- blog
Data Governance - A complete guide to what exactly it is. Why and how to implement?
Data Governance - not the ugly duckling anymore!
Unpacking Data Governance vs Data Management vs Data Enablement - What actually is Data Governance?
Data Management Body of Knowledge from Data Management Association (DAMA) defines Data Governance as the planning, oversight, and control over the management of data and the use of data-related resources. Its controls cover the full data lifecycle, from the initial creation, to the reading, usage, and eventual deletion.
Every Data Governance program has a uniquely defined scope, but common themes include data protection, data retention, proper use of data, data security and the management of data as a business asset.
Data governance provides direction and oversight for managing data which includes its quality, reporting, its security, privacy, and lineage etc.
eliiza looks at data governance as a data βenablerβ. Practical policies and standards combined with effective data management enable users to get value from data. From evidence based decision making to creating effective strategies, data enablement is the best outcome of the data available, not only at the start but throughout the lifecycle of data. From being collected at the source until archiving, data governance can strengthen the way the data is used, managed and monitored at all points.
Often people try to understand the difference between data governance, data management and data enablement. This table shows a good overview of the differences between the three:
| Data Governance | Data Management | Data Enablement |
|---|---|---|
|
Definition of roles, responsibilities, policies, standards and direction
- Strategy and policies - Framework |
Technical implementation of data governance and management of data in real time
- Implementation and operations - Strategic roadmap/action plan |
Tangible business outcomes resulting from effective data governance and management practices
- Empowered decision making - Outcomes |
Itβs no more a βnice to haveβ. Organisations have tried, are trying hard. Letβs get into a few example use cases.
More organisations have realised the importance of governing data more than ever before. Many of which have attempted to establish a data governance function earlier but have not been able to sustain the function over time or the function is not able to produce the desired outcomes.
In one such ANZ retailer, notifiable data breaches were significantly increasing. As with every customer centric organisation, the firm was in the midst of a large digital transformation and faced multiple data related issues on data accuracy, validity, sensitivity for sharing etc.
It took longer and longer to fix data related issues. That is when the organisation took a step back and observed these symptoms of what they deemed to be a possible collection of scenarios that needed an implementation of structured Data Governance at their organisation.
A workgroup, with key people from all departments, was formed to analyse the root causes.Β
Results coincided with typical issues faced by organisations with a operational data ecosystem:
- Multiple sources of business critical data. eg multiple sources and copies of customer data
- Quality of data that is captured across the ecosystem is poor
- No visibility of data flow within and across the enterprise
- The organisation has not agreed on the classification of data
- no structured responsibility and accountability to take data related decisions
- No visibility of data capabilities across the organisation leading to tools and process redundancies
- Data was always an afterthought and not considered as a key enabled or treated as an asset
The same group was formalised as the Data Governance Working Group through a Charter to address the above issues.
An unstructured approach to data governance is one of key causes of operational inefficiencies, heightened risk of data related incidents and breaches, impaired decision making and increased cost of integration.
Implementation of a structured Data Governance process will alleviate these challenges and enable creation of a foundation that organisations can use to scale their data management maturity further.
How important is Data governance? It is βextremely importantβ with βlow success rateβ
Organisations are already performing some form of data governance and management across the enterprise over time. However, most of these processes or activities were not repeatable or shared enterprise wide, and were performed in silos. Data Governance aims to harmonise these activities into well-orchestrated processes and provides empowerment to individuals by formalising those responsibilities across the enterprise.
As mentioned above, organisations have already resorted to implementing a data governance program with the foremost priority of the program being to enable business driven compliance. The other priorities in that order of preference would be to establish people & process practices and then drive data standards through technology & tools.
In a recent research with 150+ data professionals by TDWI on the state of data governance at their organisations, it was noted that while over 80% say that data governance is βextremely importantβ,Β only 8% consider their typically enterprise-wide data governance program as highly successful. Over 90% are now considering uplifting data governance as an opportunity because it further ensures compliance, and provides internal standards for improving data and its management.
Key Principles of Data Governance - think holistic, start local and adapt iteratively for effective Data Governance
While there is no one-size-fits-all approach to implementing data governance, it is important it still carries a certain SHAPE to ensure a successful implementation and/or ease of communication. So, do remember βSHAPEβ when it comes to data governance as it needs to be:
Simple: We have often seen successful programs start adopting simple incremental, non-intrusive steps as much as possible.
