One of the key tasks faced by the life insurance industry is to identify market segments who are potential customers of Ule services offered by the fums. The fmt step in this p w s is to identify and target those individuals in the population who are more likely to "respond" to mailers from the insurance fm Such identification in Ihe Mushy has typically emphasized broad bawd segmentation approaches. In this papa, we ddress this subject by suggesting several things. Fmt, we recom- mend that in-house demographic data must be supplemented by werlay- ing it with fmancial data available fium certain secondary sources. Se- cond, this comprehensive database should be used to perform exploratoly data analysis to get a pmfde of potential responders and nompmders, and to prepmcess variables to make them more appropriate for subse- auent statistical modeling. 'lhird we mommend the use of a Logit d e l to (a) identify the- independent variables which are more likely than others to a fect the urobabilih, of res~~~nse aJad (b) accuralely Met the probability with &ch different in&viduals w&ld respod It is believed that Ihe benefits of basing a segmentation and targeting approach on such robust statistical approaches will far ouhKeigh whatever maease in costs such techniques may be associated with.