The main goal in the spare parts industry is to minimize the downtime of your valuable assets such as: aircrafts, vessels, oil rigs, wind turbines, generators, manufacturing equipment and heavy machinery. Whenever an asset needs repair or maintenance, the piece of machinery undergoes a full stop which is a highly costly scenario.
As spare parts are characterized by a highly erratic, infrequent demand with a high degree of randomness, the classic time-series approach of forecasting falls short. The average consumption over a given period will produce nonsensical results. Most of the time, you have no demand for a part, but on random occasions there is an urgent demand for a particular part that is critical to have. Another complexity is introduced when a breakage of an asset needs replacement of multiple parts. This means that if you have any of those parts missing, the repair cannot be done. Thus, the quality of service is not conditioned by having one particular part, but rather whether you can fulfill the whole BOM that is needed to complete a repair. How do you cope with this degree of inconsistency?
In spare parts, the client is not a human, but the valuable asset itself. It is crucial to have an understanding of the probability of every part being needed in the next lead time window before calculating the financial impact of keeping a part in stock by comparing to the consequence of downtime if there is a stock out if that part.