This article proposes the use of the beta-binomial model to describe consumer response to mail solicitation. Besides generalizing the conventional view that consumers either are or are not responders, this model provides an intuitively appealing explanation for the well-known phenomenon of list falloff. It is shown that list falloff depends only on the dispersion of response probabillities in the population and not on the average response rate. These results are applied to the question of how many times to mail a given list. A discussion of the approach and its limitations follow.
Get full access to this article
View all access options for this article.
References
1.
BuchananB.S.MorrisonD.G.Measuring Simple Preferences: An Approach to Blind, Forced Choice Product TestingMarketing Science4198593–109Spring
2.
Buchanan, B. S. (1988), “Sampling Properties of a Compound Discrete Extinction Process,” Unpublished Working Paper.
3.
GivonM.M.Determination of Optimal Sample Size in the Beta-Binomial Brand Choice ModelJournal of Marketing Research17198058–72February
4.
GreeneJ.Consumer Behavior Models for Non-Statisticians1982PraegerNew York
5.
KalwaniM.U.Maximum Likelihood Estimation of Zero Order Choice Models Given Variable Numbers of Purchases per HouseholdJournal of Marketing Research171980547–551November
6.
KalwaniM.U.MorrisonD.G.A Parsimonious Description of the Hendry ModelManagement Science231977467–477January
7.
LindgrenB.W.Statistical Theory3rd Edition1976Macmillan Publishing CompanyNew York
8.
MasseyW.F.MontgomeryD.B.MorrisonD.G.Stochastic Models of Buyer Behavior1970M.I.T. PressCambridge
9.
MorrisonD.G.BrockwayG.A Modified Beta-Binomial Model with Applications to Multiple Choice and Taste TestsPsychometrika441979427–442December
10.
SabavalaD.J.MorrisonD.G.A Model of TV Show LoyaltyJournal of Advertising Research17197735–43December
11.
SabavalaD.J.MorrisonD.G.A Non-Stationary Model of Binary Choice Applied to Media ExposureManagement Science271981637–657July
12.
WierengaB.Paired Comparison Product Testing When Individual Preferences Are StochasticApplied Stastistics231974284–296September
