How to Reduce Machine Downtime: 15 Proven Ways and a Worked Example

Key takeaways

• Capture every stop automatically before you start fixing anything, because manual logs miss many of the short stops.
• In most plants, three or four causes account for most of the lost minutes. Find them with a Pareto chart and fix those first.
• Downtime is reduced in two ways: fewer stops (prevention) and shorter stops (faster recovery). You need both.
• Changeovers and waiting for material are downtime too, and they are often the cheapest minutes to win back.
• In the worked example below, three fixes recover 26 minutes per shift, which is more than 200 extra running hours a year.

What counts as machine downtime?

Machine downtime is any time a machine is expected to produce but does not.

It splits into two groups, and the difference matters for how you fix it.

Scheduled breaks and time with no orders are usually not counted as downtime. They are removed before you calculate availability.

Very short stops, often under two minutes, are usually tracked as micro stops. In OEE they usually count as a performance (speed) loss rather than an availability loss.

For the full definition and categories, see our guide to equipment downtime.

To choose a tool for tracking it, read our machine downtime tracking software guide.

What machine downtime really costs

Siemens' report The True Cost of Downtime 2024 estimates that the world's 500 biggest companies lose almost $1.4 trillion a year to unplanned downtime, equivalent to 11% of their revenues.

Its survey of 181 maintenance, engineering and IT professionals at large industrial organizations in automotive, FMCG, heavy industry and oil and gas found that an average large plant loses 27 hours a month to unplanned downtime.

Plants averaged 25 downtime incidents a month, based on responses collected from April 2019 to March 2023.

Your own number matters more than an industry average. Here is a simple way to estimate it.

How to calculate your cost of downtime per hour

Cost of one downtime hour = lost margin + idle labor + repair labor + parts and scrap for each stop

The figures below are illustrative and assume one stop in the hour.

If the hour is made of several stops, multiply the parts and scrap rows by the number of stops.

One rule keeps this honest: only count lost margin if the output cannot be made up.

If you recover it with overtime, the cost of the hour is the overtime, not the margin.

Measure it before you fix it: five numbers

These five numbers tell you whether you have a frequency problem or a duration problem.

Low MTBF means you need prevention. High MTTR means you need faster recovery.

Our article on MTBF and MTTR explains both in more detail.

The most common causes of machine downtime

15 ways to reduce machine downtime

The list is grouped by what each step does: see it, prevent it, recover faster, plan better.

Start by seeing it, then take the biggest causes first, wherever they sit on the list.

See it: ways 1 to 3

1. Capture every stop automatically

Manual logs catch the long breakdowns and miss almost everything else.

Take the signal from the PLC, a retrofit sensor or a camera, so every stop is recorded with its start time and duration. Our guide to machine monitoring systems compares the options.

2. Give every stop a clear reason

Build a short list of reasons that describe the symptom, not a guess at the cause.

"Capper jam" is useful. "Operator error" usually is not.

Our guide to downtime reason code design shows how to build the list.

3. Make stops visible in real time

A stop that everyone can see gets a faster response.

Use a line display or an andon system, and escalate automatically when a stop runs longer than an agreed limit.

Prevent it: ways 4 to 8

4. Attack the top causes first

Rank downtime by minutes lost over the last four weeks.

Work on the top two or three causes until they drop, then move down the list. A Pareto analysis makes the ranking obvious.

5. Base preventive maintenance on real usage

Calendar schedules service a machine whether it ran 20 hours or 200.

Trigger maintenance on run hours, cycles or counts instead. Our guide to usage-based maintenance shows how.

6. Watch the condition of your critical assets

For machines where a failure stops the whole line, monitor temperature, vibration or current.

A rising trend gives you time to plan the repair. See condition-based maintenance.

7. Let operators own basic care

Cleaning, inspection, lubrication and tightening catch most early warning signs.

Train operators to do them daily with a simple checklist. This is the core of autonomous maintenance in TPM.

8. Solve repeat failures at the root

If the same fault comes back, the repair fixed the symptom.

Run a short 5 whys on every failure that repeats three times in a month.

