Value Stream Mapping (VSM)
Value stream mapping draws, on a single page, a dated picture of everything it takes to bring a product or a service to the customer. It shows two circulations together: the flow of material and work that turns a request into a delivery, then the information flow that triggers each step. At every step it measures processing time, the time when value is actually added, against the wait time spent in a queue. The value stream is a movement, the real circulation of material and information to the customer, to be distinguished from Porter's value chain, which inventories activities without making them flow. The technique's signature is the timeline drawn beneath the row of boxes: it sums processing times on one side and wait times on the other, sets total lead time against value-added time and reveals how small a share of the delay is spent creating value.
Goal
Value stream mapping gives a cross-functional team a shared, evidenced picture of an end-to-end flow, so that it sees where time and value are lost, agrees on the waste and designs a leaner future state whose gains can be planned and measured. It answers a question that local measures leave open: why does a simple request take days to complete when the actual work fits in a few minutes. The finding comes from comparing the time spent moving the request forward with the time it spends waiting between steps.
The deliverable is a set of two maps and one list. The current-state map draws the flow as it actually runs, with its measured durations. The future-state map derives from it, once avoidable waste has been removed, and becomes the target of the improvement initiative. Between the two, the areas to address are marked on the current map as kaizen bursts, each paired with an improvement measure, and that list seeds the work plan.
Usage
When to use it
- End-to-end lead time dominates the problem: the flow crosses several roles or units and it is the total time that costs.
- Existing, repeatable process: improving a flow whose current state can be observed or estimated on real cases.
- Team to align on the waste: building a shared visual language across business and technical teams before designing the solution.
- Current and future state to compare: quantifying the gap between the process as it runs and its leaner target, to ground an initiative.
- Development flow to shorten: mapping the path from discovering a solution to putting it into production, to reduce delivery lead time.
When not to use it
- Knowledge-based or non-repeatable work (research, creative, decision): the flow does not repeat and the map lies, prefer a qualitative root-cause analysis.
- A single localised step at fault: mapping the whole flow costs too much, target the step with a root-cause analysis.
- Continuous, incremental improvement sought: VSM's one-off redesign posture sits poorly with continuous pull, prefer flow metrics and a Kanban board.
Description
Value stream mapping is one of the methods of process analysis, alongside SIPOC, which gives its one-page overview. BABOK (§10.34) places it among the Lean methods for analysing a flow, inherited from the shop floor and the supply chain, where a product is followed from supplier to supplier, setting wait time against processing time at each stage. The Agile Extension (§7.23) carries the same map to a higher altitude: in that framing the value stream is the circulation of material and information needed to bring a solution to the customer, value is read across the whole customer experience, and the waste that matters is the delay between steps and between initiatives. It is the same technique, applied now to a workshop floor, now to a software delivery chain.
The value stream and the two maps
A mapping always proceeds in two moves, and their order is the technique's central discipline. You draw the current-state map first, the flow as it works, because you can only improve what you have first seen. You then derive the future-state map, the same flow stripped of avoidable waste, which sets the target. The current-state map is the object of analysis, the future-state map is the result of design, and conflating the two, drawing the target before recording reality, is the mistake that empties the exercise of its meaning.
Drawing the current state
Preparation gathers a cross-functional team: the business experts of the flow and the technical roles it draws on, development, testing, operations, architecture, supplier, with a business analyst usually holding the facilitation. You appoint a value-stream owner who knows the current process, you choose one product, one product family or one service and fix the map's scope, then you name the value for the customer the flow must produce. A map fits on one family and one page, failing which it becomes unreadable.
You then trace the flow working backwards from the customer, observing or simulating the real path. The process boxes line up left to right. Under each, a data box carries its measures: processing time, changeover time, availability rate, number of operators. The information flow is drawn across the top of the map and shows what triggers each step, the orders and the schedules. Between the boxes, inventory or queue triangles carry the work waiting. In a service flow, these triangles carry requests stacked in an inbox or a queue, and that is where, most often, the bulk of the delay hides.
You then fill in each data box. Processing time is measured or estimated, and an estimate beats an empty cell: the finer the reading, the easier the improvement is to target. At each queue, you record the wait time. The standard notation that carries these objects, the icons for a process, a stock, a push flow, comes from Rother and Shook, and a map everyone can read depends on it as much as on the numbers it carries.
The timeline and cycle efficiency
Three durations must be kept apart, and the literature, like both bodies of knowledge, often confuses them. Processing time is the effective work time at a step, the time when the request moves toward the customer, sometimes called cycle time in the Lean tradition. Wait time is the time spent in a queue before a step, with nothing moving forward. Total lead time is the sum of the two across the whole flow, the clock the customer sees. Confusing a step's cycle time with the flow's total lead time distorts the whole analysis built on them.
The timeline is drawn along the bottom of the map, in a sawtooth: the upper segments carry the wait times, the lower segments the processing times. They are summed separately. The sum of the processing times is the value-added time. The sum of all times, processing and wait, is the total lead time. Process cycle efficiency is the ratio of the first to the second. In an administrative or service flow it often falls to a few per cent, which shows that the lever is the queue. This timeline is what separates VSM from a plain flowchart, which shows the sequence without ever saying how long the request spends doing nothing.
From current state to future state
The analysis of the current state separates the steps that add value from those that do not, then splits the latter between necessary waste, imposed by a regulatory control or a legal delay and which cannot be removed, and superfluous waste, the excess paperwork or the over-long queues that can be eliminated. The Agile Extension routes this analysis through root-cause analysis, to understand why a queue forms before claiming to remove it. You mark the areas to address with kaizen bursts on the current map, set an improvement measure for each, then draw the future-state map with the superfluous waste removed. That map becomes the initiative's target state.
