Ask most plants what an hour of downtime costs and you’ll get a shrug, or a number that only counts the machine’s hourly rate, which is usually the smallest piece of the real cost. Downtime doesn’t just stop a machine. It stops the labor standing next to it, delays whatever was downstream of it, and often gets paid for twice, once in lost time and again in overtime or expediting to make it back. The number worth knowing is bigger than the one most plants assume, and it’s simple enough to build with a pencil and the numbers already sitting in a job cost sheet.
The four pieces most plants only count one of
Lost throughput. The part not made. If the machine makes a part worth $40 in margin every 90 seconds, an hour of downtime is 40 parts not made, roughly $1,600 in lost margin, not lost revenue, margin, because material and labor for those parts weren’t spent either.
Labor paid to stand idle. An operator running that machine is usually still on the clock during a stop. If they’re paid $28 an hour loaded (wages plus benefits and overhead), that hour of downtime cost $28 whether or not a single part got made, and that’s before counting a second person, a tech, a supervisor, who might get pulled in to help.
Fixed overhead still running. Rent, utilities, insurance, equipment financing, none of it pauses because a machine did. Spreading a shop’s fixed monthly overhead across its scheduled machine hours gives a rough per-hour figure, often $15 to $40 an hour depending on the shop, that keeps accruing whether the machine is cutting or not.
Recovery cost. This is the piece almost nobody prices in, and it’s often the largest. A late job doesn’t just disappear, it gets made up, usually with overtime, an expedited shipment, or a rush order to a supplier, all of which cost more than the original plan would have. A job that slips because of two hours of unplanned downtime might cost an extra $400 in weekend overtime to hit the ship date anyway, on top of the original two hours already lost.
The math, in a form you can redo on paper
Nobody needs a spreadsheet to build a rough version of this. Four inputs, most of them already sitting in a job cost sheet or a payroll report.
- Margin per hour of production. Take the part’s margin and divide by its cycle time, or just take a shift’s total margin and divide by scheduled hours if that’s easier to pull.
- Loaded labor rate for whoever’s standing at the machine during a stop, wages plus your standard benefits and overhead load, usually 1.25 to 1.4 times base wage.
- Fixed overhead per scheduled machine hour, monthly overhead divided by total scheduled machine hours across the shop that month.
- A recovery estimate, this one is the roughest, but even a conservative guess, half the affected job’s margin, or your standard overtime premium on the hours needed to catch up, is more accurate than leaving it at zero.
Add the four and you have a real hourly downtime cost specific to that machine, not a generic industry number. Multiply it by how many hours a month that machine actually sits down, and the total is usually the number that gets a plant’s attention in a way the individual stops never did.
Why a small daily number becomes a large monthly one
The instinct is to shrug off a 20 minute stop. Twenty minutes doesn’t feel expensive. But run the same machine through a normal month and the stops add up in a way that daily thinking hides.
A single machine losing 20 minutes a day, every working day, at roughly $2,053 an hour, works out to about $684 a day, $14,400 a month, and $172,000 a year. That’s one machine, one recurring stop pattern, no dramatic breakdown involved, just a quiet daily leak nobody bothered to price. Multiply that across a shop floor with ten similar machines and the number stops being a rounding error and starts being a line item worth a meeting.
The caveats
This model overstates cost in a genuine bottleneck-free shop where a stopped machine’s work can shift somewhere else with slack capacity, lost throughput on that specific machine doesn’t necessarily mean lost throughput for the plant. It understates cost on a true bottleneck machine, the one everything downstream depends on, where an hour of downtime there can idle several other machines and people waiting on its output, which this simple model doesn’t capture at all. Bottleneck identification covers how to tell which kind of machine you’re actually pricing before trusting the number too far.
The recovery cost line is a real cost, not a certainty. Some late jobs ship late with no premium paid, at the cost of a customer relationship instead of a dollar figure. Both are real costs. Only one shows up on a spreadsheet, which is why it’s worth naming the other one out loud even as an estimate, instead of letting it hide as an intangible nobody has to justify.
Why the number is worth building even roughly
A rough number beats a precise number nobody trusts, and it absolutely beats no number at all. The value of building this model isn’t landing on an exact dollar figure defensible to an accountant, it’s converting “that machine goes down a lot” from a vague floor complaint into a specific number that competes for attention against every other line item a plant manager is weighing. A maintenance fix that costs $3,000 looks very different against “it might help” than against “the stop pattern it addresses is costing about $172,000 a year.” The second framing is what actually gets budget approved.
Once a plant has real stop data by asset, this same math is what turns a downtime Pareto from an engineering curiosity into a spending decision, ranking not by which reason happens most often but by which reason costs the most, which the downtime reason codes guide covers from the tracking side.
Quick recap
- Downtime cost has four parts: lost throughput, idle labor, fixed overhead, and recovery cost, most plants only count the first
- Build the model with four inputs: margin per hour, loaded labor rate, overhead per machine hour, and a recovery estimate
- A modest daily stop pattern compounds fast, 20 minutes a day can mean six figures a year on a single machine
- The model overstates cost where slack capacity absorbs a stop, and understates it on a true bottleneck machine
- A rough number that gets used beats a precise one nobody trusts
- Pricing downtime in dollars is what turns a maintenance fix from “might help” into a decision with a payback period
Downtime cost FAQ
How much does an hour of downtime cost? It depends on the machine and the shop, but a useful estimate adds four pieces: lost margin on unmade parts, idle labor still being paid, fixed overhead that keeps accruing, and the recovery cost of making the job up later, often with overtime or expediting. Many shops land well over $1,000 an hour once all four are counted.
What is usually left out of a downtime cost estimate? Recovery cost, the overtime, expediting, or rush shipping needed to make up a late job, is the piece most plants leave out, and it’s often the largest single line item once it’s priced in.
Why does a small daily downtime number matter? Because it compounds. A stop that costs a few hundred dollars a day, every working day, adds up to tens or hundreds of thousands of dollars a year on a single machine, a total that gets budget attention in a way the daily number never does.