Blog: David Loshin

This may seem like a big number but I am not surprised. The data quality that we find in customer source systems, including shiny new ERP implementations, continues to amaze ne. In one case at Shell the business thought it was selling 20,000 product combinations to 20,000 commercial customers, and was actually selling 5,000 products to 6,000 customers! It can be seen that the potential for cost saving in a situation like that was significant.

It is "nice" to see if anybody is interesting in study the financial impacts of poor quality data or information on business.

By simple dividing it is in average 20 000 EAR/company. It is not too much but I think it could be more because study covers only a part of the costs of quality.

To me are more important the last information about number of companies focuses on information quality measurement, etc.

This study is a very welcome document. Thanks for it.

Andy:

Your insight in the particulars of that specific instance is a good validation of the existence of insidious data quality issues that might never be exposed were it not for some analysis being done on the data. Knowing your expertise and interest in Master Data Management, I would suspect that your team has found numerous instances like this that could be used to contribute to a set of case studies. With the project at Shell, did you guys find a large number of data flaws that could be specifically tied to monetary impact?

Milan:

Thanks for your comments, and it is good to see you emerge, since I hadn't seen your name around in a while!

For a very good paper on data quality impact analysis, see this one:

http://tricare.osd.mil/rm/documents/fa/DoDGuidelinesOnDataQualityManagement.pdf

from the US Department of Defense. It is a little old (from the mid-90's), but the approach taken and described provides a good case study for assessment and impact analysis.