7 Lean Manufacturing Improvements That Cut Production Downtime

Early in my career, I thought downtime was a maintenance problem. A machine breaks, a technician fixes it, and the line starts again. Then I spent three weeks timing a packaging line with a stopwatch, and that idea fell apart. That project was my real introduction to lean manufacturing process improvement.

The breakdowns were real, of course. But they were the smallest share of the time we lost. Instead, most of the loss was in plain sight. Operators spent twenty minutes looking for change parts. Meanwhile, crews waited for a lead to sign off a first article. One sensor faulted six times an hour, and nobody logged it anymore because resetting it had become part of the job.

That is the trouble with downtime. Big failures get everyone’s attention, while small ones slowly eat your capacity.

What Downtime Really Costs Manufacturers

The industry numbers are hard to ignore. The Siemens True Cost of Downtime 2024 report estimates that the world’s 500 largest companies lose about $1.4 trillion a year to unplanned downtime. In other words, that is roughly 11% of their revenue. In a large automotive plant, a single idle line can cost around $2.3 million an hour.

Similarly, a 2025 survey commissioned by Fluke points the same way. In it, 61% of manufacturers reported unplanned downtime in the past year, at an average cost of $1.7 million per hour.

You don’t need to run a car plant for those figures to matter. Scale them down to a 40-person job shop and the math still hurts.

Where These Seven Lean Manufacturing Process Improvement Practices Come From

Below are the seven lean manufacturing process improvement practices I have seen work most reliably. I have used them on stamping lines, food packaging lines, machining cells, and in plants that told me they had “already done lean.”

None of these ideas are new. In fact, most were written down decades ago. What separates plants that get results from plants that don’t is rarely the tool. Rather, it is the discipline around the tool.

Why Lean Manufacturing Process Improvement Belongs at the Center of Your Downtime Strategy

Before the seven practices, it helps to understand what lean actually does to downtime.

Lean doesn’t treat downtime as one big problem. It breaks lost time into categories so you can match each one with the right fix. My preferred framework is the “Six Big Losses,” which sits under Overall Equipment Effectiveness (OEE).

• Availability losses: equipment failure, and setup or adjustments.
• Performance losses: minor stops, and reduced speed.
• Quality losses: process defects, and startup scrap.

So when a plant manager tells me downtime is killing them, my first question is always “which kind?” A plant with long changeovers needs a very different fix than a plant with constant micro stops. For example, a predictive maintenance platform won’t shorten a 90-minute die change. Likewise, a SMED workshop won’t stop a gearbox from failing.

Good lean manufacturing process improvement starts with that diagnosis. After all, everything else follows from it.

1. Measure Downtime Honestly Before Any Lean Manufacturing Process Improvement

This step sounds obvious. It is also the one most plants skip.

Most facilities believe they know their downtime. Usually, there is a number on a board somewhere. When I dig into it, though, that number only counts stops longer than five or ten minutes, because those are the ones that end up in the maintenance system.

Short stops, on the other hand, almost never make it into a report. An operator clears a jam in 45 seconds and keeps going. Yet OEE guidance defines minor stops as pauses of a minute or two, well under five minutes. On a high-speed line, those small stops can add up to more lost time than all the breakdowns combined.

What to Put in Place First

On almost every project, I start with three things:

• A short, structured list of reason codes. Aim for 12 to 20 codes, grouped under the Six Big Losses. If operators have to scroll through a long menu, they will pick “Other” every time.
• Automatic capture wherever you can get it. Most plants already have more data than they use. If your lines run on modern industrial automation systems, the PLC already knows every time a machine stops. Logging those run and stop signals reveals the true count of micro stops, and the operator only has to add the reason. The machine tells you when, and the person tells you why.
• A weekly Pareto chart that leaders actually review. Data sitting in a spreadsheet changes nothing. So post the top three downtime reasons on a board the plant manager walks past every morning.

On one bottling line, we found that a single capper jam caused nearly a third of all lost time. It had never shown up in maintenance records, because each event lasted less than two minutes. Once the team could see it, the fix took a week. In short, that is what honest measurement gives you: the right target.

2. Use SMED as a Lean Manufacturing Process Improvement Tool for Changeovers

In a high-mix plant, changeovers are often the biggest block of planned downtime. They are also where I have seen the fastest gains, which is why SMED is usually the first lean manufacturing process improvement tool I reach for.

SMED stands for Single Minute Exchange of Die. Shigeo Shingo developed the method in the 1950s and 1960s. The Lean Enterprise Institute describes its goal as a changeover in single-digit minutes, meaning under ten. Whether or not you hit that target, the method works on almost any equipment, not just presses.

