How to calculate the cost of downtime

The cost of downtime is everything it costs you when an installation is not running: the contribution margin you lose, extra labour, the repair itself and knock-on costs such as expedited parts, scrap and penalties. You calculate with margin rather than revenue, and only for output that is genuinely lost. Once you know that figure per hour and per incident, you can justify maintenance, spare parts and investments instead of prioritising on gut feeling. With the calculator in this article you can work it out for your own installation.

The formula for the cost of downtime

The cost of downtime consists of four items. For each incident, you add them together:

Cost per incident = lost contribution margin + extra labour costs + repair costs + knock-on costs

The most important choice is in the first item. Contribution margin is the selling price minus the variable costs: raw materials, packaging, energy and any other cost you do not incur when you are not producing. If you use revenue instead, you also count the materials you did not use during the stoppage and therefore did not pay for. The figure comes out too high, and any controller will spot it immediately.

The second principle: only count what would not have been spent without the failure. You pay your permanent operators whether or not the line is running. Those fixed costs were what the lost contribution margin was supposed to cover, so adding them again is double counting. Overtime, agency workers you pay while they stand idle and a callout of the on-call team do count as extra.

Then split the amount into costs per hour and costs per incident. Lost contribution margin and idle agency staff grow with every hour the installation stands still. Parts, an expedited delivery and start-up losses are paid once per incident, however long it lasts. With that split you can cost any failure: cost per hour × duration + fixed costs per incident.

In practice, you work it out per installation like this:

1. 1Establish whether the installation is the bottleneck, and how much buffer and catch-up capacity there is.
2. 2Convert the lost output per hour into contribution margin per hour, using figures from finance and production.
3. 3Add the extra labour costs per hour.
4. 4Determine the fixed costs per incident: repair, rush charges, scrap and start-up.
5. 5Calculate per incident and per year, and record the assumptions.

When downtime really costs money: bottlenecks, buffers and catching up

Not every hour of downtime costs output. That is the most important insight in this calculation, and the point where most estimates come out too high.

• Bottleneck or not. A plant's output is set by the installation with the lowest capacity. An hour lost on that bottleneck is an hour of output for the whole business. An hour lost on a machine with spare capacity often costs only the repair.
• Buffers. Intermediate stock buys time. If a mixer fails while three hours of product are waiting for the filling line, a two-hour failure costs no output at all. The cost only starts once the buffer runs empty.
• Caught up or lost. If you can make up the lost output later, at a weekend or with an extra shift, your loss is not the contribution margin but the cost of catching up: overtime, shift premiums and energy. Count one or the other, never both. An installation that already runs at full load has no spare capacity to catch up. At a waste-to-energy plant running at full load, the bunker absorbs deliveries during a stoppage, but lost incineration hours do not come back.
• Redundancy. If a standby pump sits alongside, failure of the duty pump costs only the repair, plus a period without a standby.
• Partial downtime. If a line runs at half speed, you count half of the lost output.
• Timing. The same failure costs more in peak season than in a quiet week.

Then there is the difference between planned and unplanned. Planned downtime goes where it costs least: in a scheduled shutdown, a cleaning window or a quiet period, with parts and people ready. Unplanned downtime arrives at a moment you did not choose, with rush costs and often secondary damage. That difference is the economic case for condition-based maintenance: a deviation you spot early turns a breakdown into a planned job.

In business continuity management this exercise is called a business impact analysis. ISO 22301 asks you to establish how the impact of a disruption grows the longer it lasts, and when it becomes unacceptable. For a critical installation that is exactly the question.

The four cost items in practice

Lost contribution margin

Take the output the bottleneck would have produced in that time and multiply it by the contribution margin per unit. In manufacturing that means units or tonnes per hour. At a waste-to-energy plant it covers the power and heat that are not delivered, plus processing capacity: the waste still has to be burned, stored or sent elsewhere. How maintenance is organised at such a plant is covered in waste-to-energy plant maintenance. At a power station, the lost amount depends on the electricity price at that moment; with heat supply, it depends on what the supply contract says about failing to deliver. In a distribution centre it is less about output and more about orders that miss the carrier's cut-off.

Extra labour costs

Overtime and premiums to catch up or to finish the repair, agency workers you pay while they cannot work, and an out-of-hours callout of the on-call team. Permanent operators you pay anyway do not belong here.

