SMED, or Single-Minute Exchange of Die, is a lean manufacturing method used to reduce equipment setup and changeover time. It does this by moving as much work as possible outside the period when a machine is stopped, then simplifying, standardising and continuously improving the tasks that remain.
This article looks at how the lean manufacturing method of SMED can be used to help manufacturers reduce inefficiencies in their manufacturing processes and support their continuous improvement strategy.
Despite its name, SMED does not mean every changeover must take one minute. “Single-minute” refers to the aim of completing a changeover in a single-digit number of minutes, less than ten. For some complex processes, that target may not be immediately realistic. The practical objective is to reduce changeover time safely and repeatably from the current baseline.
This guide explains the difference between internal and external setup, takes you through a seven-step SMED exercise and shows, using a worked manufacturing example, how shorter changeovers can release productive capacity.
Like many strategies that focus on process improvement, SMED began in Japan in the 1950s when a Toyota consultant noticed an inefficient process involving the body moulding process. Much of the waste came from the time it took to switch to new tools used on the equipment that was needed for the different phases of the body moulding process.
The changeover took between two and eight hours with Toyota also having to spend money storing vehicles being built. It was projected that Toyota could save money by speeding up the changeover process.
The improvements needed all focused on less changeover time. This meant modifying the factory equipment, and the vehicle parts and changing the order of the steps in building the car body mouldings. By the 1970s, the changeover was just three minutes.
As mentioned, SMED stands for Single-Minute Exchange of Die. The method is closely associated with industrial engineer Shigeo Shingo and the setup-reduction work developed in Japan during the 1950s and 1960s.
Although the name refers to changing a die, SMED can be applied to almost any repeatable setup or changeover, including:
Changing a fixture on a CNC machine
Switching tools or programmes between jobs
Cleaning and resetting a food or beverage production line
Changing packaging formats
Reconfiguring an assembly cell
Preparing a press, moulding machine or fabrication process for the next product
The changeover should be measured from the last good part from the previous run to the first good part from the next run. This definition matters: restarting the machine is not enough if the first pieces require adjustment, rework or scrapping before the process becomes stable.
The Lean Enterprise Institute defines SMED as a process for changing production equipment from one part number to another as quickly as possible, with a target of a single-digit changeover time. Its central principles are to distinguish between internal and external setup work and convert internal work to external work wherever practical.
During an internal changeover, the equipment is unavailable for production. Long or unpredictable setups can therefore reduce available capacity and make production schedules harder to achieve.
Manufacturers sometimes compensate by producing larger batches so that changeovers happen less often. While this can reduce the setup time attributed to each unit, it may also increase work in progress, inventory, lead times and working capital, while reducing the ability to respond to changing customer demand.
Reducing changeover time can help a manufacturer:
Recover productive machine capacity
Make smaller batch sizes more practical
Respond more flexibly to changes in demand
Reduce work in progress and excess inventory
Improve schedule adherence
Make setup performance more consistent between operators and shifts
Reduce errors through clearer, standardised procedures
Create more opportunities for continuous improvement
SMED should not be treated as a race. A faster changeover is only an improvement when safety, process control and product quality are maintained or improved.
A SMED exercise focuses on making the Changeover period as efficient as possible, ultimately leading to a changeover that takes less time.
The distinction between internal and external setup is the foundation of SMED.
Internal setup activities can only be completed while the machine or process is stopped. External setup activities can be completed before the machine stops or after it has restarted and is producing good parts.
Setup type |
When it happens |
Manufacturing examples |
|
Internal |
The machine must be stopped |
Removing the current fixture; fitting a new die; making a guarded mechanical connection; carrying out a required test cycle |
|
External |
The machine can continue producing |
Finding tools; transporting the next fixture; checking materials; retrieving the programme; preheating a die; preparing inspection equipment |
Consider an operator who waits until a machine has stopped before looking for the next fixture, tools and work instructions. None of those activities necessarily requires the machine to be idle. Preparing them in advance converts lost production time into external setup time without asking the operator to work faster.
