The present study has certain limitations, and caution is therefore advised with
respect to the generalisability of the results. First, the simplicity of some of the measurements (due to the secondary nature of the data) raises some issues with regard to the operationalisation of specific constructs. Studies that model attitudinal as well as behavioural relationship outcomes have strong precedence in the relationship marketing area. It should be mentioned that even if various authors speak about attitudinal and behavioural measurements of loyalty, the majority of measurements are attitudinal. In this study, the “intentions” approach is followed rather than a behavioural one, as in order to develop true behavioural measurements, real behaviour measurement of clients was necessary. However, the measurement of the real behaviour of industrial clients proves to be very difficult from a practical point-of-view. Even if it is possible to measure “actual” behaviour, researchers are always confronted with difficulties related to the definition of the loyal client, raising questions concerned with the link between volume of repeated purchases or time and loyalty attributes