During auditions and certifications more and more questions are derived from requests by the ISO 9000 series, they concern the application of statistical methods in the realms of the experimental design. Taking the example of a bellows for homokinetic joints it will be shown that in the area of compounding in the rubber-converting industry one should start with the experience of the chemists, this helps to confine the number of necessary experiments caused by many influencing parameters. The required fine tuning in the case of large volume compounds should then occur with the aid of tools of statistical experimental design.