Operations & process
Value Stream Mapping
A pencil-and-paper lean method for drawing every process step, inventory queue and information flow a product family passes through, so a team can see the whole flow at once, compare value-adding time with total lead time, and design a leaner future state worth implementing.
Also known as VSM, Material and information flow mapping. First set out by Mike Rother and John Shook (Lean Enterprise Institute) in 1998; the primary source is cited in full below.
Where this is contested
The method formalises Toyota's internal material and information flow diagrams; Toyota never called the tool value stream mapping, and the term 'value stream' itself comes from Womack and Jones's Lean Thinking (1996) rather than from Toyota.
- Format
- Mapping
- Level
- Business unit · Product · Team
- Best for
- Plan execution · Prioritise
- Decision stage
- Diagnose · Plan
- Difficulty
- Intermediate
- Time to apply
- Two or three days for a first current-state and future-state pair for one product family; the implementation plan then runs over months.
Plate · The model
The components
Choose the product family
Selection of one family of products sharing similar processing steps and equipment, which defines the scope of the map. The family, not the plant, is what gets mapped.
Signals of strength
Products grouped by shared downstream steps, not by sales category · One family chosen deliberately, with volume and demand noted · Scope set door to door within one facility before extending wider
Map the current state
A first-hand drawing of how material and information actually flow today, in the canonical notation of process data boxes, inventory triangles, information arrows and the bottom timeline comparing lead time with value-adding time.
Signals of strength
Data collected by walking the flow, not from the ERP system · Inventory physically counted at each queue · Information flow drawn, including schedules each process really follows · Timeline shows lead time in days against processing time in minutes
Design the future state
A redesigned map of the same stream built around takt time, continuous flow where possible, pull where flow must break, a single scheduling point at the pacemaker, and levelled production, with kaizen bursts marking the changes required.
Signals of strength
Takt time calculated from real customer demand · One pacemaker process receives the schedule · Supermarkets or FIFO lanes replace push between processes · Each kaizen burst names a specific, necessary improvement
Plan and implement
A dated, owned value stream plan that breaks the future state into loops, sequences the work, and commits to reviews, so the future state is achieved rather than admired.
Signals of strength
Objectives per loop are measurable, with owners and dates · Progress reviewed by someone accountable for the whole stream · The map is redrawn as improvements land
When it earns its keep
- Lead times are long relative to actual processing time and nobody can say where the days go. The map makes the gap visible on one page.
- You are about to launch improvement work and need to choose where kaizen effort will move the whole flow rather than polish an isolated process.
- Departments each optimise their own step, inventory piles up between them, and you need a shared door-to-door picture that cuts across functional ownership.
- You are planning a layout change, new equipment or a pull system and want to test the design of the whole value stream before committing.
And when it doesn't
- The problem sits inside a single process step, for instance a machine's changeover time or defect rate. Use direct study of that process, or a kaizen event, rather than a door-to-door map.
- The work is genuinely non-repetitive, judgement-heavy and unroutinised, such as original design or advisory work. The notation assumes a definable sequence of steps and flows, and forcing one can mislead.
- Demand and mix are so volatile that a single snapshot of flows and inventories would not represent anything. Consider simulation or direct queue analysis instead.
- You need financial rather than flow answers. The map shows time and inventory, not cost or margin; pair it with unit economics or cost analysis for those.
How to run it
Before starting, gather the inputs the analysis depends on:
- A chosen product family, identified from the product-process matrix, so the map follows one coherent flow rather than the whole factory.
- A walk of the actual flow, door to door, collecting current data first-hand: cycle times, changeover times, uptime, batch sizes, staffing and inventory counts at each stage.
- Customer demand data, from which takt time is calculated: average daily requirement and how the customer orders and takes delivery.
- Knowledge of the information flow: how forecasts and orders arrive, how production is scheduled, and what instructions each process actually works to.
- A pencil, an A3 sheet and a small cross-functional team including someone with authority over the whole stream.
- 1
Choose the product family
Group products that pass through similar processing steps and shared equipment, and pick one family to map. Rother and Shook are insistent on this: mapping everything at once produces a wiring diagram nobody can act on. One family, door to door, is the unit of analysis.
- 2
Map the current state
Walk the flow yourself, upstream from despatch, and draw it in the standard notation: a data box under each process (cycle time, changeover, uptime, operators), inventory triangles with counted quantities between processes, the customer and suppliers along the top, and the information flow showing how each process is told what to make. Finish with the timeline along the bottom, contrasting total lead time with value-adding time. That contrast, often days against minutes, is the map's central finding.
- 3
Design the future state
Redesign the stream against lean questions: What is takt time? Can you build continuous flow, and where flow must break, can a supermarket pull system replace scheduling? Which single process, the pacemaker, should receive the schedule? How can you level mix and volume there? Mark the improvements needed as kaizen bursts on the map.
- 4
Plan and implement
Break the future state into loops, typically a pacemaker loop and supplier-side loops, and write a value stream plan: measurable objectives per loop, owners, dates and review points. Then execute, and re-map once the future state becomes the new current state. The map is a working document, redrawn as the stream improves, and the point of mapping is the future state, never the current-state drawing itself.
