December 17, 2010, 5:46 PM — Scenario: A large financial services company is having fundamental problems with its data management and governance. Business leaders want to better leverage data for both operational and analytical needs, so there is widespread support for finally treating it as a true corporate asset. Consolidating data sets is now high on everyone's priority list. Both the CIO and CTO believe they need to adopt master data management (MDM) in some key areas but recognize that it's more critical to establish an overall data strategy. What's the most effective way for these IT leaders to move this forward?
Paul Bergamo
In helping many organizations reform their data management and governance practices, I've noticed how often companies fail because they are unclear on what outcomes they truly need. They also falter when IT tries to go it alone without business buy-in.
Business leadership must own and drive high-level data decisions, while IT should focus on implementing those decisions and providing technical advice and counsel.
One of my clients did a great job of handling this. The company had one very senior executive who declared ownership for a subject area and was able to make significant progress standardizing his data and its usage. He found that since he adopted this approach, both IT and business users would actually seek out his governance function to get things done.
As the CIO and CTO team up, they have to first gain business engagement (not just support) and then set expectations. They must communicate that good data strategy, data management and governance are ongoing efforts. They should then prioritize data capabilities and required outcomes, striking agreements with the business on key measures.
Establishing a good communication strategy is key to engaging the organization and helping people understand their roles. They'll also need to implement an oversight model that scales to their capabilities. The last step (and potentially toughest one, politically) will be holding people accountable for their data decisions and actions.
Paul Bergamo is a general partner at NewVantage Partners. Reach him at pbergamo@newvantage.com.
Jesus Arriaga
A few years ago, I worked as a CIO at a company facing a similar situation. Over time, we'd built systems and data sources that didn't always use the same definitions. For example, different systems of reference had different definitions of which products were included in a product category.
That made analysis challenging, since there were often different versions of the truth. Simply getting data to flow between systems was an issue, and it was evident that we needed to fix our data and develop a better way to manage it. An MDM plan looked to be the best approach.
Our approach to creating a solid MDM strategy involved 1) defining business processes, 2) understanding data integration points, 3) data cleansing and 4) ongoing governance.
Defining the business processes involved establishing a cross-functional team of key business and IT individuals who documented the processes that create and touch data. This team also established the data definitions that would be the standard throughout the company. An extensive (and time-consuming) data-cleansing effort uncovered inadequacies in certain applications that had to be fixed in order to maintain an MDM strategy. Finally, we developed a governance methodology with an IT steering committee that made the calls on any changes.
This effort created an IT environment that was easier to manage and more responsive to change as the company grew. Arguments over what analysis was accurate were no longer an issue. There was one version of the truth.
Jesus Arriaga is president of CIO Strategic Solutions. Reach him at Jesus.Arriaga@CIOStrategicSolutions.com.
Read more about data management in CIO's Data Management Drilldown.
