Acxiom, IRI Tackle Data Accuracy

Data quality is important to any business. But when data is your business, as it is for database marketers, it has to be squeaky clean.

Information Resources Inc. has 170 terabytes of data, the majority of it point-of-sale information collected from 40,000 grocery, drug, and convenience stores--600 million records every week--on what packaged goods and health care products consumers are buying. That's supplemented with data provided by 100,000 households and information about the promotional activities of consumer packaged goods manufacturers and retailers. The data is sold to consumer packaged goods manufacturers and retailers, who use it to calculate market share, plan marketing campaigns, and carry out other analysis.

IRI corrects and standardizes data as it comes in from retailers in 1,850 formats, CIO Marshall Gibbs says. The company uses IBM's DataStage for quality control and management and Ilog's JRules rules engine to apply the necessary business rules to data. IRI uses its own algorithms for detecting any anomalies.

The key to IRI's success is using Six Sigma methods to analyze data collection and management processes before implementing data quality automation software. "Automating a process works much better after you've done the process reengineering to improve it," Gibbs says, adding that the amount of bad data that gets through to customers is negligible.

Acxiom, which augments data provided by its business customers with terabytes of consumer data that it collects and maintains, monitors the quality of data coming from clients using a scorecard system and metrics developed internally in the late 1990s, says Traci Campbell, a data quality consultant at Acxiom. Incoming data is scrutinized for accuracy and completeness, proper categorization, and redundancies before any corrective action is taken.

Problems still crop up, such as when a client prepares consumer data for a project, then tries to use that data for another project for which it's missing key information or isn't properly formatted. And it's a challenge to keep consumer data fresh, reflecting up-to-date address changes, marriages, divorces, and other life-changing events, says Bob Gillen, address quality services product manager. "It's not something you fix once."

Illustration by Jay Montgomery

Return to the story:
Hamstrung By Defective Data Continue to the sidebars:
Avaya Group Safeguards Internal Information Quality
and Customers' Data Quality Needs Make Vendors Acquisitive