Examining the segment retention problem for the “Group Satellite” case

The purpose of this work is to determine how well, criteria designed to help the selection of the adequate number of market segments, perform in recovering small niche segments, in mixture regressions of normal data, with experimental data. The simulation experiment compares several segment retention criteria, including information criteria and classification-based criteria. We also address the impact of distributional misspecification on segment retention criteria success rates. This study shows that Akaike’s Information criterion with penalty factors of 3 and 4, rather than the traditional value of 2, are the best segment retention criteria to use in recovering small niche segments. Although these criteria were designed for the specific context of mixture models, they are rarely applied in the marketing literature.