By now, most of us are familiar with data quality “best practices.” Involve the business user. Correct the source. Establish data governance.
It sounds great—but it often falls flat in the real world. Why?
It’s too difficult, states Lyndsay Wise, president and founder of the independent research and analysis BI firm, WiseAnalytics.
“Many operational systems were developed years or even decades ago without processes for correcting inaccurate and inconsistent data entries,” Wise wrote in a recent TechTarget article. “A majority of companies have yet to implement master data management programs and systems that use master reference data to help identify and fix quality issues as data is entered into systems.”
As a result, many companies find addressing data quality best practices translates into high costs and a lot of work by in-house or consultant developers.
A more practical approach is to address data quality at the BI and data warehouse layer, Wise contends, rather than trying to solve data quality in siloed operational systems.
She also says many business users seem to think that BI systems magically correct data quality problems. Not true, she points out. She suggests one of two approaches for cleaning up your data:
“Managing data quality activities in one place should result in lower costs compared with modifying data in individual source systems,” Wise adds.
It’s not a best practice. Data quality expert and Obsessive-Data Quality blogger Jim Harris would probably call this “reactive data quality,” which you may need to do to repair immediate damage.
Harris has compared this to everything from tooth decay to dog whispering, but the theme is the same: When there’s no data governance or data quality at the source, you never address the root of the problem. And eventually, that causes more problems.