When Segments Misbehave: Segmenting Real-World Marketing Data

In an ideal world all segmentation methods work pretty well because we'd find a small number of evenly-sized spherical segments that we'd measure without noise.  In the real world, however, any or all of these ideal conditions may be violated: our segments might be elliptical instead of spherical, numerous rather than few and unevenly sized. Moreover our questionnaire might include a mix or signal variables that differentiate segments and noise variables that do not.  Faced with this array of adverse data and measurement conditions which methods can we rely on to find the segments in our data?  Which segmentation methods work best when we have non-ideal data conditions?  In this webinar Keith Chrzan from Sawtooth will share best practices for overcoming each of these challenges.