Attending the Data Governance Conference in San Diego June 1, 2009, through June 4, 2009, I was impressed and inspired but concerned at the same time. The success stories from a number of companies are impressive and inspiring. But like all things in this economic environment, realizing the goals and benefits of data governance is extremely difficult. I listened to multiple presentations, discussions and debates about data governance, master data management, data quality, data profiling and so on. In the end, I settled on a broad definition. Data governance pertains to principles, practices, and methods for managing and using data in accordance with the company’s governmental, fiduciary and social responsibilities.

Throughout the conference, there were many questions from participants about building and maintaining support for data governance in the current downturn.  That said, on the whole the conference was very hopeful. As I listened to a number of presentations from thought leaders, analysts and practitioners, I was struck by the passion and creativity they bring to the party.  Strange as it may seem, it also occurred to me that the approach to data governance today parallels toxic cleanup. It is messy, expensive, difficult, hard to justify and reactive. As this thought raced through my head, I began thinking about how we got into this data mess and what we coul do to get out of it. The stories being told paradoxically described progress at the same time sounding a little like playing the arcade game Whack-a-Mole.  Just when you think you have one thing solved, two more moles pop up!

Taking inspiration and lessons from the conference along with more than 20 years of data management experience, I know there are a few critical things we can do to transform our future.

Focus on Business Value, Never Say “Data Governance”!

First and foremost, don’t pitch “data governance” as such!  Too often, data governance is pitched as an initiative independent of a specific and direct business benefit.  Most people agree that data needs to be effectively managed and should be of the highest quality.  But without a specific improvement in the operation of the company, the agreement ends there. Remember that data governance is a means to an end and not the end itself.  The end is, and must remain, achieving business value. Simply put, data governance guides how to operate but not why.  Incorporating data governance practices into a business initiative is certainly challenging but it can be done effectively. The best advice I can give is to talk about the impact of data in achieving the business objective. As an example, a company wants to improve profitability. One of the initiatives is to optimize pricing of products and services based on a customer’s entire relationship with the company.  In almost every company in the world, there are significant problems with assembling a full view of the customer across all products, services and operating entities.  A 360° understanding of the customer cannot be achieved unless the necessary data is available, complete and accurate.  The trick in this situation is to explain how this data problem can be mitigated without waxing poetic about “data governance” and “master data management”! Remember, data governance may be a tool that helps to achieve improved profitability but is not worthwhile on its own.

As we look at the data challenges in the organization, a relentless focus on business impact forces our conversation toward mitigation of problems and that’s a good thing. A more subtle result of this focus is that the “business” can realize a real life manifestation of un-governed data practices that directly affect the ability to achieve objectives. Conventional wisdom teaches us that humans will not change until the pain of the status quo exceeds the pain of changing.  In other words, until business people experience and understand the pain of un-governed data in a very specific and explicit way, they will not be open to the kind of change needed for data governance practices to really take hold.

Change That Lasts!

Now that we have an example of driving change that embraces the tenets of data governance, we have the opportunity to guide a more permanent transformation of the business.  Again, we don’t want to “sell” data governance per se. By building examples and stories of other business objectives or initiatives where un-governed data impedes the path to success, we can begin a conversation about integrating “data governance practices” into standard operating procedure. Specifically this means adding rules, standards of practice and actions to existing policies, procedures and processes.  Several of the success stories presented in the conference used this approach. Changing standard operating procedure is just one of the critical steps toward making “data governance” practices a way of life.

Somebody’s Got To Do The Dirty Work!

Bringing data under control is hard, politically challenging and sometimes dirty, but it still needs to be done. Think about it. Data governance exists because there is a data problem that has not been resolved as a natural evolution of business practices and computer systems. In many respects, the data problem is like the disaster caused by dumping toxic substances into our environment. Over the years of ignoring the adage “garbage in, garbage out,” we have made a mess of our data. That problem has been exacerbated by exponential data growth of the past 10 years.  Now we’re paying the price. It needs to be cleaned up but it is hard to do, very expensive and painful. Unfortunately, there is no Super Fund for data cleanup. To obtain the funding needed to clean up our toxic data, we need to develop business cases that make sense and demonstrate a return on investment. There is enough data toxicity in most enterprises that there should be no lack of targets for a business case. It is essential to correlate the un-governed data impact with business value to effectively build a business case.  Beyond tactical projects to reverse data toxicity, we also must proactively change the way we operate in the future. Learning from our mistakes is a business case in itself!

Demand Manufacturer Change!

As I contemplated the data governance challenges described in the conference sessions, I had a radical thought. The toxic data environment is the result of many contributors. One source is the many years of building software that did not provide real data integration or interoperability between applications at the data level. In software packages today, processes can talk to each other and pass requests for action back and forth. One could say there is interoperability between application software processes. However, the data that results from these processes is stored many times in our systems and frequently with different values and meaning. This over-replication is further complicated by the heterogeneous ecosystem of software in nearly every enterprise on earth.  For too many years there has been an excessive emphasis on process integration without an equal focus on data integration.  Has the time come to demand that software vendors agree on a single standard for data interoperability? What if proprietary data structures for “master” data were eliminated or at least fully communicative? What if we did not have to invest billions to integrate and synchronize applications in addition to the toxic data clean up? We asked polluters to alter their systems to prevent future contamination of the environment. We need to ask software manufacturers to do their part to build software that coordinates, prevents and mitigates future data contamination while being interoperable. Can’t we all just get along?

Recent articles by Steve Palmer

• Business Intelligence Recipe for Success
• Business Intelligence Realization: An Agile Answer
• From Cave Paintings to CNN: The Quest for Business Intelligence