Business intelligence starts with clean, well-managed data.

 

No garbage in, no garbage out.

Technology consultant Phil Simon posted an excellent article on the MIKE 2.0 blog yesterday, asking whether a company should first invest in tools for data quality or business intelligence.  He approaches the issue from the perspective of DataFlux President Tony Fisher’s data maturity model, concluding that the question can only be resolved by determining how far along the scale a company is.

A company at the “Undisciplined” end of the scale can’t count on the accuracy of its data, so there’s no point in investing in a business intelligence tool that consumes the data.  That company would be better off investing first in a data quality solution.

On the other hand, a company toward the “Disciplined” end of the scale would have greater confidence in its data.  This company would benefit from a BI system because the data the BI system consumes is more reliable.

However, in both cases, the key to sound business intelligence is clean, well-managed data.   As Simon says (his phrase, not mine) in conclusion:

DQ tools may not have the sizzle of their BI counterparts. However, if they are purchased, implemented, and utilized effectively throughout the organization, they sure will make BI tools more potent when you get there.

In fact, most people would agree that the primary reason BI projects fail is that the data that feeds the BI system is error-prone, duplicative and inconsistent.  TEC consultant Anna Mallikarjunan summarizes the issue:

Even with a data warehouse that is well designed and equipped with the best tools for business intelligence (BI), users will encounter inefficiency and frustration if the quality of data is compromised. When embarking on a data warehousing or business intelligence project, it is essential for organizations to emphasize the quality of data that is used for analysis and subsequent decision making.

To address data quality, organizations need to approach the problem at the source.  That’s where master data management (MDM) comes in.   MDM includes both the technology and business processes for cleaning and managing master data from a variety of systems for use throughout the enterprise.

BI systems give companies insights into their operations, customers, sales, financials, and many other key performance areas.  But for reliable results, the data BI systems use must be accurate, consistent, and up-to-date.  That’s a tall order when the data in question is drawn from systems and applications across the enterprise.

Writing in Information Management, Michael Vølund, vice president of Platon USA, details how MDM addresses this challenge:

The remedy to [having various versions of the truth in multiple systems] is to start looking at how the focus on MDM can change the systems, processes and accountability in the data-producing part of the business to increase data quality and promote transparency and flexibility. For a BI practitioner, there is gold to be found in the MDM discipline.

That’s because MDM enables companies to (1) collect, cleanse, and organize their data, (2) keep constantly changing data clean , and (3) feed quality data into their business intelligence systems.  That last part is key; as noted in this blog and elsewhere, an MDM platform is only of value when its content can be easily (and even automatically) published or syndicated to consuming applications – including business intelligence, marketing and sales, dealer portals, customer support systems, and others.