The (poisoned) gift of data governance

Data Governance is not a new term. However, for most companies it is a new discipline. A discipline created out of necessity, to maintain control over the ever-growing number of data sources an organisation has access to.

At its core, Data Governance is about enabling data users in an organisation to find, understand and trust data. Capabilities which are essential to building a data driven organisation. In practice, most data governance programmes have failed to deliver on this promise.

The most common pitfalls we see include:

1. Trying to govern all data in an organisation

2. Lack of change management

3. Being perceived as the internal data police, stifling innovation

Trying to govern all data in an organisation

Let’s unwrap them one by one. Trying to govern all data in an organisation is like trying to reverse the expansion of the universe, it’s a fight you’re going to lose. As data gets more and more democratised, with overall data literacy increasing and people getting access to self-service data and analytics tools, the number of data assets grows exponentially. To use your limited data governance resources effectively, consider implementing a tiered data governance system (with for example “Bronze”, “Silver” and “Gold” tiers) where you only put the strictest data governance requirements on your most business-critical data. Your sales dashboard going to the senior management and its underlying data sets will be Gold, with clear data definitions, high data quality standards and responsible data stewards assigned in case of issues. On the other hand, a new data set created for an innovative proof of concept will not need to meet the same data governance standard.

Lack of change management

When people only see Data Governance as an annoyance, as something they have to comply with, the change management aspect has probably been forgotten in the Data Governance programme. Appointing Data Stewards and implementing a Data Catalogue are not enough. People need to:

• be made aware of the hidden cost of ungoverned data,

• see the benefits of data governance in terms of easily finding and understanding trustworthy data,

• be trained to understand the impact of data quality issues on processes and results,

• receive the time to invest in data governance related initiatives.

Being perceived as the internal data police, stifling innovation

In the absence of these elements Data Governance and its representatives will be seen as the internal data police, tell people what they must do and cannot do with data. This is a far cry from the need from which the Data Governance practice arose. At its core, any Data Governance programme must aim to unburden the people working with data, whether they are dashboard users, data analysts, data engineers or data scientists. In our experience, the best way to do this is by creating an internal data marketplace with:

• Trusted data sets

• Information about how to access the data (which tools/technologies)

• Metadata such as when the data was last updated, who to contact in case of problems and any known data quality issues.