Academic journal article The Journal of Business Forecasting Methods & Systems

How to Predict the Success and Failure of a New Mailing List in Direct Mail

Academic journal article The Journal of Business Forecasting Methods & Systems

How to Predict the Success and Failure of a New Mailing List in Direct Mail

Article excerpt

Sooner or later, most direct-mail marketers find themselves in a bind for lists. They have scraped the barrel clean of lists that represent the same "type" as the ones that worked successfully before. As such, they have to go to other lists in search for more and new customers. But it is very costly to test every commercially-available list on a hit-or-miss basis, knowing very well that most of the lists they test will fail. Based on the author's experience, only 25% of the lists tested are used again either as a second test or as a part of rollout. The purpose of this article is to show with real data how a direct-mail marketer can forecast the success and failure of a new list. With that information, he or she can not only cut down the cost of list testing but also expand the customer base.

LITERATURE REVIEW

A number of articles have been written on the subject but their approach is mostly descriptive, that is, what to look for in selecting a mailing list. Where statistical approach is suggested, the emphasis is not on list selection but on the selection of certain clusters of names from a house list. Taybi and Frankel, Fernback, Kendall, and Goldberg point out what direct-mail marketers should look at in selecting a list. Taybi and Frankel suggest that direct-mail marketers should study four different elements in selecting a list which are: 1) Customer penetration analysis, 2) continuation usage, 3) categorical analysis, and 4) overlay comparison. Customer penetration analysis calls for matching the house list with outside lists. The higher the duplication between them, the higher will be the response. Continuation usage refers to how well a given list has been used by competitors as well as by mailers with similar markets. List is good if it has been used successfully by these mailers. Such information can be usually obtained from list brokers. Categorical analysis refers to how names were acquired (direct mail, space ads, etc.), what products they purchased, and what types of promotion and offer were made in acquiring them. The authors claim that one can develop an index on the basis of this information, but don't explain how. Overlay comparison refers to how an outside list which the mailer wishes to test compares with his or her own bad debt file. The higher the duplication with the mailer's debt file, the less chances of a list to succeed. Fernback, on the other hand, suggests that one should look into recency (how recently they ordered), frequency (how many times they bought) and monetary value (how much they have spent) in selecting a list. She also suggests that one should look into the demographic characteristics such as income, age, education, family size, length of residence, as well as the psychographic characteristics (what they do, how they live, what is their lifestyles, etc.). Kendall suggests that before selecting a list the mailer should determine such things as demographics and buying habits of his or her customers, and then test lists with those characteristics.

Goldberg suggests, in selecting a list, one should look very closely the media and the offer. Among the media, he regards the direct mail as the best. He further adds that a list made up with a sweepstakes offer will not work well for a non-sweepstakes mailing.

Hubbell talks in general terms how regression model can be used in selecting a mailing list but does not show how it works and what variables are to be used or tested. Nor does he share his experience, that is, how it worked if and when he used. He says, however, that U. S. Census Bureau can be used as a source for gathering data on income, age, household size, occupation and education. A number of studies - Thrasher, Hotchkiss, Bult and Wanbeek, and Ehrman and Miescke - refer to the use of regression in selecting a list but their emphasis is not on selecting an outside list but certain segments of a house list. Thrasher shows how CART (Classification and Regression Trees) can be used in segmenting profitable names from a house list. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.