Why this question never gets a straight answer
Ask a plant what one hour of downtime costs and you usually get one of three answers: no idea, a number pulled out of thin air, or how much electricity was saved. The first two cannot support a decision. The third has the sign backwards.
The real reason is that downtime cost is not one number — it is five numbers added together, and every plant has its own basis for each. Industry, line continuity, bottleneck operations and order-book saturation differ enormously. Any generic figure is a reference point at best and has to be recalculated against your own site.
Break it into five and fill them in one at a time
1. Production loss — use contribution margin, not revenue
The most common and most damaging mistake is working straight from hourly output value. The correct basis is contribution margin: for the output you did not produce during the stop, what you actually lost is the gross profit you would have earned.
Hourly production loss ≈ hourly output x unit contribution margin
With a full order book and the line at capacity, unit gross profit is a reasonable approximationWith spare capacity, discount the loss, because the volume can be made up laterWith buffer inventory in between, factor in how long the buffer holds — if it covers the stop, do not count it yet2. Labour and overtime
Repair crew hours, including night and public holiday multipliersOvertime to catch up production afterwardsIdle labour on other operations starved of material — the item most often forgotten3. Scrap and rework
Stopping and restarting are not symmetrical: the cool-down, heat-up and re-stabilisation phases are where scrap is generated. For continuous processes — drying, heat treatment, reactors, kilns — this line can exceed the production loss.
4. Parts and subcontractors
The parts themselvesExpediting premiums and express freightEmergency service fees from contractorsLoss of equipment life, from the damage a sudden failure does to the machine, which usually surfaces as later repair cost5. Delivery and reputation
Penalties for late delivery, or the difference in cost when you switch to air freightExtra cost from emergency sourcing and order reshuffling at the customerHardest to quantify and frequently the largest. Put in a conservative figure rather than a zeroA formula you can apply directly
Cost per stop ≈ downtime hours x (hourly production loss + hourly labour and overtime) + scrap and rework + parts and subcontractors + delivery risk cost
Run two versions: a conservative one covering the first three items and a full one covering all five. Use the conservative version to justify the decision and the full version when reporting internally.
Three indicative scenarios (indicative estimates, not commitments — verify against your own site)
Worked example: a small metalworking plant, one critical operation with hourly output value of ¥8,000, gross margin of 18%, three people on the repair and four hours of downtime (indicative estimate, not a commitment):
| Item | Indicative basis | 4-hour subtotal |
|---|
| Production loss | ¥8,000/hour x 18% = ¥1,440/hour | about ¥5,760 |
| Labour and overtime | 3 people x 4 hours x ¥60 | about ¥720 |
| Scrap and rework | Generated during restart | about ¥800 |
| Parts and expediting | Bearings, seals, express freight | about ¥1,500 |
| Delivery risk | Conservative estimate | about ¥1,000 |
| Total | about ¥9,780 |
That works out at roughly ¥2,400 of downtime cost per hour for this machine. Note that the machine itself is not expensive — what is expensive is the operation it serves.
Three more directional sketches (structure only, not industry statistics; real differences are large):
Compressor house main unit trips: the whole plant loses air, there is essentially no buffer, and the hourly figure is the plant-wide hourly contribution margin — usually the largest of the threeA single circulating pump stops: with a standby that can take over, the loss is mostly the risk and labour during changeover; without one, on a continuous process, it approaches line-level lossesA critical fan stops (drying, kiln, extraction): the loss concentrates in scrap and re-heating, and it is non-linear in downtime — scrap from a four-hour stop can be several times that from a one-hour stopNow turn it around: is monitoring worth the money
Once the per-event figure is settled, the investment case follows:
Annual downtime loss = annual unplanned stop count x average cost per stopThe avoidable part: the value of predictive maintenance is not eliminating failures, it is turning emergency repairs into planned ones — scheduled into changeovers, weekends or low-demand shifts, which is where the production loss line collapsesAnnual benefit ≈ avoided stops x cost per stop x the discount factor from planning + savings on parts and contractors + reduced patrol labourInvestment = sensors and gateway hardware + installation and commissioning + annual platform feePayback period = investment / annual benefitOne caution: results vary enormously between plants and between machines, so an ROI derived from an industry average carries essentially no weight. Fill it in with your own stop count, your own cost per stop and your own order-book saturation.
Four places this calculation usually goes wrong
Using revenue instead of gross margin: it inflates the benefit several times over, and when the number fails to materialise the project loses credibilityIgnoring the separate value of a few days' warning: the same failure costs very different amounts as an emergency repair versus a planned one, mostly in production loss and expediting feesLeaving false alarm cost out of it: false alarms send people to investigate and sometimes stop the machine, which is itself a cost. Ask about alarm precision rate when evaluating a monitoring systemCounting the machine but not the knock-on effect: one hour down on a bottleneck operation can idle three operations downstreamCalculate it online, and what to do next
Online calculation: fill in your own figures against the five items above and get hourly downtime cost, annual downtime loss and payback period directly (/roi-calculator/)Not sure which machine to start with: run a diagnostic assessment first and we will tell you which units are worth instrumenting first (/diagnostic/)Free trial on 3 units: get real trend data from your own equipment before committing to anythingScan the QR code to add us on WeChat, send over your equipment list and downtime history, and we will fill this table in together
Downtime cost is not a number for frightening people. It is a number for ranking. Work out the hourly downtime cost of each machine and you will know exactly where the monitoring should go first.
Calculate your downtime cost online
Enter hourly output value, gross margin, stop frequency and repair effort. Get hourly downtime cost and payback period in about three minutes.
