Clustering analyses of segmentation surely scores over apriori segmentation, as it does not use the tried and tested demographic indicators, and purchase %age to create a model around regression co-efficients. If there are more dependent variables, and high degrees of freedom, the predictive validity of the linear equation gets skewed. It's always better to keep homogenous groups for benefit segmentation, and keeping those distinct from heterogenous effects
I would tell this to the product manager
In my marketing career, I have learnt one thing- use templates, AOP guidelines, and per cap data to justify your cause in the board room. It does not help to use the statistical data or alpha co-efficients to re-launch a product that is heterogenous in so many ways. Grouping of homogenous data around distinct benefits creates a benefit segmentation that John Lynch explained in his class. His benefits around IBM computer data did indicate the technological challenges for the segments that do not like to time travel.
![Reblog this post [with Zemanta]](https://lh3.googleusercontent.com/blogger_img_proxy/AEn0k_u4E0WYosNxJTq8gHcd49zBh-dJXLavp4VebZ6RxVxR43JUbNb4wpJHUsYVX2st_VTZl6Cr71ZM_HnHCBOKyZNk9Fa7951qBURz9zlznp6Md8DAQM0oKA9Tv8C-Rq7QdXjxcOm4tSfxulJ6fTOXWzSm=s0-d)
