There is an ill-defined grey area of expectation between properties of materials as published by suppliers (or as determined in a producer's laboratory) and what is achieved in final product performance. Philosophically, the act of taking a raw material and converting it into the test piece from which a property parameter can be determined is in itself ‘conversion to a product’ and is arguably sufficient to alter the original properties. The performance of a material as a test piece may not necessarily truly reflect its performance when in a larger mass or in a different shape.

An example may be drawn from the plastics field. Suppose a thermoplastic test bar is injection moulded in order that an impact measurement can be made. The material will fill the mould surely enough, but in doing so will have some degree of molecular orientation impressed upon it by the flow process. Consequently the test bar will have a preferred tensile strength direction along its long axis; this must modify the measured impact strength as compared with a totally isotropic material which has, for example, been poured into the mould from a liquid melt. An injection-moulded article with a range of section thicknesses, stiffening webs and surface areas in different planes will have a different level of molecular ordering to that in a test bar. Under impact conditions the laboratory test result could not accurately be extrapolated to reflect a finished article.