Published on .

Each spring and fall MRI releases audience numbers and demographic data on more than 160 magazines-and all hell breaks loose. Highly successful targeted magazines such as Business Week lose over one-fifth of their audience overnight, even though their circulation is as strong as ever. A rapidly growing magazine, Allure, is reported with a margin of error of 21%.

The reason for the under-representation is that MRI is not a magazine readership survey, but a general media and marketing survey. MRI will produce large sample sizes for general interest, mass circulation titles; for example, the sample size for Good Housekeeping, People and TV Guide average over 3,000, which produces a low margin of error, around 3%. The sample size for Allure, Mademoiselle and Self, though, averages only around 300, which produces margins of error ranging from 17.5% to over 20%.

Another problem is that the MRI survey is overloaded. Those lucky people who agree to be interviewed sit through a questionnaire that attempts to measure yellow pages, radio, cable, TV and newspapers-and lasts close to an hour. The fatigued respondents are then shown over 260 logo cards. If the respondents recognize the logo, they can state they "read or looked at" an issue, any issue, of that publication within the past year. This then qualifies the respondent as a recent reader.

This methodology has serious problems that penalize targeted magazines, which now represent the majority of magazines reported by MRI. In-home interviews do not work in the '90s because of problems in penetrating doorman buildings or security-conscious residential communities, or in capturing 18-to-24-year-old active young adults (who are rarely home). To compound the problem, MRI does not even measure college dorms.

A further problem is that when MRI can't capture enough people in New York, Boston or San Francisco, they compensate by oversampling in Dakota or wherever they can get respondents, and then make very heavy adjustments. Perhaps this is why MRI's total sample consists of only about 8% professional/managerial women when this fast-growing segment now accounts for more than 17% of the population.

No one questions MRI's integrity. The problem is their methodology when it comes to measuring targeted audiences. What is needed are larger sample sizes, a mail survey that can get into every home and a dramatically reduced questionnaire. No more than 150 magazine logos should be part of the survey.

This is precisely the approach Audits & Surveys is taking in both their primary and total audience study for magazines. Audits & Surveys may not be the perfect solution-no research study can be-but their study, or for that matter any other viable new study, at least deserves the opportunity to move forward-and the support of the advertising and publishing community. If we do not support new solutions, then we should stop complaining.

Stephen M. Blacker

VP-market research, Conde Nast

New York

Most Popular
In this article: