printer friendly version

How to make data governance practical

Gary Allemann.

For many businesses, data governance is not well received. This is because governance programmes are far too often driven by supporting functions – like compliance or IT – with a focus on the needs of these supporting functions. The result for many stakeholders is that they feel like they are being given a lot of new tasks to fit into their week, without any real benefit.

Data governance is not about data governance

Any approach that implements data governance without clear use-cases that support multiple stakeholders will struggle. This does not mean that governance should not start small – it should. By selecting one use-case where an organisation can deliver value and where it has support, it increases its chances of success as long as:

1. Its governance programme directly includes stakeholders that benefit from the use-case selected.

2. It does not force work upon people for whom the use-case is not adding value.

3. It agrees to a roadmap and plan to include additional use-cases to support additional stakeholders and keep these additional stakeholders updated on progress through its communications plan.

Data governance is typically about formalising ad hoc information gathering and decision-making processes (about data) so that they become repeatable. If too many people are engaged too early, many will lose interest because they have nothing to contribute to the information gathering or decision making for the use-case in focus, or because they cannot see how this will be applied to support them.

The right level of communication engages different stakeholders at different levels of detail, managing expectations while expecting more effort from those that are getting the immediate benefit.

What are potential use-cases?

The right use-case will vary from one organisation to another and could, for example, be one of these:

1. Delivering better insights more quickly. Decision support remains one of the most visible applications of data. Yet, for many managers, analytics takes too long and when reports are delivered, the results are treated with suspicion. Data governance use-cases that can be explored to improve the delivery of insights could include:

a.Improving access to data by cataloguing and publishing datasets, identifying owners and approval processes for data access requests, and delivering fine-grained access control policies.

b.Improving trust in data by making data quality visible, adding context such as definitions for calculations used and confirming the data lineage.

2. Operational efficiency. Almost any business process generates data. Inaccurate or incomplete data at the first point of capture has a huge impact on downstream business processes that require the missing data, as discussed in the 1:10:100 rule. Fixing data captured incorrectly costs around 10 times what it would have cost to capture correctly the first time, while leaving bad data can cause errors that may cost a hundred times the original cost of capture.

Use-cases that can be explored to increase operational efficiency typically focus on data integrity; for example, agreeing on data capture standards and policies, defining data quality rules, applying data quality metrics, and managing data quality remediation issues. These use-cases are explored in a recent Precisely webinar: Fuelling Enterprise Data Governance with Data Quality.

Data governance also helps in improving IT efficiency. By ensuring that vital data specifications are available and accessible for impact analysis, data governance practices support agile development methodologies like DevOps and DataOps. Key data governance use-cases include the delivery of quality metadata, processes to ensure collaboration across cross-functional teams, and managing requirements specifications and approvals to speed quality delivery.

3. Managing data risk. Data risk can largely be separated into reporting risk – which comes down to the delivery of trusted data in regulated industries – and data privacy risks. PoPIA implementation requires a series of data management capabilities and processes to be implemented with governance oversights. These can offer a range of possible use-cases to kickstart data governance.

Data governance is not new. In most organisations, decisions are regularly made to support data projects. Executives need to take responsibility for data breaches, or to sign off financial reports as accurate. Data governance, as a discipline, should be about reducing the overheads associated with these tasks through collaboration and reuse.

Share this article:

Further reading: