Key takeaways
Changeover time is the elapsed time between the last good part of one production run and the first good part of the next. The SMED method shrinks it by separating internal setup steps (machine stopped) from external ones (machine running), then converting and streamlining them, which directly lifts the Availability component of OEE.
Every minute a machine sits idle between two product runs is capacity you paid for and did not get. On high-mix lines, changeover is often the single largest, most controllable chunk of lost production time. This guide defines changeover time precisely, walks through the SMED method step by step, shows a worked before/after example, and explains how cutting changeover raises OEE and unlocks smaller batches, all in claims-safe language you can take to the floor.
Changeover time is the elapsed time from the last good part of the prior run to the first good part of the next run. That definition matters: it does not stop when the operator finishes the physical swap. It keeps running through every adjustment, test piece, and warm-up cycle until the line is producing sellable parts again, at rate and at quality.
This "last good part to first good part" boundary is the standard way lean practitioners measure setup, because it captures the hidden tail of trial-and-error tuning that informal stopwatch readings miss, per Vorne's Six Big Losses reference.
Changeover is also called setup time. Some teams separate "internal setup" (work done while stopped) from total changeover, but for OEE accounting the wall-clock gap is what counts, because that is the time the machine is not making good product.
The whole of setup reduction hinges on one idea: not all changeover work has to happen while the machine is stopped. Lean splits every changeover task into two buckets.
Most plants run changeovers where huge amounts of genuinely external work, walking to the tool crib, hunting for a missing bolt, reading the spec, are done after the machine stops. That converted-by-default waste is the easiest minutes you will ever recover.
SMED stands for Single-Minute Exchange of Die, a structured method for dramatically reducing changeover time. "Single-minute" does not mean every changeover must take one minute. It means the goal is to get changeover into the single digits, under 10 minutes, according to the Lean Enterprise Institute lexicon.
SMED was developed by Japanese industrial engineer Shigeo Shingo through work in the 1950s and 1960s. His central insight has two moves: first separate internal setup tasks from external ones, then convert as many internal tasks as possible into external tasks so they happen off the critical path while the machine still runs.
Practitioners typically run SMED as three sequential, repeatable phases, as laid out by Lean Production's SMED guide:
The payoff is well documented. Across a wide range of companies, Shingo's pioneering work led to changeover reductions averaging 94%, for example from 90 minutes to under 5 minutes, per Lean Production. Treat that as the historical benchmark for what is achievable, not a guarantee for any specific line.
Consider an injection-molding press changing from Mold A to Mold B. Here is the same changeover before and after a SMED workshop.
| Changeover step | Before (machine stopped) | After SMED |
|---|---|---|
| Find and fetch next mold from rack | 8 min, internal | 0 min, done externally on a staged cart |
| Gather clamps, hoses, water fittings | 6 min, internal | 0 min, pre-kitted externally |
| Pre-heat mold to temperature | 12 min, internal (heat after install) | 0 min, pre-heated externally before stop |
| Remove Mold A, install Mold B, connect | 15 min, internal | 9 min, internal with quick-release clamps |
| Adjust shot size and pressures by trial | 14 min, internal | 3 min, internal from a saved recipe |
| Run scrap until first good part | 5 min, internal | 2 min, internal |
| Total changeover time | 60 min | 14 min |
No new capital equipment was required for most of the gain. The biggest savings came from converting internal work to external (staging, kitting, pre-heating) and from standardizing settings so the operator stops guessing. That is the SMED pattern in miniature.
Changeover is one of the Six Big Losses, the loss taxonomy that maps onto OEE. Specifically, changeover (a planned stop) reduces the Availability factor, because the equipment is scheduled to produce but is not running good parts during the swap. To see exactly how Availability is calculated and where setup time lands, see our guides on the Availability metric and the Six Big Losses framework.
Because OEE = Availability x Performance x Quality, every minute of changeover you remove flows straight into Availability and therefore into headline OEE. If a line runs 8 changeovers a week and you cut each by 45 minutes, that is 6 hours of recovered run time weekly with zero capital spend. For the full formula, see how to calculate OEE and the complete OEE guide.
