The work order says the changeover took 45 minutes. From the last good part to the first good part, it actually took 65 minutes.
I’ve seen that gap on dozens of shop floors. The paperwork says one number. The floor lives a different one. The 20 minutes in between is where most of your changeover loss is hiding.
Long setups get treated like a fact of life, but they don’t have to be. They’re waste, and waste can be found, sequenced, and cut.
Single-Minute Exchange of Die (SMED) is the method for doing that. This guide covers what a changeover actually measures, why the “7 steps” you’ve heard of aren’t quite Shingo’s original method, and what I’ve learned walking plants that get changeover right and the ones that don’t.
Key Takeaways
Single-Minute Exchange of Die (SMED) is a method for cutting the time between the last good part of one job and the first good part of the next. Shigeo Shingo built it at Toyota. The name doesn’t mean every changeover takes sixty seconds. It means Shingo’s target was a single-digit number of minutes, and that number was never meant to travel to every plant unchanged.
The word “die” in SMED refers to specialized tooling used in stamping, molding, casting, and forming operations. The SMED process extends far beyond traditional metalworking presses, though. Modern plant operations apply the SMED methodology to any setup, line changeover, or conversion activity, including switching packaging lines, cleaning equipment between pharmaceutical batches, or changing formats in food and beverage processing.
Shigeo Shingo developed SMED as a Toyota consultant.
He documented average changeover reductions of approximately 94% across the companies he worked with. His work helped establish SMED as one of the most recognised approaches to set up-time reduction within lean manufacturing.
SMED sits alongside other lean manufacturing practices such as 5S, kaizen, and value stream mapping. While each methodology addresses different sources of waste, they share a common principle, which is that they identify activities that don’t create value and find systematic ways to eliminate or reduce them.
The central idea behind SMED is powerful because changeovers often contain more waste than manufacturers initially realise.
A typical changeover might involve:
Some of those tasks genuinely require the machine to be stopped. Others don’t.
For example, an operator might spend several minutes looking for a tool after the machine has already stopped. Another operator might prepare the next set of materials only after production has ended.
From a SMED perspective, these activities represent opportunities. On paper, a changeover looks like one long stop. On the floor, however, it involves minutes hunting for the right tool, minutes swapping the mold cold, minutes dialing in settings by trial and error, or minutes ramping up while scrap piles up at the start. To illustrate, a 50-minute changeover might genuinely break down as roughly 8 minutes finding tools, 12 swapping, 10 adjusting, 9 on trial runs, and 11 ramping back up to speed. None of that gets faster by telling the crew to hurry.
SMED separates what has to happen while the machine is down from what doesn’t, moves the external work outside the stop, and tightens what’s left. Stage the tools while the line is still running. Preheat the mold before the stop. Pull proven settings instead of rediscovering them by trial and error. Do that consistently, and the same 50-minute changeover becomes a 9-minute one. Not because anyone moved faster, but because the sequence changed.
Changeover time gets treated like a fact of life because it’s planned. A breakdown gets a war room. A changeover gets a shrug.
That’s backwards. Planned downtime is still downtime. If a machine runs six 45-minute changeovers across three shifts, that’s 4.5 hours of production gone, on a day nobody called an emergency.
I don’t exclude planned stops when I calculate Availability, and you shouldn’t either. Every minute the machine isn’t producing counts, scheduled or not. Excluding “planned” losses to make an OEE number look better just hides the opportunity from the people who could fix it.
Several manufacturing trends are making quick changeover increasingly important:
| Challenge | Impact of Longer Changeovers |
| More SKUs | Greater product variety means more frequent production changes and setup periods. |
| Shorter production runs | Smaller batches can become uneconomical when each changeover consumes significant time. |
| Margin pressure | Every minute of unavailable equipment represents potentially lost production capacity. |
| Labour constraints | Fewer experienced workers make efficient, standardised changeovers more important. |
The manufacturing workforce is also changing as experienced employees approach retirement. This Silver Tsunami can create knowledge gaps, particularly when critical changeover expertise exists primarily in experienced technicians’ heads.
SMED helps capture that knowledge through:
Changeover time directly affects Availability, one of Overall Equipment Effectiveness (OEE)‘s three components:
This is also where the measurement window matters, and getting it wrong is more than a rounding error. Picture a changeover report that reads 42 minutes against a 40-minute target. Two minutes over. Not enough to open a CI case, so nobody does.