Internal and External Setup

SMED splits changeover work into two types:

• Internal setup can only happen while the machine is stopped. Swapping a die, for example, is a typical internal task.
• External setup can happen while the machine is still running. That includes staging the next die, gathering tools, and printing paperwork.

From there, the process has three steps. First, film a real changeover and list every action. Second, move as much work as possible from internal to external. Third, streamline what remains inside the stop with quick-release clamps, standard heights, preset tools, and two people working in parallel.

The first filmed changeover almost always surprises the team. On one recent job, a 72-minute changeover included 19 minutes of walking. Operators kept leaving the cell for an Allen key, a gauge, a torque wrench, and shims, one trip at a time. Yet none of that was internal work. A shadow board and a changeover cart cut those 19 minutes to nearly zero before we touched a single fixture.

Lessons From Running SMED Events

A few lessons I learned the hard way:

• Involve the operators who do the changeover. They know which steps are padding and which ones protect quality. By contrast, an engineer redesigning the process from a desk will miss things.
• Standardize the new method, or it will drift back. Post a changeover procedure with photos at the machine. That document is what keeps a 25-minute changeover at 25 minutes six months later.
• Use shorter changeovers to change your schedule. Once changeovers are cheap, you can run smaller batches, carry less inventory, and respond faster to customers. Above all, that is where the real money is.

3. Build Total Productive Maintenance Into Daily Operations

Equipment failure is the downtime everyone notices. Total Productive Maintenance (TPM) is lean’s answer to it. Pay attention to the word total: TPM is a plant program, not a maintenance department program.

Both the Lean Enterprise Institute and the U.S. EPA describe TPM as a way to make equipment reliable by involving everyone, especially operators. In my experience, the part that matters most is autonomous maintenance, where operators take ownership of basic cleaning, inspection, lubrication, and tightening on their own machines.

Why operators? Because they stand next to the equipment for eight to twelve hours a day. They hear a bearing start to whine. An oil spot on the floor catches their eye, and even a small change in vibration is something they feel. In contrast, a technician who walks past once a week will never catch those early signs.

How to Roll Out Autonomous Maintenance

Here is the rollout I usually follow:

• Start with a deep clean. It sounds basic, but cleaning is inspection. During the first clean, teams routinely find loose bolts, cracked hoses, missing guards, and leaks. Tag every issue and track it until it is closed.
• Write short, visual checklists. A daily check should take five to ten minutes and fit on one laminated card, with photos showing what “good” looks like.
• Move technicians to higher-value work. With operators handling routine care, skilled technicians can focus on planned maintenance, failure analysis, and upgrades instead of running from one emergency to the next.

Where Predictive Maintenance Fits in Lean Manufacturing Process Improvement

TPM also opens the door to smarter maintenance strategies. The Pacific Northwest National Laboratory prepared an Operations and Maintenance Best Practices guide for the U.S. Department of Energy. It notes that preventive maintenance can save 12% to 18% compared with a purely reactive approach, and that a working predictive program can save another 8% to 12% over preventive alone. Moving from reactive to predictive can deliver savings of more than 30% to 40%.

This also matches what I see on the floor. The plants that get out of firefighting mode are the ones that finally find time to improve.

One word of caution, however. I have watched plants spend six figures on vibration sensors while operators still didn’t know how to check an oil level. Predictive tools certainly amplify a good maintenance culture. They don’t replace one, and no sensor can stand in for real lean manufacturing process improvement on the floor.

Over the longer run, the maintenance records your TPM program builds also feed industrial equipment lifecycle management, which tells you when an aging machine is worth rebuilding and when it is time to replace it.

4. Use 5S and Visual Management to Kill the Search

Search time is one of the most overlooked sources of downtime. It rarely gets a reason code. Instead, it just shows up as an idle machine while someone looks for something.

5S stands for Sort, Set in order, Shine, Standardize, and Sustain. Many plants treat it as housekeeping. Done well, however, it is a downtime reduction tool. EPA lean guidance frames 5S as a way to cut waste and improve workflow, largely by making problems visible.

My test is simple. Can a new operator find everything needed for a changeover in under a minute, on day one, without asking anyone? If not, downtime is waiting to happen.

5S Practices That Pay Off

These practices consistently pay off:

• Point-of-use storage. Tools and parts live at the machine, not in a central crib 200 feet away.
• Shadow boards and labeled locations. When a wrench goes missing, everyone sees it right away.
• Floor markings and min/max levels. Material handlers can tell at a glance when to restock, so machines never starve for parts.
• Visual status at every machine. With a stack light or status board, supervisors can see which machines are running, waiting, or down without walking the whole floor.

