Opinion: How to kick-off a data governance initiative

Janani Dumbleton, a consultant at Experian Data Quality, examines the first principles of setting up a data governance initiative

More and more organisations are starting to understand how critical quality data is in achieving strategic objectives, not to mention ensuring customer satisfaction. That means data that is accurate and complete, and consistent with legal requirements and business rules.

However, while there is a growing consensus about the importance of quality data, there remains a lack of understanding about how to achieve it

Indeed, organisations often focus on short-term fixes or projects to bring their data ‘up to scratch', but then fail to follow this up by implementing appropriate governance processes to maintain that new found quality. It is then only a matter of time before another data quality drive is needed.

Data governance - the process of planning, monitoring and enforcing the management of data assets - is a key component of long-term data quality (and, therefore, long-term organisational success). It helps to ensure that data is captured accurately and that this accuracy is maintained, no matter how long it is stored for, to avoid the expense of repeated data quality initiatives.

So why are so many organisations still shying away from a proper data governance structure when it makes such obvious business sense in an ever-more data-driven world?

It all comes down to responsibility: without a single person (or department) in charge of a proper governance programme, it will invariably flounder.

While some organisations are starting to create data quality roles and departments, there are few with a similar set up for data governance. And, without this structure to report and feed back into, those responsible for data quality will frequently be cast in the role of fire fighter.

Just what is a data governance framework?

While there is no one size fits all approach, there are certain elements of a data governance framework that can be applied across the board.

1) Policy

A clear statement from a suitably senior and relevant C-level leader that your organisation requires proper data governance is a pre-requisite for building up the support for any initiative on a long-term basis (rather than it occasionally being ‘flavour of the month');

2) Processes

It is essential to have clearly defined and documented processes (and associated deliverables) in place, setting out how things such as data quality reporting and data quality issue management should be handled;

3) Responsibility

Responsibility is key. Defining who is responsible for data governance and quality is the cornerstone of any data improvement.

In short, data governance is about analysing, improving, and controlling data, and establishing processes to investigate and act on data quality issues.

How do you implement a successful data governance programme?

1) Analyse

The first step is to take stock and look at what you've got. This can be broken down into two main elements:

a) Data profiling

The process of gathering and examining information about existing data is often viewed as a pure data-quality activity. But when shared with those responsible for the data this can give advanced business expertise and insight to the results - bringing wider benefits to the organisation as a whole;

b) Reviewing and approving data definitions

To truly understand and manage your data it must be defined (for example, in a data dictionary or glossary), and then held where it is readily accessible by the users.

2) Improve

Once you know what you are working with you can set about improving it. This means:

a) Collaborating

A lack of collaboration between IT and the rest of the organisation will prevent data governance initiatives from getting off the ground all together. Take time to ensure that everyone understands the wider objectives, agree how to achieve them, and, crucially - who is responsible for what.

b) Reviewing and approving business rules for data cleansing

After undertaking your analysis you will need to garner input from your stakeholders to agree the rules by which the data will be cleansed (of course, with clear responsibilities assigned, deciding who to involve should be fairly simple). It's also useful to include these data cleansing rules in your data glossary for ease of future reference.

c) Master Data Management

Focus on how your master data management. That means making sure that the processes, governance, policies, standards used are well defined and communicated.

3) Take control

This is where you have the opportunity to make a tangible difference in how governance works within your organisation. The key elements are:

a) Defining data quality rules

This is where a data governance framework starts to deliver results. This pro-active process should enable you to report on the status of your data quality at any point in time - not only serving as a monitoring system, but also providing an early warning of any potential issues (before they get too big - and expensive!)

b) Data quality reporting

Only after data quality rules are defined will you be able to instigate a process for reporting on how the data measures up against those rules.

c) Monitoring and acting on data quality reports

This is where it comes back full circle to the initial establishment of a policy on data governance. With this, and the associated processes, in place you can take steps to ensure that those that need to take the necessary action, do so.

Taking all these steps, and embedding them within your organisation will help to promote and embed data quality and governance. This will deliver long-term benefits for the organisation as a whole, avoiding the repeated expense of regular, ad hoc data quality improvement initiatives, and help you to capitalise on the benefits that data quality can bring in a sustainable manner.

Janani Dumbleton is a principal consultant, data quality propositions, thought leadership at Experian Data Quality UK. She can be tweeted at @JananiDumbleton