Process discipline is the recommended cure for inferior data

Data quality gets a bad rap, and deservedly so, judging by recent commentary from this analyst "trifecta":

A study by Cambridge, Mass.-based Forrester Research found that nearly a third of respondent companies reported they had no systematic approach to data quality. A May report by Stamford, Conn.-based Gartnersays more than 25 percent of critical data in Fortune 1000 companies is compromised due to data entry errors, incompatible data formats, inconsistent data definitions, and problems in compiling data from disparate business units. And another May report, this one from Boston-based AMR Research, made the distinction between transaction data problems and process measurement data.

AMR's report cites sloppy records management as the primary culprit for poor transaction data, resulting in potentially costly consequences, including:

• Improperly filled customer orders;
• Incorrect invoicing that delays payments from customers;
• Incomplete inventory records compensated for by excess inventory;
• Poor master data resulting in poor product quality;
• Weak financial recording processes.

"Most enterprises don't fathom the magnitude of the impact [resulting from] data quality problems," says Ted Friedman, principal analyst for Gartner. "[Data problems] cause wasted labor and lost productivity that directly affect profitability."

Remedies AMR recommends include improved process disciplines and Six Sigma techniques, with the caveat that the realm of process measurement data is "murky." Measurements such as daily yields—and product attributes such as thickness or weights—"follow a probability distribution" that can skew inventory accuracies and trend reporting.