Repair costs

Hours of your own technicians at the internal rate, outside specialists, parts and materials, and hire of equipment such as a crane or an access platform. Your own technicians are a fixed cost too, but their hours go to this failure instead of planned work. That is why you include them, just as they appear on the work order. Do not forget secondary damage: a seized bearing that also scores the shaft makes the repair more expensive.

Knock-on costs

Expedited delivery and transport, scrap of product that was in the line, start-up losses until the product is back within specification, extra cleaning, and penalties or discounts under contracts. In the process industry a short failure easily becomes a long stoppage: cooling down, cleaning, heating up again. At a waste-to-energy plant, firing up counts too. After a stoppage, waste may only be fed again once the combustion gases reach at least 850 °C, and the plant gets there on auxiliary burners that cost fuel and time (Article 50 of the EU Industrial Emissions Directive).

Worked example: a failure on a packing line

An example with fictitious, round numbers. Packing line 2 in a food factory is the plant's bottleneck, runs around the clock and demand exceeds what the line can make. A bearing in the main drive fails. Between the breakdown and the restart there are 5 hours, 2 of which are spent waiting for an expedited bearing. The two technicians each spend 4 hours on it: 3 hours on the repair during the stoppage and 1 hour on checks after the restart.

The waterfall chart below shows how the four items stack up, and how large the share of lost contribution margin is.

The average of €2,500 per hour gives a first impression, but for decisions the split is more useful. Lost contribution margin and the agency workers together cost €1,920 per hour. Repair and knock-on costs come to €2,900 per incident. Two variants show how much the assumptions matter:

• Had the bearing been in stock, the 2 hours of waiting and the expedited delivery would have disappeared: 2 × €1,920 + €400 = €4,240 less.
• Had line 2 not been the bottleneck, with enough buffer for 5 hours, the lost contribution margin would have been zero and the same failure would have cost €3,500.

From incident to year

For an annual figure you need the number of failures and their duration, which is what MTBF and MTTR give you. Suppose line 2 runs around 8,000 hours a year with an MTBF of 400 hours: you can then expect 20 failures a year. With an MTTR of 2.5 hours, that is 50 hours of unplanned downtime. At €1,920 per hour and an average of €1,000 in fixed costs per incident (most failures are smaller than the bearing), unplanned downtime on this line costs 50 × €1,920 + 20 × €1,000 = €116,000 a year.

What to do with the downtime figure

• Prioritise. In a criticality analysis, a cost per hour makes the production consequences of different installations comparable, and work on the bottleneck goes first.
• Justify spare parts. Set the €4,240 per incident from the example against the cost of keeping the bearing on the shelf. How to make that trade-off is covered in which spare parts to keep in stock.
• Weigh monitoring and redundancy. A vibration measurement that reveals bearing damage early, or a second pump next to the first: set the investment against the expected annual cost of downtime. The pump life cycle cost guide published by the US Department of Energy puts it this way: with a spare pump installed in parallel the initial cost is higher, but a failure then costs only the repair.
• Repair or replace. Downtime is a standard item in the life cycle cost of an installation. How it counts is explained in repair or replace.

Do not inflate the figure. A downtime cost that is too high makes every investment look profitable and so loses its value. Let finance supply the contribution margin and production the capacity and buffers, and present a range rather than a single precise number.

Recording downtime properly

Without records, every downtime figure remains an estimate. For each incident, record at least:

• start and end of the downtime: the moment production stopped and the moment it was back at speed, not when the work order was opened and closed;
• planned or unplanned, and full or partial, with the share of capacity that was lost;
• the asset and the cause, using fixed failure codes instead of free text;
• the costs incurred: hours, parts, agency staff and rush charges.

The same records feed availability in OEE. One source for production and maintenance avoids arguments about whose figure is right. How a fault notification records whether there is downtime is shown under notifications and work orders.

Common mistakes when calculating the cost of downtime

• Revenue instead of margin. Materials and energy you do not use, you do not pay for either.
• Counting every stoppage as lost output. Buffers, spare capacity, standby units and the option to catch up make a lot of downtime cheaper than it looks.
• Double counting. Fixed wages on top of lost contribution margin, or lost contribution margin and catch-up costs for the same output. If you use the downtime figure in a life cycle cost calculation, do not count the repair under maintenance as well.
• One figure forever. Margins, capacity and buffers change. Review the figure at least once a year and after any major change to the installation.
• An average where timing matters. A failure in peak season or when electricity prices are high costs more than the annual average.
• Only counting the big failures. Many short stoppages can add up to more than the one big breakdown everyone remembers.