Single-Minute Exchange of Die

SMED is the strictest, most ambitious version of quick changeover. Shigeo Shingo developed the methodology at Toyota over the 1950s and 1960s, eventually achieving setup times that had been measured in hours coming down to under ten minutes. The name is sometimes mistranslated as "single-minute" meaning literally one minute; Shingo's actual target was "single-digit minutes," meaning nine or less. The discipline matters because long setups force long batches, and long batches create the inventory and lead time problems that lean exists to remove. A shop that masters SMED unlocks small-batch production at a profit, which changes what the schedule can do.

"If the setup is long, the batch will be long. Shorten the setup and the rest of lean becomes possible."

How SMED works

SMED walks through three structured stages. Each stage produces measurable setup-time reduction.

Stage 1: Separate internal from external. A real changeover is video-recorded. The team watches the video and tags every action as either internal (the press has to be stopped) or external (the press can still be running). In most baseline changeovers, 30 to 50 percent of the actions tagged as internal could actually be external. The operator was simply doing them inside the stopped window out of habit. Just identifying this gap and rescheduling those actions outside the stop usually cuts setup by a third without changing any equipment.

Stage 2: Convert internal to external. The team finds ways to physically move more actions outside the stopped window. Preheating the next mold during the current run. Staging the next setup's tooling on a cart that rolls into position when needed. Kitting components for the next job during the previous one. These conversions usually take a small investment in carts, staging space, or heating equipment but pay back quickly.

Stage 3: Simplify the remaining internal work. Once internal setup has been minimized, the work that has to happen with the press stopped is simplified. Color-coded clamps so the operator does not check spec sheets. Quick-release hardware replacing bolts. Standardized shim packs that eliminate measurement. Positioning fixtures that hold parts in the correct location without measurement. The remaining internal work shrinks toward the single-digit minute target.

Most shops hit a 50 percent reduction after Stage 1 and 2 in two weeks. The 90 percent reduction Shingo aimed at requires Stage 3 and often a second pass through the whole methodology.

Where SMED fits on the shop floor of a small manufacturer

Imagine a 25-person plastics injection molding shop running 14 different products for three customers. The main press has a mold change time of about 80 minutes. The shop runs batches of 4,000 parts to make the setup math work, which creates two weeks of WIP and 22-day lead times on small orders. The owner has been turning down small-quantity orders because the math is bad.

A two-week SMED project changes the picture. The shift lead films two changeovers. Stage 1 shows 28 of the 80 minutes are actions that did not require the press to be stopped: tool fetching, the next mold's preheat, the kit pull for the next job. The team builds a small mold cart that holds the next mold and its kit, with a preheat element that activates during the current run. The press is no longer waiting for those minutes.

Stage 2 converts more internal work to external. The operator's offset documentation is moved onto a laminated card attached to each mold, so the manual lookup happens before the press stops. The clean-down between runs is pre-staged with cleaning kit and rags positioned at the press during the previous job.

Stage 3 simplifies the remaining internal work. Standard quick-release clamps replace the threaded clamps that had been used. A color-coded shim pack replaces the spec-sheet-driven shimming. Setup drops to nine minutes.

The shop now runs minimum batches of 600. Small-order lead time falls from 22 days to four. The owner starts taking the small orders the shop had been turning down. The project cost about 1,200 dollars in materials and two weeks of operator attention.

Common mistakes with SMED

• Skipping the video. Memory underestimates setup time and misses the small walking and searching that dominate. Always video-record the baseline.
• Buying tooling before standardizing. Expensive quick-change fixtures imposed on a chaotic setup do not pay back. Standardize first, invest second.
• One-time project mindset. SMED is a recurring habit. First pass takes 50 percent, second pass takes another 30. Schedule the cycle, not just the kickoff.
• Improving the wrong machine. Setup reduction on a non-bottleneck does not show up in lead time. Pick the constraint first.
• Not adjusting batch size after the win. A setup that drops from 80 minutes to nine only matters if the shop actually runs smaller batches. Update the scheduling rules in tandem.

SMED and related Lean tools

SMED is the strictest methodology inside the broader quick changeover practice. The metric it directly improves is changeover time. The downstream benefit lands on the schedule through heijunka, which uses short setups to level production across product mix, and through mixed-model production, which is impossible without short setups.