Lean Fundamentals · A Visual Explainer  

The Two-Second
Pit Stop

A pit crew changes four tires in seconds — because everything that can happen before the car stops already has. That is the entire idea behind SMED, and it is how a lean plant makes changeovers cheap enough to stop batching.

Scroll

Ask why a plant runs in big batches and the honest answer is usually the changeover. Swapping a die, a fixture, or a recipe takes so long that nobody wants to do it twice, so they make a giant run to amortize the pain. The batch isn't the disease — the slow changeover is.

SMED — Single-Minute Exchange of Die — is Shigeo Shingo's method for getting any changeover down into the single-digit minutes. It doesn't ask people to rush. It re-sequences the work so most of it happens while the machine is still running. Scroll to watch one changeover shrink.

The baselineHere's one changeover, drawn as time. Six tasks (see the key) — fetch the new die, find tools, remove the old die, mount the new one, dial in settings, run a trial. Done the usual way, every minute happens with the machine stopped: 35 minutes of pure downtime.
Move 1 · SeparateBefore changing anything, just look. Which tasks truly need the machine stopped, and which only happen then out of habit? Steps 1–3 — fetching the die, gathering tools, pre-setting the next die — could all be done while the current job still runs (now outlined in blue).
Move 2 · ConvertSo move them off the clock. Steps 1–3 lift up to the “while running” lane — staged before the line ever stops. The same work happens, just not during downtime. Downtime falls from 35 to 18 minutes without anyone working faster.
Move 3 · StreamlineNow attack what's left. Bolts become one-touch clamps, settings become numbered presets instead of trial-and-error, and two people work in parallel. The remaining downtime collapses under the 10-minute “single-minute” line — about 8 minutes.
Same crew, same toolsNobody hurried. The changeover went from 35 minutes to 8 purely by re-sequencing the work and simplifying the fasten-and-adjust steps. A changeover this cheap changes what's possible: you can switch products often, run tiny batches, and stop drowning in inventory.

The scoreboard

Same people, same machine, same tasks. Only the sequence — and a few fasteners — changed.

ResultBeforeAfter SMED
Machine downtime per changeover35min8min77% less
Setup work done while running0min17minmoved off the clock
Changeovers you'd run per shift1severalsmaller batches become affordable

Every SMED effort, on any equipment, is the same three moves in order. You can run them on a stamping press, a CNC tool change, a printing line — or a hospital OR turnover.

01

Separate

Watch a real changeover and split every task into internal (needs the machine stopped) and external (could be done while it runs). Most teams find a third to a half is external work trapped inside the stop.

02

Convert

Move that external work outside the stop. Stage materials, pre-heat and pre-assemble the next setup, lay out tools on a shadow board. Nothing about the task changes — only when it happens.

03

Streamline

Shrink whatever downtime remains: quick clamps instead of bolts, numbered presets instead of trial runs, parallel hands, standardized heights. This is the only move that touches the tools — and it comes last.

A fast changeover isn't about squeezing a few more parts out of the press. Its real payoff is that it removes the reason to batch in the first place.

Batch size follows changeover cost. When switching costs 35 minutes of downtime, you run huge batches to spread that cost over thousands of parts. When it costs 8, you can switch often and run small — the move toward one-piece flow that batching always blocked.

Inventory and lead time fall with it. Small batches mean less work parked in buffers, less cash tied up, and a customer order that flows through in hours instead of weeks. SMED is the enabler that makes a pull system and level scheduling realistic.

Flexibility becomes free. A line that changes over in minutes can chase today's demand mix instead of last month's forecast — making exactly what sold, when it sold.

And like every lean idea, it travels. A deployment that takes a weekend gets released quarterly; a deployment that takes minutes ships ten times a day. A meeting that needs an hour of setup happens monthly; one that needs none happens whenever it's useful. Wherever switching is expensive, work piles into batches — and wherever you make switching cheap, the batches melt away.

Good to know

Frequently asked

What does SMED stand for?
Single-Minute Exchange of Die — a method, developed by Shigeo Shingo, for getting any equipment changeover down into the single-digit minutes (under ten).
What is the difference between internal and external setup?
Internal setup is work that can only be done while the machine is stopped (removing and mounting a die). External setup is work that can be done while the machine is still running (fetching the next die, laying out tools, pre-setting). SMED converts as much internal work to external as possible.
What are the three steps of SMED?
First separate internal from external setup, then convert as much internal work to external as you can, then streamline both — quick clamps instead of bolts, presets instead of trial runs, parallel work.
Why does faster changeover matter if the work is the same?
Batch size follows changeover cost. A long changeover forces big batches to amortize it; a short one lets you switch often and run small batches, which cuts inventory and lead time and makes one-piece flow possible.
Is SMED only for big stamping presses?
No. The same three moves apply to a CNC tool change, a kitchen line switch, a hospital operating-room turnover, or a software deployment — anywhere switching from one job to the next is slow.
MS
Matthew Savas

Founder of Kaizumi, an AI-powered Lean training platform. More about Matthew →

Updated June 20, 2026 · Drafted with AI assistance and reviewed by Matthew Savas for accuracy.