The aim is not to move every task outside the changeover. Some activities must remain internal for technical or safety reasons. The aim is to question each activity, convert what can safely be converted, and make the remaining internal work as simple and repeatable as possible.
SMED works best as a team exercise involving the people who perform and support the changeover. Operators, production supervisors, maintenance, quality and engineering may each see different constraints and improvement opportunities.
The diagram shows the seven steps of SMED.
Start with one repeatable changeover where improvement would have a meaningful operational effect. Good candidates usually have one or more of these characteristics:
A long average changeover time
Large variation between changeovers, operators or shifts
Frequent changeovers
A machine that constrains overall production flow
A changeover associated with startup scrap or quality problems
A process the team can observe and repeat often enough to test improvements
Avoid starting with the most complex process in the factory simply because it has the longest setup. A well-chosen pilot should offer meaningful scope for improvement while allowing the team to learn and test changes safely.
Action: Select one machine, product transition and changeover type.
Expected output: A clearly defined pilot with a named owner and team.
Metric to record: Changeover frequency, median duration and variation.
Example: Select the change from Product A to Product B on a CNC milling centre because it occurs ten times a week and regularly takes between 45 and 70 minutes.
Measure what actually happens rather than relying on the written procedure or people’s recollections. Observe several changeovers if possible, using timing data and video where appropriate and agreed with the team.
Record the sequence of tasks, who performs them, how long each takes, and any time spent walking, searching, waiting, adjusting or repeating work. Include the period from the last good part of the previous run to the first good part of the next run.
Involve the operator from the beginning and explain that the purpose is to improve the process, not assess individual performance. Operators often know where the delays and workarounds are; the exercise will be weaker without their input.
Action: Break the current changeover into individual timed activities.
Expected output: An agreed current-state sequence and reliable baseline.
Metric to record: Total changeover time and the duration of every task.
Example: The team records 12 minutes removing the previous setup, eight minutes fetching the next fixture and tools, and 15 minutes producing and checking the first acceptable part.
Classify every activity as either internal or external. Ask a simple question: does the machine genuinely need to be stopped for this task?
Typical external activities include:
Retrieving tools, fixtures and gauges
Confirming that the correct material is available
Checking drawings and work instructions
Preparing cleaning materials
Locating the next programme and tool list
Inspecting the next fixture before it reaches the machine
Be honest about the current state. A task is not external merely because it could happen while the machine is running. Record how it is carried out today, then identify what needs to change.
Action: Label each observed task as internal or external.
Expected output: A time-based view of how much of the changeover currently requires stopped equipment.
Metric to record: Internal minutes as a proportion of total changeover time.
Example: Fetching the next fixture currently happens after the machine stops, so it is recorded as internal in the current state, even though the team identifies it as a candidate for conversion
Review every internal task and ask how it could be completed, wholly or partly, while production is still running.
Common changes include:
Creating a pre-changeover checklist
Delivering the next material, fixture and tools to a defined location in advance
Presetting tools and offsets away from the machine
Preheating or preconditioning components before the changeover
Preparing programmes and work instructions in advance
Carrying out quality and material checks before the current run finishes
Using duplicate or intermediate fixtures where the operational benefit justifies them
Do not transfer delays elsewhere and call them an improvement. External tasks still consume time and resources, so they should remain visible and be scheduled properly.
Action: Design a future-state sequence that moves suitable work outside machine stoppage.
Expected output: A shorter list of genuinely internal activities, supported by a defined preparation process.
Metric to record: Minutes converted from internal to external setup.
Example: The next fixture, tool cart, inspection gauge and programme are checked and positioned before the previous batch ends, removing eight minutes from machine downtime.
Once avoidable internal work has been moved outside the stopped period, simplify the tasks that must remain.
Look for opportunities to eliminate adjustment, reduce movement, make positioning repeatable and replace slow fastening methods. Depending on the process, improvements might include:
Quick-release clamps or fasteners
Standardised bolt sizes and tool requirements
Clearly marked locating points
Preset stops, guides, pins or spacers
Better point-of-use storage
Parallel working where it is safe and responsibilities are clear
Mistake-proofing to prevent incorrect positioning or assembly
A standard sequence that removes unnecessary walking and waiting
Engineering controls, guarding, isolation procedures and quality checks must not be bypassed to reduce the recorded time. Review proposed changes through the appropriate risk-assessment and change-control processes.
External work should also be streamlined. If preparation is unreliable, the internal changeover will still be delayed by missing tools, materials or information.
Action: Eliminate, combine, simplify or resequence the remaining tasks.
Expected output: A safer, simpler future-state changeover method.
Metric to record: Time removed from each internal task, together with safety and quality measures.
Example: Locating pins make fixture alignment repeatable, while a standard tool cart prevents walking back to stores during the changeover.
Document the agreed method so that the improvement can be repeated across operators and shifts.
The standard should be practical and visual. It may include:
The correct task sequence
Clear internal and external responsibilities
A pre-changeover checklist
Photographs or diagrams of the prepared setup
Tools, materials and information required
Safety and quality checkpoints
Expected task and total changeover times
A defined process for recording problems and improvement ideas
Train everyone who carries out or supports the changeover. Then observe the new standard in practice to confirm that it works under normal production conditions.
Action: Create, test and train the team on standardised work.
Expected output: A repeatable method that is understood and owned by the people using it.
Metric to record: Median changeover time and variation by shift or changeover type.
Example: A visual standard shows where the prepared fixture and tool cart must be placed, who confirms the programme, and how the first-off inspection is completed.
The first SMED exercise is rarely the end of the work. Run the new method, collect enough evidence to distinguish a real improvement from normal variation, and review any remaining delays.
Compare the results with the original baseline. Ask:
Did the change reduce the median changeover time?
Is performance more consistent?
Was safety maintained?
Did first-off quality improve or deteriorate?
Did any task or delay move elsewhere?
Can the new standard be followed on every shift?
What is now the largest remaining source of lost time?
Prioritise the next change, test it and update the standard when the improvement is proven. There is no credible fixed percentage by which every SMED cycle should improve; results depend on the process, baseline, constraints and changes made.
Action: Review the evidence, address the next constraint and update the standard.
Expected output: Sustained improvement rather than a one-off fast changeover.
Metric to record: Trend in changeover time, variation, first-off quality and productive time recovered.
Example: After the new sequence is stable, the team finds that first-off inspection is the largest remaining delay and begins a focused improvement on measurement preparation and feedback.
The following example is illustrative. It shows how a manufacturer could apply SMED to a recurring changeover on a CNC milling centre; it is not a FourJaw customer result.
The baseline changeover, from the last good Part A to the first good Part B, takes a median of 55 minutes and occurs ten times each week.
Changeover activity |
Current stopped time |
Improvement |
New stopped time |
|
Remove the previous fixture and tools |
12 min |
Standard tool cart and simplified fastening |
7 min |
|
Find and collect the next fixture, tools and gauge |
8 min |
Prepare and verify before the machine stops |
0 min |
|
Install and align the next fixture |
15 min |
Locating pins and a defined setup method |
8 min |
|
Find and load the programme and offsets |
5 min |
Prepare externally; verify at the machine |
1 min |
|
Trial cut, inspect and adjust the first part |
15 min |
Prepared inspection equipment and standard settings |
8 min |
|
Total |
55 min |
24 min |
The change reduces stopped time by 31 minutes:
55 minutes − 24 minutes = 31 minutes recovered per changeover
At ten changeovers per week, that releases:
31 minutes × 10 = 310 minutes, or 5 hours 10 minutes, per week
If the machine produces eight good parts per running hour, the released time represents a theoretical opportunity to produce approximately 41 additional parts each week:
5.17 hours × 8 good parts per hour = approximately 41 good parts
That figure should not automatically be recorded as output gained. The actual benefit depends on customer demand, labour, material availability, downstream capacity and whether this machine is a constraint. Productive time recovered is the direct result; extra output is one possible use of that capacity.