Sustain, the last S, is where most programs die. Adding more audits won’t fix it, though. Build the 5S standard into daily work instead, with a short walk at shift start and a supervisor who actually cares. When leaders stop looking, the floor stops maintaining.

5. Stop and Escalate Fast With Andon and Jidoka

I see this pattern all the time. First, a problem occurs, and the operator spends fifteen minutes trying to fix it alone. A lead tries for another ten. Eventually someone calls maintenance, and the technician is on the other side of the building. By then, a five-minute problem has become a forty-minute stop.

Lean tackles this with two connected ideas. Jidoka, often described as “automation with a human touch,” means designing equipment and processes to stop on their own when something goes wrong, so defects aren’t passed down the line. Andon is the signal, such as a light, cord, or button, that calls for help the moment a problem appears.

Building the Response System

The hardware is simple. What actually makes andon work is the system behind it. When I help a plant set one up, we define four things:

• Who responds to each type of call: quality, maintenance, materials, or supervision.
• How fast they must arrive, measured in minutes rather than “when they can.”
• What happens next if the first responder can’t solve it within a set time, so the problem escalates automatically.
• How each call gets logged, so recurring calls become improvement projects.

Culture matters even more than hardware, which is true of every lean manufacturing process improvement effort. If operators get criticized for pulling the andon, they will stop pulling it, and problems will go underground. In the best plants I have worked in, an andon pull is good news. It means someone caught a problem early, while it was still small and cheap.

6. Mistake-Proof Setups and Startups With Standard Work and Poka Yoke

A surprising amount of downtime is self-inflicted. Someone loads the wrong part. Occasionally, a setting is off by one digit. Other times, a fixture goes in backward. Then the line starts, makes scrap, and stops again for troubleshooting. The OEE framework counts part of this as reduced yield: the scrap made between startup and stable production.

Standard Work and Poka Yoke Together

Two lean tools work well together here.

Standard work captures the current best way to do a task, including the sequence, the timing, and the key points that affect safety and quality. However, it isn’t a thick binder nobody reads. Good standard work fits on one page, uses photos, and lives at the workstation.

Poka yoke, or mistake-proofing, designs the process so errors are either impossible or instantly obvious. The Lean Enterprise Institute describes it as a way to help operators avoid mistakes, such as picking the wrong part or leaving one out. Examples I use often:

• Asymmetric locating pins, so a fixture fits only one way.
• Barcode or RFID checks that confirm the right program and tooling before a machine will cycle.
• Go/no-go gauges at first article inspection, which remove the judgment call.
• Part-presence sensors that confirm a part is seated before a press closes.

Most of these are design problems, not training problems. I once worked with a machining cell that lost almost an hour a week to first-piece problems after changeovers. Every issue traced back to three setup errors, and all three were preventable. Retraining alone wouldn’t have fixed them, so we took a different route. We treated it as an industrial engineering design problem: we rebuilt the fixtures with keyed locators so they could only go in one way, and the controller got a program check before the first cycle. Within a month, those incidents were close to zero. Notably, nobody worked harder. The process simply stopped allowing the mistake.

7. Make Root Cause Problem Solving Your Daily Lean Manufacturing Process Improvement Habit

The first six improvements will cut your downtime. This last one, however, keeps it down.

Every plant solves problems. The real question is whether it solves them once or for good. Too often, a failure gets fixed and the line restarts. Three weeks later the same failure is back, because nobody asked why it happened.

Three Daily Problem-Solving Habits

Lean relies on kaizen: continuous small improvements made by the people closest to the work. The EPA describes kaizen as engaging workers in quick, focused improvements. In practice, I rely on three habits:

• Tiered daily meetings. Hold a ten-minute stand-up at the cell each shift to review yesterday’s downtime, today’s risks, and open actions. If the cell can’t solve an issue, it goes to area leaders, then to plant leadership if needed. Problems keep moving instead of sitting.
• Honest use of the 5 Whys. Keep asking why until you get past the symptom. “The motor tripped” is not a root cause. “The motor tripped because the conveyor jammed, because a guide rail was loose, because it was never on an inspection checklist” gets you somewhere useful.
• Closing the loop in the standard. Every true root cause should end in a lasting change: a new checklist item, an updated work standard, a design change, or a poka yoke. If the fix lives only in someone’s memory, it will be forgotten.

The NIST Manufacturing Extension Partnership supports small and midsize manufacturers across the United States, and it has documented many cases where a continuous improvement culture produced lasting gains in productivity and capacity. That matches my own experience, too. Tools start the improvement, while daily habits keep it going.

