Guide to managing a data quality assurance program
A comprehensive collection of articles, videos and more, hand-picked by our editors
A few years ago, the finance department of a major telecommunications company turned over its fixed asset record ledger -- an official accounting of property such as cell towers -- and a $5 billion budget sheet to independent auditors, who found discrepancies in the company's data that pointed to possible problems in its data quality process.
The project: Chris Levitt, a member of the Technology Business Management Council, a nonprofit organization that makes recommendations for best IT practices and benchmarks, was an IT risk manager for a consultancy at the time. He was brought in to lead a team of technicians and evaluators to find the gaps between the telecom's actual assets and what was documented in the ledger.
The evaluators sifted through the ledger data, comparing what was recorded with a master list of the company's property and identifying areas that didn't match up. In some cases, the amount the company paid for an asset and its depreciated value were missing, sending the team members back to the original contract data. When they couldn't find that information, they had to use secondary sources to come up with an estimated value.
Based on the findings, the technicians built algorithms to help clean up the ledger, which involved asset transfers, disposals, updates and revaluations across millions of rows of data.
More on data quality management processes
Get advice on building a business case for data quality investments
Find out how Tagged.com fixed data quality issues that were hampering its analytics efforts
Read about the importance of setting internal standards for data quality
The challenge: When a business buys an asset, the process of entering data to document the purchase starts in one department and works its way through the organization until it eventually comes to rest in the fixed asset record. But for the telecommunications company, the system broke down, leading to data quality problems and the disparity between the ledger and the results of the independent audit.
"All of the processes are in areas where the expertise is distinctly different -- from IT purchasing to IT operations to the financial analyst to the procurement program manager," Levitt said. "All of those people have a role to play and are touch points as the data moves through the organization."
The lessons: Make sure employees can talk the data talk. To help bolster the telecom's data quality process, Levitt and his team established a common data language, with definitions for every asset type and instructions on how those assets should be identified, tracked and loaded into the system.
Find a point person. Levitt's team built models to help automate as much of the data entry and updating process as possible, but also tied duties to specific employees and provided training.
Fight for executive sponsorship. As tired as that piece of advice may be, Levitt encouraged securing executive support for data quality improvement efforts, which he said can provide employees an incentive. "There is no silver bullet [to data quality]," Levitt said. "It … crosses so many boundaries you need to go outside of IT to make it work."