One of the LinkedIn users brought to our attention the fact that “according to top managers in the IT business the top 2 reasons why CRM investments fails is:
1. Managing resistance within the organisation
2. Bad data quality” (B. Hellgren, 2008) and I think that’s valid not only to CRM or BI solutions but also to other information system solutions.
Especially in ERP systems the data quality continues to be a problem and from my experience I can point here the following reasons:
- processes span across different functions, each of them maintaining the data they are interested in, being rare the cases in which a full agreement is met between functions, and here is also management’s fault.
- within an enterprise many systems arrive to be integrated, the quality of the data depending on integrations’ stability and scope; if stability can be obtained in time, scope resumes to the volume of data available, the delta needing to be maintained mainly manually.
- there are systems which are not integrated but using the same data, users needing to duplicate their effort so they often focus on their immediate needs.
- lack of knowledge about the system used, the processes in place and best practices (they can be further detailed).
- basic or inexistent validation for data entry in each important entry point (User Interface, integration interfaces, bulk upload), system permissiveness (allowing workarounds), stability and reliability (bugs/defects).
- inexistence of data quality control mechanisms or quality methodologies.
Data quality is usually ignored in BI projects, and this because few are the ones that go and search the causes, being easier to blame the BI solution or the technical team. This is ones of the reasons for which users are reticent in using a BI solution, to which add up solution’s flexibility and the degree up to which the solution satisfies users’ needs.
On the other side BI solutions are often abused, including also reports which have OLTP characteristics or of providing too much, unstructured or inadequate content that needs to be further reworked.
I think that data quality comes on manager’s dashboards especially during ERP implementations, when the destination system has some strong validation. And then there is no wonder when each record contains at minimum a defect and the whole dashboard is in nuances of red (the red must be "en vogue").