How to Sequence Lean Manufacturing Process Improvement in a Real Plant

You can’t tackle all seven at once, and you shouldn’t try. This is the order I usually recommend:

1. Start with measurement (#1). Two to four weeks of honest downtime data will show where your losses really are.
2. Attack the biggest loss category first. If changeovers dominate, run SMED (#2). When breakdowns dominate, launch TPM on the worst machine (#3). Should micro stops lead the list, dig into the top causes with root cause analysis (#7).
3. Layer in 5S (#4) and standard work (#6) as the foundation. Together, these make every other improvement stick.
4. Build the andon response system (#5) once the basics are in place. Andon only works when a team is ready to respond.
5. Prove it in one pilot area, then spread it. Nothing sells lean manufacturing process improvement to a skeptical workforce like results from the line next door.

In my experience, a focused plant can cut downtime noticeably within 90 days on a pilot line. Sustaining those gains for years is a longer job, and it depends entirely on leaders staying involved.

Final Thoughts on Lean Manufacturing Process Improvement

After years of this work, I’m convinced that downtime is mostly a management problem wearing a maintenance uniform. Indeed, the machines are rarely the real issue. What matters is whether a plant can see its losses, respond quickly, and learn from every stop.

None of these seven practices are complicated. Measure honestly. Shorten changeovers. Share ownership of equipment. Organize the workplace. Escalate fast. Design out mistakes. Solve problems at the root, every day. Plants that do these things consistently do more than cut downtime. They also build a workforce that knows how to improve on its own.

If you remember one thing, make it this: go watch your line for an hour with a notebook, and write down every stop, no matter how short. What you find will show you exactly where your lean manufacturing process improvement journey should begin.

Lean Manufacturing Process Improvement FAQs

What is lean manufacturing process improvement?

Lean manufacturing process improvement is a structured way to remove waste from production, including lost time, excess inventory, defects, and unnecessary motion. It uses tools such as SMED, TPM, 5S, standard work, and kaizen, backed by a culture of daily problem solving.

What is the biggest cause of production downtime?

It depends on the plant, which is why measurement comes first. In high-mix operations, changeovers are often the largest loss. On high-speed lines, minor stops can add up to more time than breakdowns. The Six Big Losses framework helps you sort downtime and find your true largest cause.

How much can SMED reduce changeover time?

The classic SMED target is a changeover under ten minutes. Results vary by equipment and starting point, but a first SMED event often cuts changeover time sharply just by moving work outside the stop and eliminating searching and walking.

What is the difference between TPM and preventive maintenance?

Preventive maintenance is a schedule of tasks, usually carried out by the maintenance department. TPM, however, is broader. It includes preventive maintenance, but it also gives operators ownership of daily equipment care and focuses the whole plant on eliminating equipment losses.

How is OEE related to downtime?

OEE multiplies availability, performance, and quality to show how much planned production time is truly productive. Breakdowns and changeovers lower availability, while minor stops and slow running lower performance. As a result, tracking OEE with reason codes shows exactly where your downtime comes from.

How long does it take to see results from lean improvements?

Focused pilots, such as a SMED event or a TPM launch on one machine, can show results within weeks. Plant-wide results usually take months, and sustaining them requires daily habits like tiered meetings, standard work audits, and ongoing root cause analysis.

Do small manufacturers benefit from lean?

Yes. Small and midsize manufacturers often benefit the most, because every lost hour of capacity hits them directly. Programs like the NIST Manufacturing Extension Partnership exist specifically to help smaller manufacturers apply lean methods.

References

Downtime Cost Data

1. “The True Cost of Downtime 2024: A Comprehensive Analysis” (summary of the Siemens report). AEMT
2. “Unplanned Downtime Costs Manufacturers Up to $852M Weekly” (2025). Fluke Reliability

Lean Methods and Maintenance Practice

3. “How to Address the Six Big Losses.” OEE.com
4. “Single Minute Exchange of Die.” Lean Enterprise Institute
5. “Total Productive Maintenance.” Lean Enterprise Institute
6. “Lean Thinking and Methods: TPM.” U.S. EPA
7. “O&M Best Practices: Maintenance Approaches.” Pacific Northwest National Laboratory
8. “Lean Thinking and Methods: 5S.” U.S. EPA
9. “Jidoka.” Lean Enterprise Institute
10. “Poka Yoke.” Lean Enterprise Institute
11. “Lean Thinking and Methods: Kaizen.” U.S. EPA
12. “Manufacturing Extension Partnership.” NIST