All the regulations around data compliance and protection such as European Union’s GDPR, California’s CCPA, Australia’s Privacy Amendment (Notifiable Data Breaches) to Australia’s Privacy Act, and other similar regulations, force companies to go all into data governance to meet regulatory requirements. Besides regulatory compliance, companies are aware that they need to have a sound data governance framework if they want to have readily available, relevant and high-quality data for their projects.
When we talked to Nicola Askham, The Data Governance Coach, at last year’s Data 2020 Summit, we discovered that one of the most common challenges that firms come across when drafting a data governance policy is that they don’t know how much data governance is enough.
As Nicola Askham presented in her interview, some companies are too aspirational. They go from having no control over their data to wanting to have everything in place, which is a huge step to take.
While at the same time, there’s the opposite end of the extreme as well. Some people worry that if they make it too much hard work nobody will sign it off. So they include very little in their policy and end up not helping the organisation to mature in terms of their data governance.
How to know how much data governance to apply?
For starters, there is no such thing as a standard data governance programme. Every company is unique, with a unique structure, needs and goals, so an off-the-shelf data governance programme will only portray like you’re putting control and governance just for governance sake, warned Nicola Askham.
McKinsey educates that data governance programmes can vary dramatically across organisations and industries. For example, banks were the first forced to adopt a data governance programme due to the BCBS 239 (the Basel Committee on Banking Supervision’s standard number 239) and other regulations that required them to have sophisticated governance models.
Other industries and organisations don’t necessarily face such strict regulations. So McKinsey advises to design programs that correspond with the level of regulation and the level of data complexity. An example is a business spanning across multiple geo-locations that needs a more complex data governance model, in comparison to a business located in one geography.
To provide a more detailed outlook of the different types of data governance, McKinsey has created a chart of data governance archetypes considering the level of data-related regulation in the industry and the level of data complexity.
Mapping the regulations of their industry and level of data complexity on the McKinsey chart can help companies figure out which data governance archetype is relevant for their unique needs.
For example, a global bank with multiple product lines would need the most comprehensive governance model driven by a dedicated data governance council often involving C-suite leaders, with a high degree of automation with metadata recorded in an enterprise dictionary or data catalogue; data lineage tracked back to the source for many data elements; and a broader domain scope with ongoing prioritisation as enterprise needs shift.
By contrast, a small tech company would need targeted data governance focused on value creation, whose council includes C-suite leaders only periodically; metadata tracking that could even start in Excel; limited lineage tracking; and narrower domain scope that enables priority use cases, states McKinsey.
Minimal data governance: How to succeed with it
McKinsey’s categorisation of data governance models provides the much-needed direction when it comes to designing a data governance programme. It’s unrealistic to expect that we can govern all data we have in our organisation perfectly and all at once. This issue becomes even more relevant as we often witness people being asked to implement a data governance programme with little or no investment, but expect great results. Explore the Data 2030 Summit agenda
Nicola Askham, The Data Governance Coach, saw this from her experience with helping organisations with their data governance. She said that as a data governance expert, she learnt to be very practical and pragmatic when doing data governance. As she related during her Data Innovation Summit 2020, not only you can’t do data governance over every data, it’s not useful because not all data is the same value.
Especially in COVID-19 times and budget cuts, organisations need to carefully consider where to implement data governance. At the same time, the demand for data governance is increasing because organisations need relevant, high-quality data now more than ever to make the right decisions, emphasised Nicola. “People are not spending money on data governance, but they want more of it,” she summarises. However, data governance should deliver some value to the organisation. There is no point in doing data governance just for the sake of data governance.
Companies are more than ever looking to do data governance with little or no budget while at the same time get value. The current conundrum inspired Nicola to come up with an approach which enables people to be pragmatic while delivering results and paving the way for future governance efforts.
The minimal data governance entails a sensible and practical approach to proactively manage data quality. This approach is usually taken initially because the circumstances required it, but at the same time delivers real value to the organisation.
As Nicola presented, the best way to embark on this minimal data governance is to follow Simon Sinek’s methodology, the author of the book “Start with the Why”. It implies starting with the core ‘why’ we are doing something and then moving towards ‘how’ we are doing it and eventually, ‘what’ we are doing to achieve it. Sinek’s approach is vital not only to minimal data governance but also when going in for the full scope of it. Explaining to stakeholders why you are doing it will help get them on board much easier.
However, if doing a minimal approach to data governance it’s important to manage expectations with stakeholders, Nicola highlighted. So after presenting the benefits of data governance and why you are doing it, you have to explicitly explain the reason for the focused delivery and manage their expectations for the time of delivery. Although it’s called minimal data governance, it still takes a long time if it’s going to be done properly. And if it’s not done correctly on the first go, it will only make matters worse for the future, Nicola stated.
Going for the minimal approach also requires prioritising from the long list of benefits and defining your scope. Minimal data governance helps tightly define the first efforts of governance, and when delivered successfully, we can roll it out further into other sets of data.
After defining the scope, we have set the terrain for designing the data governance framework which should illustrate the policy, processes and roles & responsibilities. The minimal governance framework should be designed for the intermediate benefits, but also be the groundwork for the future. It should be created to evolve and grow just as the organisation grows so it can be extended in upcoming years.
Future-proof your data governances
Join the most important discussion for your data governance. The Data 2030 Summit is an event dedicated to data strategies for AI-driven organisations. The programme is divided into 3 days, each covering an important pillar of a Modern Data Management platform: Data Governance & Data Quality, Cloud-Enabled Modern Data Platform and DataOps (+ Master Data Management track).
Data 2030 Summit gathers the Data Management community and merges two regions (the Nordics and MEA) for a unique cross-regional benchmarking and networking, local and regional insights and benefits, all in a global knowledge-sharing platform.
This article was originally published at Hyperight Read.
