What Is SMED?
SMED — Single-Minute Exchange of Die — is a method for radically reducing the time it takes to change a machine or process over from making one product to making another. It was developed by Shigeo Shingo while consulting for Toyota, where some stamping-press changeovers were cut from hours to minutes. "Single minute" does not mean one minute; it means a single-digit number of minutes — under ten — for a changeover that previously took far longer.
Changeover time is the elapsed time from the last good part of one run to the first good part of the next. Every minute of changeover is non-productive, and crucially, long changeovers force large batch sizes: if a setup costs an hour, you run a big batch to spread that cost over many parts. Big batches inflate inventory, stretch lead time, and bury defects. SMED breaks that logic by making setups so quick that small batches become economical, which is what unlocks just-in-time flow and product-mix flexibility.
Internal vs. External Setup
The single most important distinction in SMED is between two kinds of setup work:
- Internal setup (IED) — tasks that can only be performed while the machine is stopped, such as removing the old die and mounting the new one.
- External setup (OED) — tasks that can be performed while the machine is still running the previous job, such as retrieving the next die, gathering tools, and preheating molds.
In an unimproved changeover, much of what could be external is done internally — operators search for tools and fetch the next die after the machine has already stopped. Recognizing and reorganizing this distinction is where most of the gains come from.
The SMED Stages
Stage 0: Study the Current Changeover
Before improving anything, record the existing changeover in detail — ideally by video — with no distinction yet between internal and external work. List every element, its duration, and who performs it. This baseline reveals where time actually goes, which is rarely where people assume.
Stage 1: Separate Internal from External Setup
Classify each element as internal or external. Then ensure all external tasks are actually done before the machine stops or after it restarts — staging the next die at the press, laying out tools on a setup cart, and confirming materials are present. Simply doing external work externally, with no other change, commonly cuts changeover time by 30–50%.
Stage 2: Convert Internal to External
Examine each remaining internal element and ask whether it can be converted to external. Examples include preheating dies offline instead of warming them up on the machine, using duplicate fixtures so one is prepared while the other runs, and pre-assembling tooling modules. Conversion is the heart of SMED and where creative engineering pays off.
Stage 3: Streamline All Elements
Finally, attack everything that remains — both internal and external — to make it faster:
- Quick clamps replacing bolts (functional clamps, cam locks, one-turn fasteners)
- Standardized die heights and shut heights to eliminate shimming and adjustment
- Locating pins and stops so the new die seats in the correct position without measurement
- Parallel operations — two operators working different sides of the machine simultaneously
- Eliminating adjustment — the largest hidden cost, replaced by pre-set, repeatable settings
| Stage | Action | Typical Effect |
|---|---|---|
| 0 | Document baseline changeover | Reveals where time is spent |
| 1 | Separate internal from external | 30–50% reduction |
| 2 | Convert internal to external | Further large reduction |
| 3 | Streamline all elements | Approaches single digits / OTED |
From SMED to OTED
The aspirational next step is One-Touch Exchange of Die (OTED) — changeover in under one minute, ideally with a single motion. OTED is reached by engineering out all adjustment: standardized tooling, quick-release mechanisms, pre-set stops, and dies designed for drop-in mounting. Some processes go further toward "no-touch" automated changeover.
Why Setup Time Drives Batch Size
The economic batch quantity rises with setup cost. When setup is expensive, the math pushes toward large lots; when setup is cheap, small lots become viable. By collapsing changeover time, SMED collapses the economic batch size, enabling level scheduling (heijunka), lower inventory, and quick response to demand shifts. Because changeover is itself a repeatable task, you can measure and target it the same way you would any operation using a standard time analysis, then track improvement against that standard.
Common Pitfalls
- Skipping the video baseline — improving from memory misses the biggest time sinks.
- Stopping after Stage 1 — separating internal and external is only the first win; conversion and streamlining deliver the rest.
- Tolerating adjustment — "tweaking until it runs right" is often the single largest element and should be engineered out.
- Not standardizing the new method — gains erode unless the improved changeover becomes the documented standard with trained operators.