Value Stream Map

What this tool does

CM8-240 is a working value stream map. Value stream mapping is the material-and-information mapping method Toyota developed to see a whole flow of work at once — every step a product passes through from order to delivery, with the time it spends being worked on and, more importantly, the time it spends waiting. You record one row per process step; the tool builds the timeline, works out the total lead time, the process-cycle efficiency and the value-add share, ranks the queues that deserve attention first, and — once you map a future state — shows what your redesign is worth in lead time.

One file can hold several maps: every step carries a value stream name, and steps that share a name belong to the same map. The tiles, charts and summary tables always report the current state of the stream with the most steps, so the headline figures describe one honest flow rather than an average of several.

Everything runs inside this single file. There is no account, no upload and no network request of any kind, so your cycle times, defect rates and staffing figures never leave the computer you are using.

Process time, wait time, lead time

Three times matter, and keeping them apart is most of the discipline:

• Process time — the minutes of actual work done on one unit at the step. The machining cycle, the keying of the order, the inspection itself.
• Wait time — the minutes one unit queues before the step starts. The order sitting in an inbox, the blank sitting on a rack, the finished part waiting for the carrier.
• Lead time — wait plus process, summed along the stream: how long one unit takes to get from the first step to the last.

On almost every unimproved stream, the waiting dominates — hours of work inside days or weeks of lead time. That is the single most useful thing a map shows, and it is why this tool asks for the wait in front of every step rather than letting you record only the work. If a step has no wait recorded, ask whether that is observation or optimism.

The three value classes

Every step is classified with one honest test: would the customer pay for this minute?

• Value-adding — the step transforms the product towards what the customer ordered. Machining the bore is value-adding. The customer would pay for it, knowingly and willingly.
• Non-value-adding (waste) — the step exists but the customer gets nothing from it. Storage, double handling, rework, moving things between buildings. The aim is elimination, not improvement — a faster way to do something unnecessary is still unnecessary.
• Necessary non-value-adding — the customer would not pay for it, but today you cannot remove it: a regulatory inspection, a credit check, programming forced by the current technology. These are reduced by changing the constraint, not by exhortation.

The tool will not let you mark a queue or a storage step as value-adding — the customer pays for transformation, not waiting — and it treats an operation entered with zero process time as a misclassified delay. Beyond that, the classes are your judgement. Be harsh: teams that classify honestly are usually shocked by how little of the work is value-adding, and the shock is the beginning of the improvement.

Process-cycle efficiency — the lean number

Process-cycle efficiency is the share of the lead time during which anything is actually happening to the product. It is the value stream number, and for most streams that have never been mapped it comes out between 1% and 5% — the product spends more than 95% of its life waiting. That figure reliably surprises people, and it is meant to: it says the fastest way to shorten the lead time is not to work faster but to wait less.

The tile shows the PCE for the current state of the largest stream, and the donut chart shows the same arithmetic visually — waiting in grey, value-adding work in green. World-class streams reach 25% or more, but for a first improvement cycle, doubling a 2% PCE is a serious achievement. Set your own target in the Settings tab.

Walking the stream

A value stream map is drawn at the place the work happens — the gemba — not in a meeting room from memory or from the standard times in the planning system. The method is simple:

1. Pick one product, or one order, and follow it — or its paperwork — through every step from the trigger (the order arriving) to the end (dispatch, or cash).
2. Time the work yourself with a watch. Standard times record what the system believes; the map needs what actually happened today.
3. For the waits, count what is in the queue in front of each step and note how fast the step consumes it, or simply ask "when did this one arrive here?" and look at the date on the ticket.
4. Write down what you see in the notes — batch sizes, who decides when the step runs, why the queue forms. The numbers say how big the problem is; the notes say why it exists.

Walk upstream against the flow if you can — it is easier to see queues when you meet them head-on — and record the date of the walk against every step, so a future re-walk is comparable.

Mapping rules

• Map one product family per stream. A map that averages a machined bracket with a fabricated frame describes neither. If two products share every step and similar times, they are one family; otherwise give them separate stream names.
• Use real observed times, not standards. The planning system's times are targets; the map records reality. When they differ, the difference is itself a finding.
• Time one unit, consistently. Process time is per unit, not per batch. If a step runs batches of fifty, the unit's wait includes the time it spends inside the batch waiting for its turn.
• Keep the same units. Everything here is minutes. Convert days of storage into minutes (a 2-day wait in a 24-hour warehouse is 2,880 minutes) rather than mixing units in your head.
• Number the steps. The step order drives the map and the tables; leave gaps (10, 20, 30) if you expect to discover steps you missed — most first walks do.

