Data Governance Part 1 Design of the Process

Data warehouse projects present a specific set of challenges for an organization.  While most IT projects exist within a particular part of an organization, and are managed within that part of the organization, a data warehouse is there to answer business questions across different parts of the business, and will therefore have many more interfaces, involve more people and will be affected by changes in many systems. 

These challenges must be addressed and orchestrated through data governance.  The real challenge is how to set up an effective governance process without encumbering the organization with more levels of bureaucracy that inhibits rather than facilitates good business intelligence.  I expect we should first start by defining data governance.  We can then start exploring the issues.

WhatIs.com has one definition, data governance (DG) refers to the overall management of the availability, usability, integrity, and security of the data employed in an enterprise. A sound data governance program includes a governing body or council, a defined set of procedures, and a plan to execute those procedures. 

This definition seems to include all the elements.  However I think the definition is fairly idealistic and requires much more discipline than most organizations possess, and I would like to address data governance in a more pragmatic way.

Often the temptation is to read the book on governance or the definition and set up the process according to the "text book".  These approaches are doomed to failure because they do not address the specific business issue or more likely do not fit within the culture.  I do not deny that data governance must cover all the areas listed above but these need to evolve rather than be worked all at once.  There's a fine line between attaining pragmatic results and boiling the ocean.

A better approach is to design a governance process that mirrors some other cross functional management processes in the organization that are in place and are working.  These processes can take many different forms and require creativity in designing them to be simple yet effectiveBy choosing existing processes as models we avoid the difficulty of getting buy-in from the organization to follow the process. 

The design of the process should not try to address every single data governance issue in the early stages.  The initial implementation should address the pressing issues of the initial projects.  For example if the initial projects only involved a limited set of business users, others do not need to be involved.  Of course the better approach would be resolving all the issues at once but it is unlikely we will have the influence or leverage to get compliance from the whole organization.

The business leader, who is most affected by these projects and has the most to gain, should be asked to sponsor the process.  The leader will need lots of support and assistance but must take the lead in working the cross functional issues.  The leadership must promote teamwork and cooperation.   Finding models in the organization that already work will load the process for success.   

 A process that works might come from another cross functional project.   Looking at what works in the organization can really help create a successful process.  Another approach might be to integrate the process with an existing process.  An assessment of the readiness of the organization to deal with cross-functional issues is a critical step.

One must be careful not follow the lead of a vendor or consultant.  A business leader must own the process, not a vendor or a consultant.  By having the business own the process (or identifying a process that the business already owns) we can address the issue of cultural fit versus change that cannot be owned or addressed by a third party.

In summary, we must be pragmatic in designing a successful data governance process that fits within the culture of the organization and addresses the most pressing issues.  Careful design can be the difference between success and failure.

I will continue the discussion of data governance in the future blogs.  I invite comments or questions. 

Leave a Reply

captcha *