VSM: mapping the value stream from end to end

In one line

VSM (Value Stream Mapping) is a hand-drawn picture of the whole process from end to end —material and information, from the supplier to the customer— that shows on a single sheet of paper where inventory piles up, how long each step takes and which one is the real bottleneck.

What it is

VSM is a tool that comes from Lean/Toyota. According to the cited source, it is "a conceptual tool that outlines the value chain, emphasizing the flow of materials and information, which start at the supplier and head toward the customer". The underlying idea is simple: instead of looking at a single isolated process (one machine, one workstation), VSM makes you look at the complete chain of a product family, from end to end, to find where the real waste is.

A VSM is always built in two versions, not just one:

• Current-state map (current VSM): a portrait of how the process works today, with data measured on the shop floor —not assumptions or catalog data—. It is the baseline against which any improvement will be compared.
• Future-state map (future VSM): the same process redesigned, applying Lean tools (continuous flow, supermarkets, pull systems, leveling) to eliminate the waste that the current map exposed.

The source describes the method in four steps, grouped into three phases:

1. Choosing a critical production area (or product family).
2. Building the current-state map: you walk the process in person, interviewing the people in charge of each area and recording the data by hand —it is intentional that the first map is drawn by hand and not in a digital tool, because it forces you to walk the plant instead of working from memory.
3. Analyzing the current-state map: you calculate the takt time and compare it against the cycle time of each process to locate the bottleneck.
4. Future-state map: you decide where to put continuous flow, where to put a supermarket with a pull system (kanban), which process will set the pace for the whole line (the pacemaker process) and how to level production at that point.

The basic symbols

A VSM is always read on two overlapping "lanes" —information flow on top, material flow on the bottom— with a handful of symbols that repeat in any map:

• Process: one rectangle for each operation or workstation, with its shop-floor data inside. In the real case below, each rectangle carries the TC (cycle time), the TCP (changeover time between models or sizes) and the TF (reliability or availability of the process).
• Inventory: a triangle with the letter I, placed between two processes, with the quantity or days of stock accumulated there. It is the symbol that jumps out fastest in a VSM: every large triangle is a waste alarm.
• Information flow: thin lines at the top of the map, connecting the customer, the production control area (or MRP) and the suppliers —orders, forecasts, purchase orders—. It is the flow that is almost never drawn in a classic process diagram, and it is one of the reasons VSM became so widely used.
• Material flow: arrows at the bottom of the map that physically connect one process to the next, plus a truck icon at each end (supplier and customer) to mark the frequency of shipments —for example, "every 3 weeks to the shipping port".

What it is for

A process improved in isolation can look better in its own indicator and still not move the final result one bit if the real bottleneck is somewhere else in the chain. VSM serves exactly that purpose: to show the complete chain on a single sheet of paper, so the improvement priority is decided with data —where the most inventory is sitting, which process is furthest from takt time— and not by intuition or because "that process is the one that bothers us the most".

How it is applied

The connection between VSM and takt time/cycle time is the heart of the analysis, and it always follows this sequence:

1. Calculate the available production time (shift, breaks, actual paid time).
2. Calculate the takt time: available time divided by the customer demand in that same period. It is the pace at which one unit should come out to meet the customer's needs, without overproducing or falling short.
3. Measure the cycle time (TC) of each process and write it inside its rectangle on the map.
4. Compare each TC against the takt time line, normally in a bar histogram: the process or processes that end up above that line are the bottleneck —they cannot keep the pace the customer asks for—; those that end up well below it have idle time and are candidates to be merged with another step.

This site has an article dedicated to that comparison in detail: Takt time vs. cycle time.

Once the bottleneck has been identified on the current map, the future map answers a fixed set of questions: In which processes can continuous flow be introduced (joining two processes with no inventory in between)? Where is a supermarket with pull flow needed (kanban) because continuous flow is not possible? Which process will be chosen as the pacemaker (the single point where scheduling arrives, with everything upstream working in continuous flow toward it)? How is the product mix leveled at that pacemaker process?

Real example

The case documented in the source is a textile and garment manufacturing company in Lima, Peru, that makes printed polo shirts and dresses for export (customers such as Polo Ralph Lauren, Lilly Pulitzer and Tommy Bahama). It worked 25 days a month, 2 shifts of 8.5 hours with 2 breaks of 20 minutes —available time per shift: 28,200 seconds— and produced 12,000 garments/month (480 garments/day).

Takt time calculation: 28,200 s ÷ 240 garments per shift = 117.5 seconds/garment. That was the pace at which each garment had to come out to meet demand.

