Organizations are now realizing the criticality of having proper processes in place for governance of data. However this realization has not been very encouraging in India. As I talk to different CIOs, I realize that hardly 5-6% of companies are thinking about following data governance practices.
A few decades back, governance of data was not a big issue, as the data in an organization seldom crossed over from megabytes to gigabytes. Also, the systems were handled by a relatively few experts who knew the business rules well. Today we are talking about managing terabytes and petabytes of data, which is definitely not easy. So concerns surrounding governance for data increase manifold. Hence it’s high time that data governance is taken seriously.
The foremost requirement for successful governance of data is business and IT collaboration. The IT team can create rules from a technical angle like database configuration and naming conventions, but for most part they need help from business.
Before starting your data governance project, it is important to draw strategies. Though there are no fix all guidelines, you can definitely look at tools available in the market. Here are some tips that can prove handy for proper governance of data.
1. Channelize the data: Today data is flowing from different channels. For example, a retail company can have multiple channels for selling its products - like Point Of Sale, web portal or third party services like amazon.com. How can a company find out whether a specific customer is the same one who is buying from different sources? Tracking this channel is a critical part of data governance.
2. Data lifecycle: Data is born in the organization, lives its life, and dies over a period of time. This process has to be mapped. Governing the entire lifecycle of data helps in meeting company's regulatory compliance and also updating (or purging) data that has lived its life.
3. Data quality: Ensuring data quality is critical for proper governance of data. During the process of data entry, same data may be defined in different manners. For example, New York can be written as NY or N.Y. In such a scenario, searching specific information on New York becomes difficult.
Strategic business decision making depends on the quality of data. Data has to be cleansed, standardized and changed on regular intervals that help in establishing a still version of truth. There has to be a single version of truth for information on customers, products and services.
4. Validate the current data sources: Giving validity scores to data can make things easy. Trust score should be give to data coming from different sources. This will help in governing the data coming from multiple sources. http://searchdatamanagement.techtarget.co.uk/resources/UK-data-quality-management-and-governance-strategy
About the author: Sanjay Raj is the practice director of business intelligence and data warehousing for Syntel Inc.
(As told to Yuga Chaudhari)