There is a second, strategic prize. When changeover is expensive, the natural defense is to run large batches to amortize the cost, which inflates inventory, lengthens lead time, and hides quality problems. When changeover is cheap, you can run smaller batches more often, moving toward true flow and faster response to demand. SMED is therefore not just an Availability lever; it is the enabler of lean flow itself.
Cutting changeover also improves how usefully your capacity is consumed, a topic we cover in capacity utilization and machine utilization vs OEE.
Use this checklist to run a focused SMED event on your worst-offending changeover:
SMED workshops fail quietly when the gains are not sustained, and they are not sustained when changeover time is not measured continuously. If your only changeover number is reconstructed from a clipboard after the shift, you will never see drift, you will never know which line or product or crew is slipping, and you cannot prove the improvement held.
This is where real-time OEE tied to a closed-loop work-order system changes the game. Fabrico is a unified System of Action that connects directly to machine PLCs to capture cycle times and downtime as they happen. Changeover becomes a distinct, tagged downtime reason measured automatically from the last good part to the first good part, so its true cost shows up in your Availability and OEE numbers without manual logging.
Because Fabrico pairs live OEE with a full CMMS, a changeover that exposes a recurring problem, a worn clamp, a missing fixture, a chronic adjustment fault, can be turned into a prioritized, parts-ready digital work order on a technician's phone, with QR-enforced checklists so the standardized changeover method is actually followed. That is the fault-to-fix loop applied to setup, and it is how a one-time SMED win becomes a permanent capability. Fabrico is EU-built with EU data residency, a useful trust asset for manufacturers with sovereignty requirements. [INSERT VERIFIED PROOF POINT - operator to confirm]
To connect setup reduction to a broader reliability program, see the TPM 8 pillars guide, which frames focused improvement of losses like changeover within total productive maintenance.
If you want to see changeover tracked as a live downtime reason and converted into action on the floor, book a Fabrico demo and bring your worst changeover.
Changeover time is measured as the elapsed time from the last good part of the prior run to the first good part of the next run. This boundary matters because it includes setup, adjustment, trial pieces, and warm-up, not just the physical tool swap, so it captures the full cost of the changeover as the machine experiences it. Standardizing this measurement is essential for consistent tracking and for crediting OEE improvements correctly.
Internal setup includes steps that can only be performed while the equipment is stopped, such as removing and installing a die. External setup includes steps that can be performed while the machine is still running the previous batch, such as staging the next tool, kitting hardware, or pre-heating. The core of SMED is to first separate the two, then convert as many internal steps as possible into external steps so they no longer add to downtime.
SMED stands for Single-Minute Exchange of Die. The goal is to reduce changeover time to the single digits, meaning under 10 minutes, not literally one minute. It was developed by industrial engineer Shigeo Shingo through work in the 1950s and 1960s, and documented programs have averaged a 94% reduction in changeover time, for example from 90 minutes to under 5 minutes.
Changeover is one of the Six Big Losses and is a planned stop, so it reduces the Availability factor of OEE. Because OEE equals Availability multiplied by Performance multiplied by Quality, every minute of changeover you eliminate flows directly into Availability and raises overall OEE, recovering scheduled production time with no capital spend.
When changeover is slow and costly, plants run large batches to spread that cost across more parts, which inflates inventory and lengthens lead time. When SMED makes changeover fast and cheap, the economic penalty for switching products shrinks, so you can run smaller batches more frequently, move toward one-piece flow, and respond faster to real demand.
SMED gains erode when changeover is not tracked continuously. Reconstructing the number from a clipboard after the shift hides drift and makes improvement impossible to prove. Real-time OEE software that connects to machine PLCs records changeover automatically as a distinct downtime reason from the last good part to the first good part, so its true cost stays visible and standardized methods can be enforced through digital work orders.
Login
Blog