What is downtime in manufacturing?

Downtime is any period your equipment was scheduled to produce but didn't. It's the single biggest drain on most production lines, and the easiest loss to act on: every minute of it has a cause you can name and remove. The hard part isn't fixing downtime; it's measuring it honestly in the first place.

Planned vs unplanned downtime

Not all downtime is equal. The first cut every plant should make is planned vs unplanned, because they're managed in completely different ways.

1. Planned downtime

Time you scheduled the machine to be down: changeovers, planned maintenance, cleaning, meetings, breaks, no-demand idling. It's expected, but "planned" doesn't mean "free." A two-hour changeover you do twice a day is 25 hours a week you could attack with SMED.

2. Unplanned downtime

Time the machine stopped when it was supposed to be running: breakdowns, jams, starvation, material shortages, operators pulled away. This is the expensive, surprising kind, and the kind a good monitoring system catches automatically, the moment a line goes quiet.

How downtime hits OEE

Downtime is the headline driver of the availability factor in OEE. Availability is simply the share of planned production time the machine was actually running:

Every stop you fail to capture inflates run time and quietly overstates availability, which is why the most common cause of a "surprisingly good" OEE is simply missing the small stops. The brief jams and micro-stops that never make it onto a paper log don't vanish; they reappear as a mysterious gap between planned and actual output.

What counts as downtime, and what doesn't

Where you draw the line decides what your number means. Two honest conventions:

• Stops shorter than a threshold (say, 30 to 120 seconds) are usually booked as a performance loss (a minor stop / slow cycle), not an availability loss, because counting every two-second jam as "downtime" makes the data unusable. They still cost you; they're just measured in the right bucket.
• Unscheduled time (the line was never planned to run, with no shift or demand) isn't downtime against OEE at all. It belongs to the wider TEEP picture. See OEE vs TEEP vs NEE.

Pick a threshold, write it down, and apply it the same way on every line. The exact number matters far less than applying it consistently so lines are comparable.

The true cost of downtime

The stopwatch cost is only the start. A genuine downtime hour also carries lost throughput you may never recover, scrap from ramp-up and ramp-down, overtime to catch up, and the knock-on starvation of every downstream station. A rough but honest figure:

On a bottleneck machine, that number is the whole plant's lost output: a stopped bottleneck is lost forever, because nothing downstream can make it back up. Put a figure on yours with the downtime cost calculator.

How to reduce it

You can't fix what you can't see, so the order is always the same: measure first, then attack the vital few causes. Capture every stop with a reason automatically, rank the reasons by total lost time with a Pareto, and work top-down. The full playbook is in how to reduce unplanned downtime, and the causes themselves map cleanly onto the Six Big Losses.