Second article in our series on Data Sharing. In the introductionwe recalled the current importance of Data Sharing for organizations, as well as the latest trends. We now turn to the key issues that need to be taken into account in any Data Sharing approach.
Contents
1. The regulatory challenges of Data Sharing
Data-sharing initiatives necessarily take place in a complex legal context. The RGPD is the mainstay when it comes to handling personal information, but it's not the only text to consider. The Open Data Directive, sector-specific regulations (e.g. in healthcare, finance, or transport), and more recent texts such as the Data Governance Act or the Data Act, have an impact on how datasets can be collected, hosted, processed or exchanged.
The aim behind these measures is to ensure trust between public and private players, while protecting fundamental rights (privacy, freedom of enterprise, etc.). The regulatory challenge is therefore twofold: to prevent abuse (or misappropriation of data) and to encourage a stable framework in which everyone knows how to share their data.
2. The essential data foundations for any data-sharing project
A data sharing project relies on robust data foundations. One of the building blocks is the technological platform or "data stack": we often speak of data lakes, data warehouses, or even hybrid architectures capable of storing and exploiting massive volumes. API or microservice layers are also required to enable real-time or batch data exchanges, as well as an orchestrator (ETL/ELT) and security features (encryption, strong authentication, etc.).
The other building block is data governance a more organizational mechanism designed to ensure the reliability, consistency and traceability of shared information.
It defines how to record the origin of datasets (lineage), and how to document their schema, quality and access rules. Without this transparency, a data lake can quickly become a "data swamp", where it is impossible to determine the reliability of any given piece of information. The principles of interoperability, promoted by the European Commission, encourage the adoption of common standards (formats, protocols) to facilitate portability and coupling between different ecosystems.
Finally, we must always bear in mind the transition to scale (industrialization), which requires us to consider efficiency and maintainability. The greater the number of datasets and producers & consumers, the greater the need for automation and solid processes: it becomes essential to deploy an in-house catalog or "data marketplace", to make available well-packaged datasets (metadata, description, usage examples), and to check regularly that governance remains adapted to changes in the scope.
3. Acculturation: from awareness to expertise
Data acculturation needs to start at all levels, from the business to senior management. Understanding why and how to share data, what the opportunities are (innovation, efficiency, new offers) and what the risks are (non-compliance, security, etc.) is a prerequisite for avoiding reluctance and resistance. We need to overcome the "silo" culture, and get business managers to share their data rather than holding on to it as an internal competitive advantage.
Once awareness has been established, the next step is to raise the level of skills: recruit or train data engineers, data scientists, data analysts, and a range of cross-functional functions such as data stewards.
But data culture should not be limited to these specialized roles: it also concerns the "business" functions, which become capable of identifying concrete data sharing use cases and implementing them with the help of technical teams.
4. The business model: value creation and distribution
Sharing data involves effort (in terms of collection, cleansing and compliance) and risk (loss of a strategic asset or fear of disclosure). For this to work, participants in a data ecosystem (public, private, associations, etc.) need to see a return on their investment, which can take several forms.
- Internal efficiency improvement: less duplication, richer insights
- A new commercial service:selling or monetizing access to datasets
- Reciprocal exchanges: I contribute my data, and benefit from others' in return.
- A general interest rationale: the State or local authorities could provide support in exchange for data, in particular to combat pollution, optimize mobility, or improve urban planning.
5. Cybersecurity to protect shared data
- Appropriate governance: with, for example, the implementation of a "contract of trust" clarifying the scope of each stakeholder, the way in which authorizations are managed, the division of responsibilities in the event of an incident, procedures for reporting a breach, and so on.
- Technical measures: data encryption protocols, rigorous flow controls, traceability of actions.
- Team awareness and training play an essential role: it's often human error that causes data leaks.
Data Sharing: definition, trends and success stories
Data sharing is the pooling of data between different entities, to create value that could not be achieved by remaining compartmentalized. Find out more in this article
Opening up data assets with an AI agent: challenges and steps
While it may seem like a long way to go before you can deploy an AgentAI on your data assets, many companies already have sufficient data assets, but they are often poorly organized, poorly managed and poorly managed.
Data governance: definition and key principles
Data governance is a major lever for improving performance and accelerating business transformation. Today, data is a major corporate asset, with
