A methodological investigation of alternative analytical models for the purpose direct marketing prospect identification and evaluation
OA Version
Citation
Abstract
Direct response, and more specifically direct mail, is frequently positioned as the vehicle of the 90's for media targeting and individual level micro-marketing. The efficiency of direct mail however, faced with increased competition and mail box clutter, as well as escalating postage, printing and paper costs, is in serious jeopardy. The ability to better target individual prospects (current non-customers) who have both interest in a product and the propensity to purchase through the mail has seen little refinement beyond subjective or experientially based judgements. Perhaps the richest source of data on individual consumers has traditionally come from projectable sample surveys. Yet, surveys, by definition, provide information on only a small sample (often 1500 people or less) of prospects and, therefore, have not been utilized for purposes of list selection. On the other hand, increased availability of commercial databases, with both extensive population coverage and exponentially accumulating data, is seen as an under realized opportunity for targeting guidance. Coupling these databases with sample surveys presents a unique and potentially powerful solution to the dilemma of declining direct mail efficiency. A successful linkage of such attitudinal and behavioral intention information, from sample based survey research to readily available national databases, allows for the quantitative assessment and subsequent ranking of direct mail prospects. The obvious benefits of such ranking includes both increased efficiency and effectiveness of direct marketing efforts, with ultimate benefits to bottom line profitability. With applications for a variety of marketing efforts, testing environments include target group membership identification, direct response purchase propensity and alternate positioning affinities. The project outlines a systematic approach to the mathematical linkage of survey results to available demographic databases, and the subsequent development of procedures for individual level identification and evaluation of direct mail prospects for specific marketing objectives. Alternative analytic procedures for both model selection and strategies for overcoming the limits and problems associated with missing values are explored.