Why One Model May Not Suffice
Friday, March 5th, 2010In today’s tough economic times, marketers are continually searching for ways to gain efficiencies and improve return on investments. Dramatic improvements in predictive models are possible through identifying variables that may not only work differently for different segments, but might only be relevant to certain segments, and not to others.
A while back, we wrote an article that dealt with the possibility, actually the probability, that different segments of your customer file may behave differently with regard to their relationships with well known key predictor variables such as Recency, Frequency, Monetary Value and various measures of Product Purchase data.
(more…)