Cycle time measured at each process in the chain, compared against that takt time of 117.5 s:

Process	Cycle time	How does it compare with takt time (117.5 s)?
Spreading	59.46 s	Below — idle time
Cutting	49.20 s	Below — idle time
Sewing	616.86 s	Far above — main bottleneck
Embroidery	73.08 s	Below — idle time
Printing	68.64 s	Below — idle time
Finishing	162.84 s	Above — second bottleneck
Shipping	30 s	Below — idle time

The current map also showed the inventory accumulated between processes: 3 days of fabric supplier delay plus 1 day of rest before Spreading; 1 day of stock after Cutting; 1 full week after Sewing; 3,500 basic garments and 500 fashion garments after Embroidery; 4,500 basic and 500 fashion after Printing; and, after Finishing, 3,500 basic (on the way to the metal detector) plus 2,000 fashion (on the way to Shipping). With that, the current map measured a lead time of 20.87 days and a value-added time (the sum of the cycle times that actually transform the garment) of 1,060.08 seconds.

When building the future map, the team noticed that Spreading and Cutting had similar cycle times —as did Embroidery and Printing— and merged them into continuous-flow cells: Spreading+Cutting went from 108.66 s down to 76.06 s combined, and Embroidery+Printing went from 141.72 s down to 99.2 s. Sewing became the pacemaker process (the one closest to the point where it makes sense to schedule production), with a pitch of 117.5 s × 10 garments per bundle = 1,175 s (19 min 35 s). The final result of the future map: lead time of 16.7 days, versus the 20.87 days of the current map.

Template to use

Drawing a real VSM takes a pencil, paper and walking the plant —you do not need software for the first map—. This template summarizes the order of work:

1. Choose the scope

• Product family or critical area to map (not the entire factory at once)
• Define where the map starts (supplier) and where it ends (customer)

2. Gather the information flow (top of the map)

• How customer orders arrive (frequency, forecasts)
• How they are translated into orders to suppliers (MRP, production control)

3. Walk the material flow, process by process (bottom of the map)

• For each process: cycle time (TC), changeover time (TCP), reliability/availability (TF), number of operators
• Between every two processes: accumulated inventory (triangle I), in units or in days

4. Calculate the available time and the takt time

• Available time = paid shift hours minus breaks, converted to seconds
• Takt time = available time ÷ customer demand in that same period

5. Analyze the current map

• Plot the TC of each process against the takt time line (histogram)
• Mark the process or processes above that line: they are the bottleneck
• Mark the largest inventory triangles: they are the most visible waste

6. Design the future map

• Where can processes be joined in continuous flow (no inventory in between)?
• Where is a supermarket with a pull system (kanban) needed?
• Which process will be the pacemaker, the single point that receives the schedule?
• How is the product mix leveled at that pacemaker process?

7. Quantify the expected result

• Current lead time vs. future lead time
• Value-added time (sum of the TC that actually transform the product)

So you do not have to build it by hand, there is a lead time and value stream spreadsheet in Excel with the formulas already done: with the map data (TC, changeovers, availability, inventories) it calculates the takt, the processes above takt and the lead time. The values it comes with are examples and you have to delete them.

Benefits

• It shows the complete chain on a single sheet of paper, not an isolated process — it prevents you from improving a step that is not the real bottleneck.
• It connects directly with takt time: with a single histogram, it shows exactly which process has the pace problem.
• It provides a common, visual language across areas (production, quality, logistics) that would otherwise argue about the same problem with different data.
• The future map is not a loose idea: it comes from answering concrete questions (continuous flow, supermarket, pacemaker, leveling) about the current map, which has already been measured.

Limitations to keep in mind

• It is a snapshot of a specific moment in the process — a demand peak, a change in product mix or a different supplier can make the map data obsolete in a short time.
• The result depends on the field data (TC, TCP, TF, inventories) having been measured on the real shop floor and not estimated from memory — a poorly measured TC drags the whole takt time and bottleneck calculation off course.
• It does not replace a detailed time study inside each process (that is solved by another tool, time measurement) — VSM works with the already consolidated cycle time of each process, not with the detail of each movement inside it.
• Drawing a complete and correct VSM of a real plant takes more time than it seems at first glance, especially the first time a team does it.

In summary

VSM is not just another process diagram: it is the end-to-end view —material and information, from the supplier to the customer— that shows where inventory piles up and, by comparing the cycle time of each process against takt time, in which one the bottleneck really is. The current map measures reality; the future map decides, with that same data, where to put continuous flow, a supermarket or a pacemaker process to bring the process closer to the pace the customer asks for.