What MTTR, MTBF and availability mean
In one line: MTBF (mean time between failures) tells you how often a machine fails, MTTR (mean time to repair) tells you how fast you restore it, and availability is the share of time the machine is ready to run. MTBF measures reliability and higher is better; MTTR measures maintainability and lower is better; availability combines the two into a single percentage.
These are the three numbers a maintenance head is asked about in every review, and they are routinely confused. The trap is to treat them as one "uptime" figure. They are not — MTBF and MTTR are two independent problems with two different fixes, and availability only tells the whole story when you can see both of the numbers underneath it. A machine can have excellent availability because it rarely fails, or because it fails often but is fixed in minutes; those are very different plants to run, and only the split reveals which one you have.
The rest of this guide takes each number in turn with its formula and a worked example, then combines them, clears up the MTBF-versus-MTTF confusion, puts a rupee figure on downtime, and shows how Fast Maintenance Software computes all of it automatically from your breakdown tickets.
MTBF explained — the formula and a worked example
MTBF = total uptime (operating hours) ÷ number of failures. It is the average length of the run a repairable asset gives you between one failure and the next, so it is a direct measure of reliability. A rising MTBF means the machine is failing less often; a falling MTBF is an early warning that something is degrading.
A press that runs 600 hours and fails 4 times
Over a review period a press clocks 600 operating hours and suffers 4 breakdowns. Its MTBF is 600 ÷ 4 = 150 hours — on average the press gives you 150 hours of running between failures. Note what MTBF ignores: it says nothing about how long each repair took. It is purely a count of how much good running you get per failure. To make it move, you have to make the machine fail less often — which is reliability and preventive work, not faster spanners.
One caution on the inputs: MTBF should be built from operating hours, not calendar hours. A machine that is scheduled off over the weekend was not accumulating exposure to failure while it sat idle, so counting that idle time inflates MTBF and flatters reliability. This is one reason a system that timestamps every breakdown against real running time gives a truer number than a hand tally ever will.
MTTR explained — the formula and a worked example
MTTR = total repair (downtime) time ÷ number of failures (repairs). It is the average time to restore the asset once it has stopped, so it measures maintainability — how quickly and cleanly your team gets a downed machine running again. Lower is better. Where MTBF is a reliability question, MTTR is a logistics-and-response question: how fast someone is dispatched, whether the diagnosis is right first time, and above all whether the spare is on the shelf.
The same press: 20 repair hours across 4 failures
Those same 4 breakdowns took 20 hours of repair time in total. MTTR is 20 ÷ 4 = 5 hours per repair. Whether MTTR is "good" depends entirely on context — 5 hours may be fine for a heavy assembly that needs a crane and an external technician, and terrible for a machine whose only real repair is swapping a stocked belt. What MTTR does not tell you is how often you are paying that cost; for that you need MTBF beside it. The two only make sense together.
A practical note: decide once what your "repair clock" includes. Strict MTTR counts only hands-on repair time; the more useful operational figure — sometimes called mean time to restore — runs from the moment the machine went down to the moment it was handed back to production, including waiting for a technician and waiting for a spare. That wider clock is where a missing part shows up, which is exactly the number you want to attack. Whichever you choose, be consistent, because a metric that changes definition month to month cannot show a trend.
Availability — and the combined worked example
Availability = MTBF ÷ (MTBF + MTTR). It is the fraction of the total cycle — running time plus repair time — during which the asset was available to run, expressed as a percentage. Because MTBF sits on top and both numbers sit underneath, availability rises either when the machine fails less often (MTBF up) or when each repair is quicker (MTTR down). That is the whole reason to keep the two component numbers visible: availability alone cannot tell you which lever moved.
600 hours, 4 failures, 20 repair hours — then halve MTTR
Bring the press figures together. MTBF = 600 ÷ 4 = 150 hours. MTTR = 20 ÷ 4 = 5 hours. Availability is therefore 150 ÷ (150 + 5) = 150 ÷ 155 ≈ 96.8%.
Now suppose you stock the two spares that caused most of the waiting, and the average repair time halves from 5 hours to 2.5 hours. MTBF has not changed at all — the press still fails exactly as often — but availability becomes 150 ÷ (150 + 2.5) = 150 ÷ 152.5 ≈ 98.4%. You bought roughly 1.6 percentage points of availability by touching only the repair clock, without a single reliability improvement. That is the clearest possible illustration of why MTTR and MTBF are two different problems: one worked-example machine, one lever pulled, two very different levers left to pull.
