change
Draw a future state and test a specific change
A future-state map describes an improved flow to work toward. Choose a measurable change, assign responsibility and check the effect using Kaizen and Standard Work. Compare delivery, quality and workload across the whole stream.