Reading the result
Three artefacts: a current-state map showing material flow, information flow and the lead-time timeline for one product family; a future-state map showing a leaner design with the required improvements marked as kaizen bursts; and a value stream plan that sequences and assigns the work.
- Read the timeline first. The ratio of value-adding time to total lead time is the headline; if lead time is twelve days and processing time is under an hour, the opportunity is in the queues, not the machines.
- Read the information flow against the material flow. Where multiple schedules, expediting and forecasts push material into queues, that is where pull can replace push.
- Treat the future-state map as the real output. A current-state map without a future state and a plan is wallpaper.
A worked example
A uPVC window manufacturer maps its casement window value stream
A Midlands manufacturer of uPVC windows for housebuilders and trade installers quotes ten working days but is routinely late, and finished frames queue everywhere. The operations manager maps the casement window family, roughly 60 percent of volume, from profile delivery to despatch, walking the flow with a stopwatch and a clipboard over two days.
- Choose the product family
- Casement windows share the same route: profile cutting, reinforcement, welding, corner cleaning, bead cutting, hardware and glazing, then despatch. Demand averages 90 frames a day across two shifts, giving a takt time of about 10 minutes. Vertical sliders and doors run through partly different equipment and are excluded from this map.
- Map the current state
- The walk finds six process steps with a combined processing time of 42 minutes per frame, but 11.5 days of lead time. Inventory triangles tell the story: two days of cut profile before welding, four days of welded frames awaiting corner cleaning and hardware, three days of finished frames awaiting glass deliveries. The information flow is the shock: sawing, welding and hardware each work to separate weekly schedules from the office, plus a daily shortage list, so every department builds batches to its own plan.
- Design the future state
- The future state welds, cleans and fits hardware in a continuous flow cell paced at takt, scheduled at a single pacemaker, with a small supermarket of cut profile pulled by the cell and a FIFO lane into glazing sized to the glass supplier's daily delivery. Kaizen bursts mark the three changes this depends on: welder changeover under ten minutes, relayout of cleaning and hardware benches into the cell, and a firm daily glass delivery slot. Projected lead time falls to about three days.
- Plan and implement
- The plan runs in two loops. Loop one, the pacemaker cell, gets ninety days: changeover reduction first, then the relayout, owned by the production manager with a weekly review. Loop two, supplier loops for profile and glass, follows with renegotiated delivery frequencies. Success measures: lead time from 11.5 to 3 days, work in progress down two thirds, on-time delivery above 95 percent.
The read. The map shows the lateness was never a capacity problem: 42 minutes of work was taking eleven and a half days because three uncoordinated schedules pushed batches into queues. The future state removes two of the three schedules rather than speeding up any machine. The honest caveat is that everything depends on the glass supplier's daily delivery holding, and the plan treats that as a named risk with an owner, not a hope.
Pitfalls
- Mapping from the ERP system or from memory in a meeting room. The method's discipline is direct observation; maps drawn without walking the flow record the official process, not the real one.
- Stopping at the current state. Teams frame the current-state map and never design a future state, which is the only part that changes anything.
- Mapping the whole plant instead of one product family. The result is an unreadable diagram and no takt time worth calculating.
- Treating the snapshot as permanent truth. Inventories and demand move; a map counted in a quiet week can understate the problem badly, so note when and how the data was taken.
- Applying the manufacturing notation unmodified to office or service flows, where the 'inventory' is invisible work in inboxes and the biggest delays are approval loops the standard icons do not capture well.
What the critics say
The tool was built for repetitive discrete manufacturing, and cross-sector reviews find its concepts and metrics translate unevenly to services, healthcare, construction and knowledge work, where flow units, demand and value are harder to define and the mapping conventions need substantial adaptation.
Shou, W., Wang, J., Wu, P., Wang, X. and Chong, H.-Y. (2017) 'A cross-sector review on the use of value stream mapping', International Journal of Production Research, 55(15), pp. 4433-4454.
A value stream map is a static snapshot averaged over one observation period. It cannot represent variability, product mix effects or dynamic behaviour, which is why researchers have paired VSM with discrete-event simulation to test future states before committing to them.
McDonald, T., Van Aken, E. M. and Rentes, A. F. (2002) 'Utilising simulation to enhance value stream mapping: a manufacturing case application', International Journal of Logistics: Research and Applications, 5(2), pp. 213-232.
The single-family, single-facility scope can suboptimise: improving one mapped stream may simply move constraint and inventory to shared resources or to unmapped streams, a limit Rother and Shook acknowledge by treating the extended, multi-plant stream as a later exercise.
Sources and further reading
- Rother, M. and Shook, J. (1998) Learning to See: Value Stream Mapping to Add Value and Eliminate Muda. Brookline, MA: Lean Enterprise Institute. ↗
- Womack, J. P. and Jones, D. T. (1996) Lean Thinking: Banish Waste and Create Wealth in Your Corporation. New York: Simon & Schuster.
- Hines, P. and Rich, N. (1997) 'The seven value stream mapping tools', International Journal of Operations & Production Management, 17(1), pp. 46-64.
- Shou, W., Wang, J., Wu, P., Wang, X. and Chong, H.-Y. (2017) 'A cross-sector review on the use of value stream mapping', International Journal of Production Research, 55(15), pp. 4433-4454.