The mirror-image lesson matters too: if instead you had left MTTR at 5 hours and doubled MTBF to 300 hours through preventive work, availability would climb to 300 ÷ 305 ≈ 98.4% as well — the same destination by the other road. Which road is cheaper and faster for a given asset is precisely the decision the numbers exist to inform.
MTBF vs MTTF — the distinction that trips people up
MTBF is constantly confused with MTTF (mean time to failure), and the difference is not pedantic — it decides which metric is even valid for a given item. The dividing line is whether the item is repairable.
| Aspect | MTBF — Mean Time Between Failures | MTTF — Mean Time To Failure |
|---|---|---|
| Applies to | Repairable assets — machines, pumps, presses you fix and return to service | Non-repairable items — bearings, bulbs, filters you replace, not repair |
| Measures | Average running time between successive failures across the asset's life | Average life until the single failure that ends the item |
| Formula | Total uptime ÷ number of failures | Total operating life ÷ number of like units (average life to failure) |
| After failure | Repair, then the clock starts again on the next interval | Discard and replace — there is no "next interval" for that unit |
| Use it for | Both are reliability measures — pick MTBF for assets you repair and keep, MTTF for consumable parts you throw away and swap | |
So a CNC machine has an MTBF, because you repair it over and over; the specific bearing inside it has an MTTF, because when it fails you fit a new one rather than mending the old. Conflating them produces nonsense — you cannot have "time between failures" for a component that only ever fails once. In day-to-day maintenance of plant and equipment, MTBF is almost always the number you want, with MTTF reserved for reasoning about the life of the consumable spares feeding your reorder levels.
Still working these numbers out in a spreadsheet after month-end?
We can show you MTBF, MTTR and availability computed live per machine from breakdown tickets — plotted over time, no manual tally — in a 30-minute demo on your own assets.
A fuller worked example — with an INR cost lens
Numbers earn attention when they carry a rupee sign. Take three machines from one line over the same period, each with its operating hours, failures and total repair time. Compute MTBF, MTTR and availability as above, then translate the downtime into money using a simple downtime cost = downtime hours × ₹ per hour. The per-hour rate below is indicative for illustration — confirm your own rate, built from lost contribution, idle labour and any penalty for missed dispatch.
| Machine | Uptime (h) | Failures | Repair (h) | MTBF | MTTR | Availability | Downtime cost @ ₹8,000/h |
|---|---|---|---|---|---|---|---|
| Press A | 600 | 4 | 20 | 150 h | 5 h | 96.8% | 20 × 8,000 = ₹1,60,000 |
| CNC B | 600 | 2 | 16 | 300 h | 8 h | 97.4% | 16 × 8,000 = ₹1,28,000 |
| Compressor C | 720 | 6 | 12 | 120 h | 2 h | 98.4% | 12 × 8,000 = ₹96,000 |
Read across and the priorities invert what a single number would suggest. Compressor C has the best availability (98.4%) yet fails the most often — its MTBF is the lowest at 120 hours; it looks healthy only because each repair is quick at 2 hours. It is a reliability problem hiding behind a good MTTR. CNC B is the opposite: it rarely fails (MTBF 300 hours) but each stop is long at 8 hours — a maintainability problem, most likely a spare that is not stocked. Press A carries the largest downtime bill at ₹1,60,000 despite middling metrics, because it combines frequent-ish failures with slow-ish repairs.
The money column is what turns this from a scorecard into a plan. Attack CNC B's MTTR by stocking its long-lead spare and you cut a big chunk of ₹1,28,000; attack Compressor C's MTBF with preventive work and you stop paying ₹96,000 repeatedly. Same three formulas, but reading them together — reliability, maintainability and cost side by side — is what tells you where the next rupee of maintenance effort should go. (Figures illustrative; use your own downtime rate and period.)