But if the clock actually runs from the last good part of the previous job to the first good part of the next job that holds quality, the real number is often much higher. Trial shots, early scrap, and settings adjustments made after the machine “restarted” on paper can easily add another 20 minutes that a stop-to-start clock never captured. That 42-minute report becomes a 62-minute changeover, which is 20 minutes over target instead of 2. What you measure determines what you prioritize, and the wrong window means the real opportunity never reaches the board.
Reducing planned changeover downtime can therefore improve OEE while increasing available production capacity.
The bigger opportunity is operational intelligence, which entails understanding which changeovers run long, why they take longer, and whether improvements hold over time. This is where traditional SMED methodology connects with modern manufacturing intelligence.
Even after a line clears its first good part, it’s often not back to normal. Machines commonly run below target speed for a while once the changeover “ends” on paper, as operators fine-tune the process and work through micro-stops nobody has bothered to name.
That phase shows up constantly, even when nobody calls it out:
None of that gets logged as changeover time, and none of it gets logged as a defined loss either. It gets absorbed into generic “minor stops,” folded into the overall OEE number, and discovered, if at all, weeks later at month-end review instead of the same shift it happened.
Most plants handle it like this:
Plants that get ahead of it do a few things differently:
Most plants run a SMED project at some point. Few ever put ramp-up on the same agenda, which is usually exactly why it’s sitting there as the next improvement case nobody has opened yet.
Shigeo Shingo’s original SMED framework has four parts: a preliminary stage, followed by three stages of improvement.
You’ll often see this called the “7 steps of SMED.” That version, which this guide also uses below, takes Shingo’s Stage 3 and splits it into separate internal and external streamlining steps, then adds three practices most plants layer on top to make the gains stick: documenting the new standard, training the crew, and measuring results over time. Those additions aren’t part of Shingo’s original three stages, but they’re what keeps a changeover improvement from sliding back to where it started.
Before modifying any setup routines, continuous improvement (CI) teams must record the current baseline process in detail. This step involves capturing video footage of a complete changeover from the final part of the previous production run to the first quality-verified part of the new run.
Every physical movement, tool fetch, mechanical adjustment, and equipment delay must be logged timestamp by timestamp. CI leads map who performs each action, what tools are used, and exact durations. Without an accurate, unedited baseline of current operations, setup teams risk optimizing minor steps while missing major structural delays.
The fundamental principle of the SMED lean manufacturing process is differentiating between internal and external elements:
During this stage, the team evaluates every logged baseline task to determine its true execution requirement. Plants running their first SMED event routinely find that a meaningful share of the tasks they’d always treated as “internal” were actually external tasks performed at the wrong time.
Once you know which is which, move as much internal work as possible into the external window. The goal: maximize what happens while the line is still running.
Tactics that work:
Move this work outside the stop, and you cut changeover time without spending a dollar on new equipment.
For what’s left inside the stop, cut the wasted motion. No searching. No trial and error.
Common optimization actions include:
Cut the guesswork here, and the machine gets back to full speed faster once it restarts.
External work still needs organizing, or it creates its own delay.
Key practices include:
Get this right, and the crew starts the internal work the second the machine stops. No hunting.
Write it down. A standard changeover template with the sequence, who does what, and how long each step should take.
Without a written standard, every shift runs the changeover its own way. That’s a system problem, not a people problem, and it’s the fastest way to lose a gain you already made.
Run drills on the new standard before it goes live. Let the crew practice in a controlled setting, not for the first time during a real changeover.
Training builds the skill. A system keeps the record. You need both. A crew trained once and never tracked again drifts back to the old way within a few months, and nobody notices until the changeover times creep back up.
Cutting changeover time once is an event. Keeping it cut is the job.
Track every changeover against the target after the new standard goes live, not just the first few after the kaizen workshop wraps. Compare actual against baseline. Catch variance early. Feed it back into the next kaizen round.
That ongoing review, not the original event, is what keeps the gain from eroding.
| Step | Activity | Focus | Key Output |
| Preliminary | Observe & Document | Baseline Recording | Activity log and video baseline, timed from last good part to first good part |
| Step 1 | Separate Elements | Task Classification | Categorized list of internal vs. external tasks |
| Step 2 | Convert Elements | Task Re-engineering | Shifted prep work done prior to machine stop |
| Step 3 | Streamline Internal | Waste & Motion Elimination | Reduced downtime via quick-releases and fixed stops |
| Step 4 | Streamline External | Logistics & 5S | Pre-staged kitting carts and standardized shadow boards |
| Step 5 | Document Standard | Standard Work Creation | Visual SOPs and explicit SMED task templates |
| Step 6 | Train & Practice | Skills Development | Cross-trained shifts working to a verified sequence |
| Step 7 | Measure & Improve | Continuous Tracking | Sustained setup speed and iterative kaizen feedback |
See how leading manufacturers use Shoplogix to monitor performance in real time, reduce downtime, and improve operational efficiency with actionable production insights.
A faster changeover doesn’t just look good on a report. Positioning SMED as a strategic business initiative allows operations teams to unlock capacity, decrease working capital requirements, and better serve customer demand.
Every hour cut from a setup is an hour of production you didn’t have yesterday. The cheapest new capacity in your plant is the line you already own, running right. No capital request needed.
Traditional manufacturing accounting encourages long production runs to amortize lengthy setup times over higher unit volumes. However, large production runs generate massive work-in-process (WIP) and finished goods inventory holding costs.
Cut the changeover, and smaller lots start making financial sense. Less inventory. Less cash tied up in product that hasn’t shipped yet.
Changeover time is an Availability loss. Cut it, and Availability goes up. That’s the fastest lever most plants have for moving OEE, faster than chasing speed or quality gains that take longer to show results.
A fast changeover is about sequence, not speed.
Get the sequence right, and switching between SKUs stops being a scheduling headache. Shorter lead times. Better on-time delivery. The kind of flexibility that lets a plant say yes to a short, high-margin order instead of turning it down.
Standardized changeover procedures reduce trial-and-error adjustment during production restarts. Traditional setups often generate off-spec scrap in the early minutes of a new run, as operators dial in settings manually. Structured SMED practices establish fixed mechanical stops and pre-validated parameters, cutting down on that startup quality loss.
The same ramp-up period that shows up as a Performance loss on OEE is often where startup scrap concentrates too. That’s one more reason to track it as its own metric instead of folding it into the changeover number.
Paper logs and stopwatches make SMED initiatives hard to sustain. Data goes stale, gets recorded inconsistently across shifts, or lives in a spreadsheet nobody opens after the kaizen event wraps up. Shoplogix’s SMED Insights package turns changeover tracking into a live, always-on part of the plant’s daily management system.
SMED Insights tracks every changeover from last good part to first good part, automatically, on every job and every line.
When a line conversion exceeds its target time, automated alerts keep minor delays from turning into major schedule disruptions:
To ensure changeover improvements remain permanent, Shoplogix connects performance tracking directly to accountability workflows:
A dashboard nobody reacts to is decoration. Most SMED programs stall for exactly that reason. The data to answer “why did that changeover run long” exists. It’s just scattered across screens, and getting a straight answer usually means waiting on an analyst.
Pulse AI is the AI assistant built inside Shoplogix. It answers changeover questions in plain language, grounded in a plant’s own SMED Insights data.
A supervisor prepping for shift handover can ask “why did the changeover on Line 3 run long this morning” or “how does today’s changeover compare to last week’s” and get an answer in seconds, with the numbers behind it. Pulse AI doesn’t change the data or take action on its own. It answers the question and points to the open Action Plan tied to the issue, so the follow-up is already teed up.
Your operators already know which job combinations run long. They’ve known for years. Pulse AI just makes that pattern visible to everyone on the team, not only the person who’s been running that line for a decade.
Pulse AI runs on a plant’s own live Shoplogix data and follows existing user permissions. The answers reflect that line’s own history, not a generic industry number. It’s never used to train models on a customer’s data. See how Pulse AI works.
A SMED event (or kaizen event) is a focused, short-term project targeting a specific machine or line to streamline its changeover process. A SMED programme is an enterprise-wide continuous improvement strategy that establishes standard quick changeover procedures, continuous monitoring, and structured training across all lines and facilities.
Internal elements are setup tasks that can only be performed while the machine is completely stopped, such as swapping tooling or removing dies. External elements are tasks that can be performed while the machine is actively running, such as staging tools, pre-heating molds, or gathering raw materials prior to line shutdown.
No. Although SMED originated in metal stamping operations using physical dies, the quick changeover methodology applies to any setup, line conversion, or sanitization activity across all manufacturing sectors, including packaging, food and beverage processing, pharmaceuticals, and discrete assembly.
Getting SMED right turns changeover from an invisible cost into capacity you already own. Separate internal and external work. Get the sequence right. Track every changeover from last good part to first good part instead of guessing at it.
The cheapest capacity in your plant isn’t a new line. It’s the one you’re already running, done right.
See how manufacturers use Shoplogix to gain real-time production visibility, resolve issues faster, and empower operators through our library of on-demand demos.
Skip to content
