1. Knowing when to act on the process: the raison d'être of statistical process control (SPC)

A manufacturing process is implemented to obtain a product with a characteristic within defined tolerances. The process parameters are set to obtain the target value for the product characteristic. In practice, however, these process parameters can only be controlled with a degree of uncertainty: they vary randomly around the desired values. The combined effect of random variations in the many process parameters results in random variations in the product characteristic.

In general, the process applied will enable the vast majority of products to be produced within the required tolerances (cf. Statistical quality control: the notion of process centring and dispersion