Recover faster: ways 9 to 12

9. Keep critical spare parts on hand

Waiting for a part can turn a 20-minute repair into a two day stop.

Rank your assets by criticality and stock the parts for the top ones. See spare parts inventory management.

10. Put the machine history one scan away

Technicians lose time looking for manuals, drawings and what was done last time.

A QR code on the machine can open its full record on a phone. Our guide to QR codes for maintenance explains the setup.

11. Write troubleshooting steps for the top faults

For your ten most frequent faults, document the checks in the order that finds the cause fastest.

New technicians then repair at close to the speed of experienced ones.

12. Agree on response times

Define who responds to a stop, and how fast.

Track the reaction time separately from the repair time. They usually have different causes.

Plan better: ways 13 to 15

13. Shorten changeovers with SMED

Changeovers are planned downtime, and they are often the largest single block of lost time.

Move every step you can outside the stop, and standardize the rest. See our SMED guide.

14. Stop waiting for material, people and approvals

Kit materials before the shift starts and use a clear shift handover checklist.

Make sure the person who can release a first-off part is available when the line restarts.

15. Schedule maintenance together with production

Maintenance that waits for "a free moment" usually waits until the machine breaks.

Agree fixed maintenance windows with production and keep the maintenance backlog under control.

Worked example: cutting 26 minutes of downtime per shift

A bottle filling line runs two shifts a day. Each shift has 450 minutes of planned production time.

In this plant, stops longer than two minutes count as downtime. Shorter stops are tracked as speed losses.

After 20 days of automatic data capture, the average downtime per shift looked like this:

Availability was (450 - 72) ÷ 450 = 84.0%.

The three fixes

• Changeover: a SMED workshop moved parts preparation outside the stop. 24 minutes became 14.
• Capper jams: a worn guide rail was replaced and a weekly chuck check was added. 16 minutes became 6.
• Waiting for material: materials were kitted before each shift. 9 minutes became 3.

Labeler faults ranked third by minutes, but kitting materials was cheaper and faster, so the team fixed waiting first. The labeler went into the next cycle.

The result

• Downtime per shift fell from 72 to 46 minutes.
• Availability rose from 84.0% to 89.8%.
• With performance at 90% and quality at 98%, OEE rose from 74.1% to 79.2%.
• Over 480 shifts a year (240 working days), 26 minutes per shift is 208 extra running hours. That is more than five weeks of single-shift production.

Notice what did not happen: nobody bought a new machine.

The biggest item on the list, the changeover, was planned downtime that nobody had questioned.

A 90 day downtime reduction plan

1. Days 1 to 30, measure: connect the bottleneck line, agree on stop reasons and a two minute threshold, and collect a clean baseline.
2. Days 31 to 60, fix the top three: run a Pareto, assign an owner to each cause and track the minutes every week.
3. Days 61 to 90, make it stick: turn the fixes into preventive maintenance tasks, checklists and standard work, then extend to the next line.

Mistakes that keep downtime high

• Fixing the loudest problem instead of the biggest one. The breakdown everyone remembers is often not the largest loss.
• Too many reason codes. Operators pick "Other" when the list is long, and the data becomes useless.
• Ignoring planned downtime. Changeovers and cleaning are losses you can reduce, not facts of life.
• Measuring maintenance only by repair speed. A fast repair on a machine that fails every day is still a problem.
• Changing everything at once. Change one thing, measure, then change the next, or you will not know what worked.

How Fabrico helps

Fabrico is an OEE platform with a full CMMS built in, so downtime data and maintenance work live in one place.

It records stops from PLCs, IoT sensors and AI cameras, detects micro stops, and shows downtime, MTBF, MTTR and real-time OEE by machine.

From the same platform, your team can open a work order, run recurring preventive maintenance plans, track reaction time on emergency work and manage spare parts stock.

Technicians scan a QR code to open the machine's history on iOS or Android, and the AI assistant answers questions about a machine's stops in plain language.

Want to see where your own downtime goes? Book a demo.

To connect downtime to output, read our guides to production efficiency and overall equipment effectiveness.