The purpose of reliability modelling is to express the specified requirements, functions, and operating and maintenance conditions for a system in such a way that the reliabilities of the items comprising the system can be assessed and combined to predict the system reliability, indicate shortcomings and assess logistic implications.

To be effective, models must represent, as closely as possible, the various features of a system and the conditions in which it is expected to operate. The most useful models, however, are those which strike a good balance between an accurate representation of a real-life situation and the need to provide results in a reasonable time, with due regard to the quality and quantity of input data and the required accuracy of the results. It is often better to make approximations based on reasoned assumptions than to attempt to use sophisticated mathematical techniques that are inconsistent with the quality of input data. Always remember that, although prediction is a quantitative process, its primary objective is to identify weak design areas for improvement. Thus, the emphasis is often on the comparison of values rather than the absolute values themselves.

It should be remembered that, in addition to the above, a reliability model is really a model of the system failure definition. In other words, there can be as many reliability models for a particular system as there are system failure definitions. For example, if a two-speaker stereo system was considered to have failed when stereo sound could not be heard from both speakers, then the reliability model would consist of blocks representing the speakers in a series configuration. However, if system failure was defined as a total loss of sound, then, for the same system, the speakers would appear in a parallel configuration. Thus, a key starting point in reliability modelling is to construct a set of system failure definitions. This is intuitive because to state that a system had a mean time to failure one year would be meaningless unless what is meant by system failure was clearly defined.

No one would presumably doubt that a transport vehicle had failed if the engine failed to start, or, despite all efforts, the vehicle could not be moved. It would be hoped that the mean time to failure for such a definition of failure would be acceptably large. However, if system failure was defined as the inability to travel faster than 70 mph, then the corresponding mean time to failure would likely be much shorter.