Clean and accurate list data is a critical component of any successful mailing. But list hygiene is even more crucial with ethnic lists, since they have unique data capturing requirements. For one, “Hispanic names are usually longer [than typical Anglo names], and computers tend to truncate them,” explains Sybil Heine, senior account executive of data management and merge/purge house Marketing Communication Systems, based in Ivyland, PA. “The typical merge/purge program [doesn’t accommodate longer names], making deduping very inaccurate.”
When dealing with files of Hispanics, Asians, and African-Americans, systems need to recognize ethnic names such as Choi or Fuentes, and account for the possible use of multiple surnames. In the Hispanic and Asian cultures, for instance, the last surname is usually the married name, while the other two are the father’s and mother’s, says Cindy Lin, assistant vice president of New York list firm Database Management.
But mailers, list firms, and service bureaus can take steps to help keep files of ethnic names accurate. For example, experts suggest that catalogers extend the number of characters in a field from the typical 15 to 20 to accommodate longer surnames; add a field for the language in which customers and prospects prefer to be marketed to; and always ask a married woman which surname is her married name.
General merchant Fingerhut mails six Spanish-language catalogs a year, and has built an infrastructure specifically to accommodate Hispanic names. “We keep track of who would like to do business with us in Spanish” via an additional computer field to indicate a customer’s language preference, says Sigrid Page, manager of Hispanic marketing. The Minnetonka, MN-based mailer has more than 1 million Hispanic names on its 31-million name list.
On the supplier side, San Francisco-based ethnic list compiler Segment Data Management uses an algorithm that analyzes the first and last name in combination to determine ethnic origin. For instance, “the surname Kim is almost 90% Korean, but Chen can be Korean or Chinese. If the first name is Jung, that’s Korean; if the first name is Yepao, that’s Chinese,” says Elcid Choi, director of marketing.
And Database Management uses a scoring model that assigns a weight factor against the likelihood that a record may be a duplicate, comparing first names, last names, and addresses for similarities.