Holistic: Pin down the scope of data governance, and carefully outline the immediate actionable needs of the organisation
Adaptive: Donβt boil the ocean. Think big by considering the needs of the wider community, but scale it iteratively.
Pragmatic: Aim for actionable work steps and deliverables and refrain from sticking to rigid methodologies that cannot be realised
Enabler: Consider Data Governance as an enabler rather than a blocker that hinders the organisationβs overall performance and agility.
Keeping SHAPE in mind, there seems to be these contemporary structures of operating models, outlined below, in the various organisations that are looking to establish successful data governance programs.
Top-down & Bottom-up:
With SHAPE characteristics aside, every Data Governance program can be structured in a few different ways aligned to the culture of the organisation. Clients nowadays adopt a combined top-down & bottom-up approach that has often proved effective. Recently, another large customer centric organisation and also an NGO adopted this approach to fruition. The leadership team can drive the strategy and pass that down to the so-called data stewards and data owners who provide them the tools for operative functions.
The stewards on the other hand provide valuable real time information from operative functions that help the leadership team in creating effective strategies.
Depending on the structure and operations of the company, data governance can vary significantly. For multiple project based operations for example, it is seen to be ideal to have a steering committee overseeing a project with multiple working groups sitting below it for specific outcomes.
Federated in a data mesh:
Data mesh is the most talked about approach to data management and ownership in recent times. In a data mesh structure data is decentralised and each domain takes end-to-end responsibility for their data. Data governance is the key to an effective data mesh structure. With clearly defined roles and responsibilities and centralising data governance, the data needs for the whole organisation can be met in the best possible way. The policies, standards and guidelines provided by a centralised data governance supports the various domains to make the best use of their data. This can cover things such as data sharing agreements, data quality standards, data lineage requirements etc.
Common mistakes when implementing data governance - beware there are landmines along the journey:
Even after considering all the above aspects, many organisations struggle with obtaining value out of their data governance program. Few of the landmines, our clients often encounter, are:
- Lack of Business Ownership: Data Governance is often seen as a technology problem. However, Technology is only a custodian of data. Data governance is a business problem. Business divisions are the owners of data. Business ownership and participation is vital.
It is possible that a best technological solution and foundation can be pre-empted and designed as a means to promulgate data governance in an organisation, however if the data owners do not adopt them and effectively contribute with domain knowledge related data, the return on investment from the technology investments cannot be realised. - Inadequate Senior Management Support and Sponsorship: Our experience notes that most of the strategic data and governance initiatives fail due to lack of senior management support, sponsorship and understanding of what and why of data governance. In most cases, it is noted that C-level sponsorship for a data governance program is vital to ensure the processes and actions are adopted across the enterprise.
- Lack of Data Leadership and Accountabilities: This happened in one such financial services organisation, a structured approach to data governance was developed, however, the organisation had failed to invest in establishing a dedicated Data Governance leadership for accountability without adequate study on the utilisation and workload of the team. This led to lack of focus on actions, lack of progress on initiatives and eventually diminishing the value that was originally envisaged from Data Governance. Again, it is vital to pin down responsibilities and accountabilities to individuals.
- Inability to articulate or demonstrate Business Value of Data Governance: Without effective articulation of business value of data governance, it would be challenging to achieve buy-in from business stakeholders who are the actual owners of data. Their buy-in is critical to the success of the program. With the advent of cloud, agile project management methodologies and devops, the pace of change has picked up heaps. This leads to an increase in the expectation of outcomes within organisational stakeholders.
Donβt leave it too late, time to get the Data Governance duck in line.
Data governance is a journey and a team sport which requires culture change as much as it requires process & technology change and should not be treated as an individual project. With the Government’s new privacy breach penalties increasing up to $50 million, it takes time and effort to inculcate the culture in people to treat data like an asset and protect it at all costs.
We specialise in this data governance journey and travel this road often with our clients. We understand the know-how to avoid the landmines which come along the way and specialise at taking an informed approach that better suits your organisation.Time for the shameless plug! π - if you need a reliable soundboard, reach out to us and letβs travel down this road together.