So far as organizational strategy goes, the one that has been in focus of late is data governance. Ever since data analytics as a technology and technique was added as a major weapon to an Enterprise’s competitive armory, governance of the data, too, has gained importance.
Yet, many Enterprises are found to be faltering on this front. There are surveys in the past that have shown that about 80 per cent of organizations that have implemented data governance have either failed or are struggling with it.
Lack of understanding, communication gap, failure to align the program with an enterprise’s overall strategy are some of the common reasons for such failures. Before delving at length on the reasons, let’s first understand what is it that data governance really entails.
Often, organizations confuse data governance with data quality. The latter is the “How” of the data analytics process while governance is the “Who and What”. Data quality is the maintenance of quality standards at an individual project-level, while governance refers to the overall data management in an enterprise – right from the availability stage to data usability, integrity, and its security. Data quality must be a part of the overall data governance system.
According to Stewart Bond, Director of Data Integration and Integrity Software Research at IDC, data governance is not a technology but a process supported by technology. Organizations that believe there is a magic box IT solution will fail, but organizations using an approach that includes business resources working with IT to make tough data decisions in a structure called data governance will be successful, he adds.
Clearly, data governance is policy, not a solution. It nests within any organization that has deployed business analytics as part of its overall strategy – in fact, one of the reasons for data governance failure is it not being aligned with an enterprise’s business strategy. Governance is about ensuring the right implementation of business rules and controls around your organization’s data, and involves the whole-hearted participation of all the departments of a company, especially IT and business. Any attempt to run it in a vacuum or silo means its imminently doomed.
In an article by the Society For Information Management, Michael Goul, an association dean for research and professor of information systems at Arizona State University’s W. P. Carey School of Business, makes an interesting point. While agreeing that data science retains the potential of revolutionizing commerce, he opines that a lot of companies are rushing headlong into the field without putting proper governance systems in place, which, in some cases, could lead to disaster.
Obviously, any organization that has started dealing in data to derive insights needs a data governance program in place. So where does it all start to go wrong, then?
What leads to the failure of a data governance program?
The nature of data is such that it is a shared asset, consumed by many, and so needs ownership. Security, too, is of importance here.
A well-thought out data governance plan must have a governing body, and a defined set of procedures with a plan to execute them. To begin with, one has to identify the custodians of an enterprise’s data assets. Accountability is key here. The policy must identify who in the system is responsible for various aspects of the data – from quality to accessibility to consistency.
Then come the processes. A set of standards and procedures must be defined and developed in relation to how the data is to be stored, backed up, and protected. Not to be left out, a good data governance plan must also have an audit process in place to ensure compliance with government regulations.
Even if an enterprise falters in one of the above areas, it’s data governance plan is bound to fail. But even if an organization gets its right and the governance program takes off, it does not end there. Data governance is an on-going project and not a one-off project. Chances are high that it may falter down the line.
Here are some of the reasons why:
- Lack of metrics
- Absence of documentation
- Loss of impetus
As any good analyst will tell you, metrics is the key. If an Enterprise does not know where it’s headed with its data governance plan, reflected in black and white, it’s bound to stutter. Things like targets achieved, dollars saved and risks mitigated need to be measured and recorded.
Failure to record is one of the highest contributory factors for the loss of momentum of any data governance program, say experts. Continuous tracking of the achievements and failures, and updating the data governance council of these will allow a data governance to be a living, breathing creature within an organization. The first sign of success must not be interpreted as time to slow down but instead, used to add to the momentum.