It is useful, every now and then, to think about the variety of ways we try to segment markets. Consider the analytical groups of general market surveys: current users, past users, heavy users, light users, luxury brand users, economy brand users, brand loyal users, brand switchers, those who never used, etc. Every one of those analytical groups defines a market segment, but we're not often happy to leave things so coarsely segmented. No, we go on to check each analytical group on how it intersects with a lot of other segments. Not only do we check how one kind of analytical group intersects with another (e.g., heavy users by past users), but we usually think up some more segments to relate to the analytical groups. For instance, most large surveys relate demographic segmentation to the groups. Many use “psycho-graphic” measures like “venturesomeness” or “hypochondriasis.” Some use very sophisticated techniques of multi-dimensional scaling to derive complexly defined but highly discriminating market segments to be checked out against all the other ways of segmenting.