Use this checklist before, during and after a changeover.
Has the next job, product and changeover standard been confirmed?
Are the correct materials, tools, fixtures and gauges available?
Have tools, settings, programmes or components been prepared where possible?
Is the work area ready and are roles clear?
Have any safety, quality or maintenance issues been addressed?
Is the agreed task sequence being followed?
Are unnecessary walking, searching, waiting or adjustments occurring?
Are parallel tasks safe, practical and clearly allocated?
Are isolation, guarding and quality requirements being followed?
Are unexpected problems and delays being recorded?
Has the first good part been produced and approved?
What was the total changeover time?
Was the result within the expected range?
Was there startup scrap, rework or repeated adjustment?
What should be changed before the next changeover?
Has the standard been updated if a proven improvement was adopted?
Download FourJaw’s free SMED exercise posters to reinforce the seven-step process on the factory floor.
Average changeover time alone can hide important differences between jobs and shifts. Use a small group of measures to understand speed, consistency and quality.
Metric |
What it tells you |
|
Median changeover time |
The typical time from the last good part to the first good part |
|
Changeover variation |
Whether the method is stable and repeatable across events |
|
Productive time recovered |
How much additional machine availability the change has created |
|
First-time-right rate |
How often the first setup produces an approved part without further adjustment |
|
Start-up scrap and rework |
Whether faster changeovers are affecting quality |
|
Internal setup time |
How much of the changeover still requires stopped equipment |
|
Schedule adherence |
Whether more reliable changeovers are helping production meet the plan |
A SMED exercise requires two kinds of evidence: observation of the work and reliable production timing.
Direct observation reveals the manual tasks, movement, waiting and decisions within the changeover. Machine monitoring software provides an objective record of when equipment was running, idle or stopped, helping the team establish a baseline and track whether the new method has reduced downtime over repeated changeovers.
FourJaw automatically captures machine status, utilisation and downtime. Operators can add context by selecting predefined downtime reasons, helping production teams distinguish changeovers from breakdowns, material shortages and other stoppages. Teams can then compare changeover duration across machines, shifts and periods, investigate recurring delays and check whether an improvement has been sustained.
Machine data does not show every manual activity inside a setup, so it should not replace observing the work or speaking with operators. Used together, machine data and team observation answer two different questions:
Production data: When did the machine stop, restart and return to productive running?
Observation: What happened during that time, and why?
Learn more about using downtime monitoring to identify and prioritise production losses.
The purpose is to improve the process, not pressure people to move faster. Focus on preparation, task design, equipment, information and standardisation.
Ending the clock at startup can hide adjustment time, trial cycles and quality losses. Measure through to the first approved part from the new run.
An internal task converted to external work still needs an owner, time and resources. If preparation is not built into the production routine, the delay will return.
Do not remove necessary checks, bypass controls or create unsafe parallel work. Track safety and first-off quality alongside time.
Test a change under real conditions before making it the new standard. Once it is proven, document and train it promptly so the gain is not dependent on one person.
Single-digit setup is the origin and aspiration behind the name, not a universal pass-or-fail threshold. A safe, sustained reduction from the relevant baseline can still create substantial value.
SMED gives manufacturers a structured way to reduce the non-productive time between one production run and the next. Its central idea is straightforward: prepare what you can while the equipment is running, simplify what must happen while it is stopped, and turn the improved method into repeatable standard work.
Start with a recurring changeover where shorter, more consistent setup would improve production flow. Observe it with the people who perform the work, establish an accurate baseline, test one improvement at a time and measure the effect through to the first good part.
The objective isn't to just get a faster changeover. It is instead to get a safer, more reliable and more flexible production process that releases capacity and helps the factory respond to demand.
We hope you found this article insightful. Thanks for reading.