Data governance: everything you need to know to support your data strategy

What is data governance?

Data governance refers to all the bodies, rules, roles and responsibilities related to data handling in a corporation, from its collection and processing to its activation. Governance is the bedrock of a company’s data strategy since it regulates key issues around data: security, usability, integrity and reliability.

Providing this governance, framing data collection, usage and sharing guidelines across the company are at the heart of the Chief Data Officer’s tasks.

Data governance: the bedrock of data-driven strategies

Today, data impacts every level within a company, meaning that it needs to be managed with a comprehensive and transversal strategy. This is why data governance is essential, as it controls data availability, interpretability, integrity and security, all for the benefit of the business.

The way data is accessed, shared and centralized bears both technical and organizational challenges. Governance means establishing rules about data management, knowing who the data “belongs to” at every stage of its lifecycle, as well as who is responsible for it and how, for whom and for which purposes it can be collected and exploited.

A data-driven strategy runs the risk of not reaching its full potential without organizational adjustments within the company. Establishing a data governance strategy means mobilizing the skills, tools, processes and rules which make it a genuine business asset, in the same way as a production or supply chain.

How to develop an appropriate data governance strategy?

Developing a data governance strategy is a task that requires a deep understanding of the structure and inner workings of the company. Governance is not set in stone and may evolve over time. It should keep pace with the development of the data strategy and of use cases implementation. As a result, governance needs to be pragmatic, iterative and guided by business objectives.

This transversal task requires the definition and setting-up of various rules, procedures, indicators and reference frames. As a result, daily usages of data within the company will be identified and data coherence, homogeneity, reliability and availability will be ensured.

It is worth noting that some organisations choose to start by generating value from their data rather than building governance upfront. The idea is to impulse the demand for data by creating a need internally, and to then organize data management accordingly. The potential benefit of this strategy would be to align data governance with the company’s business needs, making it coherent with the business vision.

Unfortunately, there is no reference model for data governance, as the right model strongly depends on the company’s initial organisational structure. As a consequence, there are as many data governance strategies as there are companies!

fifty-five’s best practices for a robust and efficient data governance

• Build upon a cross-functional and transversal governance body which combines business, technical, analytical, legal, statistical and IT skills. A core cross-departmental team is a relevant way of ensuring a holistic vision of the company’s activity and as a result, an adequate framework. This team would bring a comprehensive understanding of the company’s challenges during decision-making.

• Capitalize on specific skill set of each department or contractor, as their know-how may be rare and highly relevant.

• Clearly define the roles and responsibilities of each stakeholder at every stage, both internally and externally. This will empower each party and make it aware of its responsibility for its tasks. The chosen governance strategy will therefore make decision-making easier and ensure that data use cases are aligned with the company’s global strategy.

• Identify and prioritize business needs in order to build use cases which can later be translated into technical requirements. All stakeholders should be involved in the process of selection and design of the use cases to provide their specific insights. The business teams should be included in this process as early as possible to ensure the strategy fits the business needs. Legal and security teams should give the greenlight for data usability while technical teams should be involved in more operational tasks. The implementation of use cases must, of course, be precisely documented in order to guarantee their sustainability.

• “Start small, grow smart”: by starting with a simple and pragmatic architecture and by enhancing it with iterations, controlling data quality and deploying security measures will be much easier. The objective is to target the scalability of use cases and not their comprehensiveness. A project’s success depends on a team’s ability to optimize and adapt use cases throughout the data management process. In this respect, assessing the cost of technical solutions shouldn’t be done based on a POC but on a global company-level.

• Ensure data security and comply with GDPR personal data requirements: using Privacy by Design, obtaining users’ consent to use data for explicit purposes, and being able to determine the impact of a data breach immediately should be made mandatory by the data governance.

• Bear in mind that technical solutions, whether they be ad hoc or “all inclusive”, are only relevant if they are managed by skilled people who can use them as effectively as possible for business purposes.