6. Mistakes to avoid

Avoid trap samples. Trap samples include, for example, a manufacturing part number for which not all the data is available, either because some of the data was omitted, or because at the time of sampling we were unaware that one of the data items, considered unimportant at the time, was later identified as influential. In this case, it's best to sample again. It is also important to avoid "cherry-picking" the sample to be used as a reference, either by taking samples from rejects, or by tightening the selection criteria.

Avoid trivializing the use of standards. Standards should always be considered as "precious", either because of their intrinsic value, or because they represent a stake in the estimation of your production. You should therefore only use them when you really need them, and take care to preserve them and keep them in good condition.