Fitness for purpose is the principle universally accepted among analyt
ical scientists as the correct approach to obtaining data of appropria
te quality. Yet few analytical scientists or end-users of data are in
a position to specify exactly what quality of data is required for a s
pecific task. A definition of fitness for purpose based on minimal exp
ected loss is proposed in this paper. This idea enables one to develop
optimal strategies for apportioning resources between sampling and an
alysis, and for balancing technical costs with end-user losses due to
error.