Reading the ladder chart

The first chart is the classic VSM timeline ladder stood upright: one bar per step in order, the wait in front of the step stacked under the work done at it, with the work coloured by value class. Read it in two passes. First the shape — where are the tall bars, and how much of each is wait rather than work? Then the colour — how much of the work itself is green? A stream can have impressive machines and still be mostly grey bar, which is the map politely saying the machines are not the problem.

The lead-time-by-step chart carries the same data with the wait + process breakdown written above each bar, and prints well when the ladder is crowded.

Designing the future state

The current-state map is a diagnosis; the future state is the design you commit to build, usually within six to twelve months. Enter it as a second, complete set of steps for the same stream with the state set to future — the whole stream again, not just the steps you changed, so the totals compare honestly. Three principles do most of the work:

Takt. Takt time is the available working time divided by the customer demand in that time — the rhythm the stream must produce at to match demand exactly. Steps much faster than takt are overproducing into queues; steps slower than takt are the constraint. Design each step towards takt rather than towards its own maximum speed.

Flow. Wherever possible, make the product move from one value-adding step straight into the next — smaller batches, steps moved adjacent, one-piece flow where the process allows it. Every queue you remove takes its minutes straight out of the lead time.

Pull. Where flow breaks — different shifts, shared machines, distant suppliers — let the downstream step pull what it needs through a small controlled buffer, rather than letting the upstream step push whatever it produced. A kanban-controlled buffer of hours beats an uncontrolled stock of weeks.

Once future-state steps exist, the comparison chart and the future-state tile show the lead time current against future, per stream — the headline your improvement plan will be judged by.

The kaizen priorities

The Kaizen priorities table is the improvement backlog the map produces automatically: every waste step and every queue, ranked by the minutes it costs, with its share of the total lead time. Work it from the top. On most first maps, the top two or three rows account for more than half the lead time, and none of them require capital — a scheduling rule, a batch size, a collection time. Attack the biggest queue first, re-walk the stream, and re-enter the times: improvement you have not re-measured is improvement you are taking on faith.

FAQ

How many steps should a map have? Between five and fifteen for a first map. Fewer and you have hidden the queues inside the steps; many more and you are drawing a process flowchart, not a value stream map. Detail belongs in the notes.

Can I map an office or service process? Yes — the method transfers directly. An invoice, a quotation or an insurance claim queues in inboxes exactly as a casting queues on a rack, and office streams routinely show PCE below 1%.

Uptime and defect rates — where do they matter? They are meaningful on operations and inspections, where they explain queues: a machine at 82% uptime needs a buffer in front of it, and a 3% rework rate is why the inspection step still exists. Recording them on a storage step tells you nothing.

Whose times should I record when they vary? The typical unit on a typical day. If variation is severe, note the range in the step's notes — and treat the variation itself as a finding, because queues are largely made of variation.

Why does everything report the "largest stream"? Summing lead time across different streams would add times from different products, which means nothing. The tool therefore reports the stream with the most current-state steps; use the filters to look at any other stream on its own.

Saving your work

Steps, settings and the report header are written to this browser's local storage as you type, and the toolbar shows the time of the last save. That storage belongs to one browser on one computer: another browser, a private window, a second machine or a clean-up tool that clears site data will not have it.

Treat Export .json as the real save — one file containing everything, which Import .json restores anywhere. Export CSV gives you the register for spreadsheet work. Reset asks twice, then erases everything this tool has stored. There is no undo.

Accuracy & disclaimer

The arithmetic here is deliberately simple — waits and process times summed, and one division for the PCE — and the tool does it faithfully. Everything that matters sits underneath: whether the times were observed or guessed, whether the value classes were judged honestly, and whether the walk captured a typical day. The times entered are a snapshot, and lead times move with demand, batch sizes and season.

A value stream map is a hypothesis for improvement, not an accounting document. It will not agree with your costing system and is not meant to; use it to decide where to act, verify by re-walking the stream, and treat the future state as a design until it has been built and re-measured.