MTTR and MTBF answer different questions, how long a failure costs and how often one comes, and it is only together that they describe an asset. A machine with a long MTBF and a long MTTR fails rarely and hurts badly; one with a short MTBF and a short MTTR is a nuisance rather than a crisis. This page is about reading mtbf and mttr as a pair: what the combination says about availability, what it says about the cost of a breakdown, and how a maintenance manager uses both to set the interval on a preventive task.
Availability from the two numbers
Availability is the share of time an asset is ready to run, and from these two figures it is MTBF divided by MTBF plus MTTR. An asset with an MTBF of 85 hours and an MTTR of 3.4 hours is available a little over 96% of the time, which sounds fine until it is the one asset the whole line waits on. The formula is worth knowing mostly because it shows which lever to pull: on that asset, halving MTTR moves availability less than doubling MTBF does.
The cost of a breakdown from MTTR, the frequency from MTBF
The worked example on this site prices a breakdown at 3.4 hours of a technician against 1.1 hours for the same work done as a planned job, at a $48 loaded rate, which is $110.40 a breakdown that planning would have avoided. MTTR is where that 3.4 came from. MTBF says how many such breakdowns a month to expect; 84 across the example site's 180 assets. Multiply and the reactive share of the month has a price, which is what the preventive programme is bought against.
Setting a preventive interval from the pair
If an asset's MTBF is 85 operating hours and the preventive task against it runs every 200, the task is not preventing the failures it was written for and the interval is too long, or the task is the wrong task. If MTBF is 900 hours and the interval is 100, the task is being done nine times for every failure it might catch and the interval is too short. The MTTR decides how much that matters: a short MTTR on a cheap asset may be worth accepting rather than preventing.
Why the two need the same records
Both figures come from the same reactive work orders: MTBF from their count against operating hours, MTTR from their times. A history that does not separate reactive from planned work, or that does not record when the asset was reported down and when it was handed back, cannot produce either number honestly, and a team that argues about the figures is usually arguing about the records.
Questions people ask about mttr and mtbf
Which matters more, MTTR or MTBF?
For availability, MTBF usually has the larger effect. For cost, it depends on what the asset does when it is down. Read both against the asset's role rather than picking one.
Can I calculate MTBF and MTTR in a spreadsheet?
Yes, from a work order list with the asset, the type, the reported time and the returned time. The maintenance log worksheet on this site is that list; a CMMS produces the figures without the spreadsheet.
How often should I look at them?
Monthly for the assets you are actively working on, quarterly for the rest. More often than that and single events dominate the averages.