What remains is to plan the improvement by listing what supports it: information systems, training, changeovers. At this stage the agile frame adds what shop-floor Lean does not provide: changes to the supporting technology become backlog items that feed an agile initiative. Once implemented, the future state becomes the new current state, and the loop begins again, a cycle of continuous improvement. The Agile Extension also recognises as a mapping object in its own right the development value stream itself, from business analysis to release by way of planning, design, build and test, each step carrying its wait time and its processing time: you then map the delay that runs from discovering a solution to delivering it, in order to reduce it.
Two traps lie in wait during execution, distinct from the contraindications because their remedy remains VSM itself. Mapping paralysis consists in perfecting the current-state map instead of moving on to improvement: you time-box the current state, a good-enough map beats a perfect one that is never acted on. Overload threatens the moment you try to capture everything: the map becomes an unreadable inventory, and the remedy is the discipline of one family and one page. In an agile setting, this economy goes as far as drawing the map on a whiteboard, freehand, which keeps it alive and correctable on your feet.
AI considerations
Two uses are solid and a third is to be avoided. The first is feeding it real data: where each step leaves a timestamp in a management system, a model mines the event logs to establish the wait and processing times observed, queue by queue. This is the best fit, because the real wait is what participants underestimate from memory, and an automatic reading restores it without indulgence. The second is calculation and simulation: summing the durations, keeping cycle efficiency up to date at every change and replaying several future states, removing one queue, merging some steps, to quantify the gain before deciding.
What must stay with human judgement is the reading of waste. Deciding that a wait is necessary, because a regulatory control or a legal delay imposes it, or on the contrary superfluous, is not in the data: it is a business reading. The map must reflect the real flow, the one observed in the field. The system itself records the ideal path it believes it follows, so that a model fed on logs alone will reproduce the tool's blind spots. And the team's agreement on the waste to eliminate is a collective act that no model signs in its place.
Examples
The technique reads best on its own artifact. The current state chosen is a service flow common in Switzerland: a health insurer (caisse maladie) reimbursing an outpatient treatment invoice received by post. The flow has five processing steps, each preceded by a queue.
| Step | Processing time (value added) | Upstream wait |
|---|---|---|
| Scanning the mail | 4 min | 3 h |
| Completeness check | 6 min | 5 h |
| Benefits check | 8 min | 6 h |
| Validation | 3 min | 3 h |
| Payment and statement | 2 min | 3 h |
| Total | 23 min | 20 h |
Value-added work totals twenty-three minutes. Cumulative waiting reaches twenty hours. The total lead time seen by the insured person is therefore 20 h 23 min, about two and a half working days, and cycle efficiency comes out at 23 min over 1'223 min, that is 1.9%. The invoice spends nearly 98% of its life waiting. What the map shows, and what a table of average delays hides, is that the five levers are the queues: shaving a minute off a check changes nothing, removing a queue changes everything.
A future state sketched on the same map confirms it. Receiving invoices electronically removes the scanning step and its three-hour wait. Merging the completeness check and the benefits check into a single verification erases a five-hour intermediate queue. Validating directly the invoices whose amount stays below a threshold removes the three-hour validation wait. All three moves act on the queues: they take away eleven of the twenty hours of waiting, the lead time falls from two and a half days to about one working day, and cycle efficiency rises accordingly.
Cost
| Phase | Level | Justification |
|---|---|---|
| Preparation | Medium | Gathering the cross-functional team and the value-stream owner, fixing the scope and recording the real flow in the field rather than from memory. The work is bounded but assumes access to the field and the mobilisation of several functions at once. |
| Execution | Low to medium | A current-state map on one family and one page is drawn in a single workshop, often on a whiteboard. The cost rises if the families multiply or if measured times are wanted rather than estimated ones. |
| Documentation | Low | The deliverable fits on one or two pages, the current map and the future map. The value is in the immediate use: the maps serve until the future state is reached and the loop restarted. |
Tools
Value stream mapping is drawn by hand first. In an agile setting, a whiteboard and sticky notes are enough for the workshop: the map stays alive, the team corrects it on its feet. A spreadsheet then holds the data boxes and computes the sums and cycle efficiency with no other tool. For a wide flow or a map to maintain over time, diagramming software carries the standard notation, with the VSM symbol libraries of Lucidchart, draw.io or Visio, or a dedicated Lean tool such as eVSM that recomputes the timeline at every change, while online whiteboards such as Miro or Mural make the remote workshop possible. Where the information system timestamps each step, a process-mining tool reconstructs the real wait times from the event logs, replacing estimation with measurement.
Sources
- IIBA, A Guide to the Business Analysis Body of Knowledge (BABOK Guide) v3, §10.34 Process Analysis: the definition of the technique, its place among the process-analysis methods and the stated limits of its use.
- IIBA, Agile Extension to the BABOK Guide v2, §7.23 Value Stream Mapping: the value stream as the circulation of material and information bringing a solution to the customer, the current and future states, the development value stream as a mapping object and the passage of removed waste into backlog items.
- Mike Rother and John Shook, Learning to See: Value Stream Mapping to Add Value and Eliminate Muda, Lean Enterprise Institute, 1999: the codification of the technique out of the Toyota Production System, the standard notation and the current-/future-state method.
- James P. Womack and Daniel T. Jones, Lean Thinking: Banish Waste and Create Wealth in Your Corporation, Simon & Schuster, 1996 (rev. 2003): the value stream as the second principle of Lean, the conceptual foundation the mapping puts into practice.
- Lean Enterprise Institute, Value-Stream Mapping, Lean Lexicon: the reference public definition of the technique and its standard icons.