Quality control in the recording and matching of personal names

Abstract

This paper addresses means of reducing the confusion and errors that can arise as a result of differences in the way that an individual's personal name may have been recorded. This may result in not being able to trace a customer, not being able to identify that customer on external files, duplication of mailings and incomplete information on customer records. Name and address matching is a process that has been used since the early days of the direct marketing industry to address these problems. However, with the evolution of rapid addressing systems, the process of deduplicating records on a customer file has increasingly focused on the reconciliation of different versions of the same address to the comparative neglect of strategies for reconciling different versions of the same name. By contrast, it is difficult to find any literature that describes strategies or tools for assuring the quality of the name element — and in particular the personal name element — of a name and address record. The paper outlines the principal sources of inconsistency in the personal name field and discusses strategies for verifying the validity of entries found in this field. It outlines the ways of distinguishing valid variants of a common name from values resulting from data entry errors, whether mistakes in phonetic interpretation, in transcription or in transliteration. Though the findings of the paper are relevant globally, the paper focuses principally on the United Kingdom.