Mapping the whole flow
Value Stream Mapping
Value stream mapping is a pencil drawing of the whole route one product family takes. Information flow runs across the top, material flow across the middle, and a timeline runs along the bottom. The timeline is the point. On a bracket line making 920 pieces a day it totals 16 working days against 104 seconds of work.
Matthew SavasFounder of Kaizumi. Reviewed 17 August 2026.
Purpose
What a value stream map is for
A value stream map shows you two numbers side by side: the total time a job takes from start to finish, and how much of that time someone actually spends working on it.
For example, look at a steel bracket. Raw metal arrives at the loading dock, and finished parts leave on a truck 16 working days later. During those 16 days, people or machines spend only 104 seconds doing active work on the part.
In this example, 16 working days equals 883,200 seconds of plant operating time. That means the bracket gets only one second of real work for about every 8,500 seconds it spends in the building.
The rest of the drawing shows the steps, the piles of inventory waiting in between, and the production schedules sent to each area. These details help you understand why the work takes as long as it does. If you draw the steps but skip the time calculation at the bottom, you just have a standard flowchart.
This time comparison shows why asking people to work faster rarely solves the real delay. The 104 seconds of active work is already very small. Most of the 16 days are spent sitting in queues between steps. That waiting time is where you can make real improvements.
These gaps are common across all kinds of workplaces. An insurance claim might take only 2 hours of active review across 11 working days. A patient in a hospital often spends most of their stay waiting for tests, beds, or the next doctor's visit.
Three bands
What is on a value stream map
A value stream map is divided into three horizontal bands on a single page. Which band a mark sits in is part of what it means. Each band shows a specific part of the system:
- The top band shows information flow, moving from right to left. Customer orders enter on the right side. From there, schedules go out to suppliers on the left and down to each internal work step. Drawing this flow from right to left reflects how customer demand pulls work through the system. It is also the order you follow when you walk the process in person.
- The middle band shows the flow of work, moving from left to right. Boxes represent each work step. Triangles represent work waiting in queues between those steps. This band looks like a standard process map.
- The bottom band shows time. It looks like a stepped ladder. The high rungs show the time items spend waiting between steps. The low rungs show the active working time for each step. The totals at the far right of this line give you the overall lead time and active work time.
A standard process map only shows the middle band. Because it leaves out information flow and waiting time, it often cannot explain why a request takes weeks to finish even when each step takes only a few minutes.
One map covers one product family or service line from start to finish on a single page. If you cannot fit it on one sheet, you are probably mapping an entire department or facility instead of a single flow of work.
The value stream is the complete sequence of activities needed to deliver a product or service. The map is simply a drawing of that path.
Simulation
See a value stream map in action
A value stream map captures a moving operation on a static page. Watching the flow move in real time helps you see how the symbols connect to daily work.
The simulation below shows a map of an electric bicycle assembly line. It includes four main work steps, along with the cycle time and changeover time for each one.
Between the steps, you can see inventory buffers shown as supermarkets and inventory triangles. Each buffer displays how many minutes of supply are waiting there. These are the same types of numbers you would record during an in-person walk of your own process.
When you add up all the waiting time in the buffers, a bicycle frame takes about half an hour to move through the entire plant. Only two to three minutes of that total time is spent on actual work. This follows the same pattern as the bracket example, but on a scale you can watch in a few minutes.
The whole plant as four process boxes and the stores between them.
Open the simulation on this term →
Symbols
What each value stream map symbol means
Every symbol on a value stream map represents a specific operational question. You record the answer as a number directly under the symbol. This turns the map into a clear summary of operational data.
| The mark | What it draws | The question it asks | The number under it |
|---|---|---|---|
| Process box | One step where material flows without stopping | Where does the part stop being one thing and start being another? | Cycle time, in seconds |
| Data box | The facts sitting under a process box | Can this step hold the beat? | Changeover minutes, uptime, operators, batch size |
| Inventory triangle | A pile of parts standing still | How long would this pile last? | Pieces, turned into days of supply |
| Supermarket | A store the next step pulls from | What is the most this store may hold? | Maximum pieces |
| FIFO lane | A queue in strict order, with a hard ceiling | How many parts before the step upstream must stop? | Maximum pieces in the lane |
| Push arrow | Material moved because a schedule said so | Who asked for this? | Nothing. That is the point |
| Withdrawal arrow | Material moved because somebody used it | What was taken? | One container |
| Kanban post | Where signals collect between rounds | How often does the runner come? | Round frequency |
| Outside source | A supplier plant or a customer plant | What crosses the fence, and how much? | Volume per period |
| Truck | A shipment | How often does material actually move? | Daily, or weekly |
| Information arrow | A schedule, an order, a phone call | How often is this said? | Weekly, daily, every 20 minutes |
| Kaizen burst | A change you intend to make | What will it be worth? | The target number |
People often draw the inventory triangle without recording a number underneath it. A triangle by itself only shows that work is waiting, which everyone already knows. Writing 4,600 pieces gives you a concrete count, and converting that count into 5 days gives you data you can act on.
Mapping a department instead of a product family. A map follows one family through the whole route. Map the stamping department and you get a picture of a building, not a stream. Sort your families first with a product family matrix.
Taking the counts out of the ERP. Count with your own eyes, on the day you walk. The system's number is what should be there, which is a different question.
Splitting the walk between two people. The days live in the handoffs. Split the stream and nobody writes them down. One person walks it end to end, at the gemba, with a pencil.
The timeline
How to calculate days of supply
To calculate how long work waits between steps, you convert the number of items in each queue into days of supply. To do this, divide the items you counted by the average daily customer demand.
First, determine the daily customer demand. In this example, the team ships 18,400 brackets over 20 working days in a month. That means customer demand is 920 brackets per day.
The operation runs two shifts of 460 minutes each, which totals 920 minutes or 55,200 seconds of working time per day. Dividing the available working seconds by daily demand gives the takt time: 55,200 ÷ 920 = 60 seconds. You use this daily demand rate to evaluate every queue on the map.
Days of supply = P ÷ D- P
- Pieces on hand. What you counted on the floor, on the day you walked it.
- D
- Demand per day. What the customer takes in a day, not what the schedule says. Here it is 920.
Repeat this calculation for each queue to get the waiting times for the upper rungs of the timeline ladder. The lower rungs do not require conversion; you simply copy the active cycle times directly from the data boxes under each step.
Flow efficiency = VA ÷ LT- VA
- Value-creating time. The sum of the low rungs, in seconds.
- LT
- Lead time. The sum of the high rungs, in the same unit. One working day here is 55,200 seconds.
The current-state ladder, rung by rung
Lead time 16 days. Value-creating time 104 seconds. Flow efficiency = 104 ÷ (16 × 55,200) = 0.012%.
Keep the totals for waiting time and active work time separate. Do not add them together into a single sum. Comparing them side by side is what gives you insight into the process.
Adding up the items in the four queues gives 14,720 brackets waiting across the line. This inventory represents the same 16 days of waiting time, counted in physical parts instead of days.
Accounting reports track this through inventory turns once a quarter or once a year. On a value stream map, you see the actual quantities sitting at each specific step on the day you walk the process.
Counting what should be there. The map is a photograph of one day. Use what you counted, not what the system carries.
Dividing seconds by days. 104 seconds over 16 days is not 6.5. Sixteen working days is 16 × 55,200 = 883,200 seconds of running time. 104 of them are work. That is one second of work in every 8,492.
You can apply this same method to your own process. It is best to draw your first map by hand with pencil and paper while observing the work. You can use this blank map template to get started with the standard layout.
Our digital map builder lets you enter your steps, process times, and waiting times. It automatically draws the timeline ladder and totals your current state alongside your planned future state. You can also visit the value stream map builder page directly.
Future state
The future state value stream map
A current-state map shows how the process works today. Drawing it identifies where delays occur, but drawing it does not fix them on its own.
The second drawing is the future-state map. It shows how you plan for the work to flow after you make improvements. You place a jagged callout, called a kaizen burst, over each specific area you plan to change.
It is best to design the future state on the same day you complete the current-state map, while your observations are fresh, and assign target dates to each change.
| Rung | Current | Future | What changed |
|---|---|---|---|
| Coils | 6 days | 2 days | A daily milk run replaces the weekly coil delivery |
| Stamped | 5 days | 1.5 days | A SMED workshop cuts stamping's changeover from 45 to 9 minutes, so it can run small batches |
| Welded | 3 days | 0.1 days | A capped FIFO lane replaces the pile |
| Finished | 2 days | 1 day | Shipping pulls from a finished-goods supermarket |
| Lead time | 16 days | 4.6 days | 11.4 days out — a 71 percent cut |
| Value-creating | 104 s | 100 s | Assembly rebalanced from 62 to 58 seconds |
| Flow efficiency | 0.012% | 0.039% | The ratio the whole drawing exists to move |
Look at the results in the last three rows. Total lead time dropped by 11.4 days, while active work time changed by only 4 seconds. The improvement did not come from making people work faster. It came from reducing the inventory on the floor from 14,720 brackets to 4,232, which eliminated days of waiting.
Only one process step required an adjustment to its cycle time. Assembly took 62 seconds, which was slower than the 60-second customer takt time. Rebalancing the work between the two operators brought that cycle time down to 58 seconds so it could keep pace with demand.
All the other improvements focused on when and how work moves between steps, rather than how fast individuals perform their work.
