Blog: Claudia Imhoff

« 2005 Editor's Choice Awards from Intelligent Enterprise | Main | HP hires NCR's Top Dog! »

Data Quality in Business Intelligence

I had the opportunity to do a web cast for DataFlux this week entitled “Reaping the Dividends of Data Quality”. In it, I talked about the five building blocks for successful data quality management. Here is a brief overview of those blocks.

The five building blocks are:

Data Profiling – Gaining an understanding of the existing data relative to quality specifications. This is your starting point from which improvement (and ROI) is measured. From this block you should be able to answer these questions: How complete is the data and how accurate is it? Consider this your baseline measurement from which you base your data quality improvement.
Data Quality – Gaining an understanding of the causes of quality problems. This block relies heavily upon the usage of data quality technology. The results yield an analysis of the root causes of data quality problems and inconsistencies. Once these are known, you can then begin to “fix” the problems, choosing from one of four options.
Data Integration – Collapsing disparate versions of data into a single one. This block demonstrates the recognition that the same data exists in multiple locations and systems with variable content in each system. It is in this block that you standardize the multiple versions (e.g., customers, products, geographies, etc.) to single version of the truth.
Data Enrichment – Incorporating additional external data to gain further insight. Her you combine your integrated internal customer, product or other data with third party data to increase your understanding of your customers (e.g., their demographics, credit history, etc.), competitors, total industry sales, and so on.
Data Monitoring – The data management effort requires an investment that requires a justification. Therefore, specific, tangible improvement measurements are often necessary to show the worth of this investment. To demonstrate this requires appropriate tracking techniques. There are three categories of data monitoring techniques – data auditing, data trending, and data alerts and controls. Use these to determine if your efforts are indeed paying off.

We are fortunate today to have technological help for each of these blocks. I hope this overview gives you some idea of what I think are the important building blocks for a successful data quality management program. If you want to hear more, you can hear an archive of the entire hour talk.

Please let me know how your own data quality management program has done.

Yours in BI success,

Claudia

Share: