The amount of total observed variance of an output that is attributable to the testing methodology, rather than the product or process itself. Often divided into repeatability and reproducibility components.

Technical point: as a rule of thumb, a good testing variation is generally <32% of the total observed variance.