The scheduling method changed as well, making the other improvements possible. In the current state, a central schedule pushes separate weekly targets to all four steps. In the future state, you send a schedule to only one step: assembly, which acts as the pacemaker.
Each upstream step then produces parts only when the pacemaker step consumes them. This replaces four separate push schedules with a coordinated pull system. Leveling the mix of work at the pacemaker step through heijunka keeps buffer sizes small and predictable.
Every kaizen burst on the map represents a project with an assigned owner and completion date, often carried out during a kaizen event. Our step-by-step mapping guide walks through building both maps in detail, and our batch versus flow explainer explains why large batches create so much waiting time.
Offices
Value stream mapping in an office
In an office, clinic, or service team, you usually cannot count physical piles of parts. Instead, the waiting work sits as digital queues in inboxes, ticketing systems, or spreadsheets.
Because of this, mapping an office process involves two main differences:
First, you record waiting time directly from timestamps rather than converting physical counts into days. You can look at when an item finished one step and when someone started working on it at the next step. The gap between those two dates or times is your waiting time.
Second, you add a field called %C&A (Percent Complete and Accurate) to every data box. This measures the percentage of work that arrives with all necessary information, so the next person can process it without sending it back for corrections.
Ask the people downstream of each step one question. What share of the work arrives complete enough to use without sending it back? A 70 percent intake means three files in ten bounce. Multiply the percentages along the stream and you have the stream's first-pass yield.
In an insurance claims process, for example, a file might require about 2 hours of active work across 11 working days. Assuming an 8-hour workday, 11 days equals 88 working hours, which gives a flow efficiency of 2 ÷ 88 = 2.3 percent.
If you multiply the accuracy rates across five steps, the rolled first-pass yield is 0.70 × 0.90 × 0.85 × 0.95 × 0.99 = 50.4 percent. That means only about half the claims move through the entire process cleanly the first time. The rest must loop back to earlier steps to correct missing or inaccurate information.
Comparing the two examples, the claims office has a flow efficiency of 2.3 percent, while the bracket line has a flow efficiency of 0.012 percent.
That makes the office about 190 times better on this measure, but not because anybody manages it better. A queue of digital files takes up no floor space and costs nothing to store. It never shows up in a cost report, so nothing pushes it down.
This is why tracking rework is essential for office maps. In a manufacturing plant, excess work sits out on the floor as physical inventory, making it one of the easiest of the seven wastes to see. In an office, inefficiency usually hides in rework loops, where staff handle the same file multiple times because it arrived with errors.
To see full examples of non-manufacturing processes, read our insurance claims mapping guide and healthcare value stream guide. Our guide to patient journey mapping uses the same principles to map the path of a person receiving care rather than a physical part.
Direct answers
Value stream mapping questions people ask
- What is a value stream map?
- A value stream map is one page showing every step one product family passes through. Information flow runs across the top, material flow across the middle, and a timeline along the bottom. The timeline totals two numbers: lead time and value-creating time. On the bracket line here that is 16 working days against 104 seconds.
- How is a value stream map different from a process map?
- A process map is the middle band on its own. A value stream map adds the information flow that sets the work off, and a timeline that totals waiting against work. That is why a process map can show every step and still not explain why the job takes six weeks.
- How do you calculate days of supply from an inventory count?
- Divide the pieces you counted by average daily customer demand. A pile of 4,600 stamped brackets on a line that ships 920 a day is 5 days of supply. Count what is on the floor on the day you walk, not what the system says should be there.
- What is flow efficiency, and what is a normal number?
- Flow efficiency is value-creating time divided by lead time, in the same unit. The bracket line runs 104 seconds inside 16 working days, which is 0.012 percent. A claims process runs about 2 hours inside 11 working days, which is 2.3 percent. Anything above a few percent in a stream nobody has worked on is unusual.
- What is the difference between a supermarket and a FIFO lane?
- A supermarket holds a fixed amount of several part numbers, and the next step takes what it wants, when it wants it. A FIFO lane holds one sequence with a hard ceiling, and when the lane fills up the step feeding it must stop. Use a supermarket where the next step picks. Use a FIFO lane where the order must hold.
- What is the pacemaker process?
- The pacemaker is the one process you schedule. Everything after it flows. Everything before it replaces what the pacemaker took. On the bracket line the pacemaker is assembly, which is why the future state has one schedule arrow instead of four.
- Can you map an office or a hospital process?
- Yes, and the drawing is the same. Two things change. You record wait time between steps instead of turning a pile of stock into days. And you add percent complete and accurate to each step, because the waste in an office hides in rework loops instead of in piles.
- Do you need software to draw one?
- No. Draw the first one in pencil, on one sheet, standing where the work happens. Software is for the version you circulate afterwards. Draw straight into software and you will end up with a tidy map of a stream nobody walked.
Keep going