This article illustrates how to determine the number of names to test mail in order to infer whether the rollout response rate will likely exceed some desired level. Also shown is how to predict the likely response rate of a full panel after a test mailing. The article is aimed at direct marketers with little statistical background.
Two example calculations are provided in detail: one for determining sample size and one for determining confidence intervals. Throughout the text, emphasis is placed on presenting, in a readable fashion, the statistical rationales that underlie all formulae. The sensitivity of the sample size formula is also explored.
Get full access to this article
View all access options for this article.
References
1.
BlattbergRobert C.Decision Rules and Sample Size Selection for Direct Mail TestingDirect Mail Marketing Manual1984Direct Marketing Association, IncNew YorkRelease 610.1, issued 1979
2.
ChouYa-IunStatistical Analysis1975Holt, Rinehart and WinstonNew York
3.
CourtheouxRichard J.Financial Optimization of Direct Marketing Test QuantitiesBurnsKaren L.Mailing Lists: A Practical Guide. Monograph81984Direct Marketing Association, Inc.New York215–236
4.
GaethGary J.ThranMichael K.Improving Low Response Rate Satistics in Direct Mail Market ResearchJounral of Direct Marketing Research1198629–40Number 1, Summer/Fall
5.
HarperRoseMailing List Strategies: A Guide to Direct Mail Success1986McGraw-Hill Book CompanyNew York
6.
HodgsonRichard S.Direct Mail and Mail Order Handbook3rd ed.1980The Dartnell CorporationChicago
7.
HrenRichardBeyond Significance: Determining Sample Size for Direct Marketing Tests using Power AnalysisJournal of Direct Marketing Research119865–28Number 1, Summer/Fall
8.
NeterJohnWassermanWilliamWhitmoreG.A.Fundamental Statistics for Business and Economics1976Allyn and Bacon, IncBoston
9.
PassavantPierre A.Direct Marketing Economics and BudgetingDirect Mail Marketing Manual1984Direct Marketing Association, IncNew YorkRelease 600.1, issued 1979