How to improve each number — the two levers
Because MTBF and MTTR are independent, improving them is two separate programmes of work. Raising MTBF is reliability engineering; lowering MTTR is response and logistics. A good CMMS supports both, but through different features.
| Lever | Raise MTBF — fail less often | Lower MTTR — restore faster |
|---|---|---|
| The goal | Fewer failures per operating hour (reliability) | Less time lost per failure (maintainability) |
| What moves it | Preventive maintenance, condition checks, root-cause fixes, operator care, correct running conditions | Spare-part availability, faster dispatch, right-first-time diagnosis, clear work instructions and checklists |
| Biggest single win | A disciplined preventive calendar on the critical assets | The right spare on the shelf — stocked against reorder levels |
| Software capability | PM schedules, checklists, breakdown history and MTBF-over-time trends to target the repeat offenders | Spare BoM per asset, reorder alerts, and downtime capture that exposes where the repair clock is spent |
| Watch out for | Over-servicing low-criticality assets — spend PM effort where failure hurts most | Chasing speed at the cost of a proper fix, which just shortens MTBF instead |
The recurring insight is that a missing spare is usually the largest single item on the repair clock, which is why spare-parts inventory management and machine downtime tracking do more for availability than heroics on the shop floor. And because the two levers reach the same availability by different routes, the honest question for each asset is which lever is cheaper to pull — a decision you can only make when both numbers, and the cost behind them, are in front of you. For the strategy behind the MTBF lever specifically, see preventive vs breakdown maintenance.
How a CMMS computes these automatically
The reason these numbers are so often stale or wrong is that hand-tallying them from a breakdown diary is tedious and error-prone. A CMMS removes the tally entirely. In Fast Maintenance Software, every breakdown is a ticket that captures a time-down and a time-restored timestamp against a specific machine. From those timestamps and the failure count, the system derives MTBF, MTTR and availability itself — no spreadsheet, no month-end reconstruction.
Because the metrics are downstream of the tickets, their quality depends on clean inputs — an accurate asset register and honestly captured timestamps. Get those right and the reliability numbers maintain themselves, cloud or on-premise, with no separate reporting effort. For the wider set of numbers a maintenance review needs, see essential maintenance KPIs and reports.
What good looks like — setting honest targets
The most common question after the formulas is "what should my MTBF and MTTR actually be?" — and the honest answer is that there is no universal benchmark. A world-class MTBF for a hard-run press is a poor one for a lightly-loaded conveyor; a 2-hour MTTR is excellent where the fix is a stocked belt and unremarkable where it needs a crane and a vendor. Absolute targets copied from someone else's plant are worse than useless because they hide whether your assets are getting better or worse.
Set targets three ways instead. First, use trend, not absolute — the question that matters is whether this machine's MTBF is higher and its MTTR lower than last quarter, which the over-time charts answer directly. Second, set targets per asset class, so a critical A-class machine carries a tighter goal than a run-to-failure C-class one. Third, tie the target to cost — the point of lifting availability is to remove a downtime bill you can name in rupees, not to hit a round percentage for its own sake. A machine already at 98% availability with a small downtime cost rarely deserves the next reliability project; a 95% machine bleeding money does.
Track the split, watch the trend, and let the rupee cost of downtime rank the work. That is what turns MTTR, MTBF and availability from three numbers you quote in a review into three levers you actually pull.
Frequently asked questions
What is MTBF and how is it calculated?
MTBF (mean time between failures) is the average operating time a repairable asset runs between failures — a measure of reliability, so higher is better. You calculate it by dividing total uptime (operating hours) by the number of failures in that period. A machine that runs 600 hours and fails 4 times has an MTBF of 600 ÷ 4 = 150 hours between failures.
What is MTTR and how is it calculated?
MTTR (mean time to repair) is the average time it takes to restore an asset after a failure — a measure of maintainability, so lower is better. You calculate it by dividing total repair (downtime) time by the number of repairs. If four repairs took 20 hours in total, MTTR is 20 ÷ 4 = 5 hours per repair.
How do you calculate availability from MTBF and MTTR?
Availability is MTBF divided by the sum of MTBF and MTTR: Availability = MTBF ÷ (MTBF + MTTR). With an MTBF of 150 hours and an MTTR of 5 hours, availability is 150 ÷ 155 ≈ 96.8%. Halving MTTR to 2.5 hours lifts availability to 150 ÷ 152.5 ≈ 98.4%, because each failure now costs less time to fix.
What is the difference between MTBF and MTTF?
MTBF (mean time between failures) applies to repairable assets — the average time between failures across a machine's working life. MTTF (mean time to failure) applies to non-repairable items you replace rather than fix, such as a bearing or a bulb — the average life until its single failure. Use MTBF for machines you repair and MTTF for throwaway parts.
How do you improve MTTR and MTBF?
You improve them with two different levers. Raise MTBF — so the asset fails less often — through preventive maintenance and reliability work that removes the root causes of failure. Lower MTTR — so each repair is faster — through spare-part availability, clear work instructions and quicker response. Availability rises when you pull both levers together.