Most of the equipment your business relies on will need to be replaced at some point. Calculating how likely failures are and when they'll occur is a crucial part of forward planning. You'll need those projections to plan for future capital expenditures, to schedule maintenance or preventive upgrades, and -- often -- simply to purchase equipment that's appropriate to your needs. One statistical tool that's used to assess a product's reliability is its mean time to failure, or MTTF.
Time isn't always the determining factor in an MTTF calculation. Instead, it's a measure of use that's appropriate to the product. Actual hours in operation is suitable for a computer chip or one of the hard drives in a server, while for firearms it might be shots fired and for tires, it's mileage. To arrive at an MTTF, you'd simply test a set number of units for a predetermined time. Multiply the number of units by the time -- or whichever measure you're testing to assess reliability -- to arrive at a number of unit-hours. Divide the unit-hours by the number of failures, and that's your MTTF. If you tested 100 units for 100 hours and saw two failures, you'd have an MTTF of 5,000 hours.
Use and Limitations
That doesn't mean you'd automatically get 5,000 hours' use -- roughly six months -- out of those components. For one thing, the numbers might vary if you tested a second batch of components. More importantly, the MTTF is a figure that might be skewed sharply by factors such as a high failure rate within the first several hours of operation. The MTTF is a useful quick calculation, but more powerful and flexible statistical tools such as the Weibull failure curve provide a better guide to a product's reliability.