Some people may think that developing a segmentation model is plain and simple. This paper shows that there are a lot of choices and options, and not all of them are obvious. What should be the main focus of the model, what sort of data collection procedure should be used, what sort of samples should be drawn, what sort of statistical tools should be applied when analysing the data, and how should the model be presented to the organisation? These were only some of the questions we confronted during our trip down the development lane. In the end we came up with a model consisting of the seven segments. Since we segmented the whole market (i.e. the Norwegian population between 16-60 years of age), the model gives both a good description of segments with already high penetrations, and indications of which segments to concentrate on next - for future penetration. We also give an example of how data from the model has been used in product development. A simulation was run in December 1995 and empirical data from April 1996 reveals the models superiority in predicting the outcome.