Introducing a data product approach in healthcare

Challenge

Getting a grip on an unclear data landscape

More than ever, companies are realizing the untapped potential of the data they collect. Our client, a major player in the Belgian healthcare industry, also came to this conclusion. Rather than just improving the current offering to its customers, it took up the ambitious plan to also create new products and services using the collected data.

In their search for the right approach, the company encountered significant obstacles. Their data landscape was unclear, influenced by a history of developing numerous custom applications, and knowledge was scattered throughout the organization. The envisaged shift in strategy also brought about new challenges, including to securely store sensitive Personally Identifiable Information (PII) data in the cloud, establish a systematic approach to data governance, and smoothly integrate new technologies into their current framework. To make this ambitious plan a reality and to tackle the challenges, it called on the expertise of Datashift.

Approach

Establishing the foundations

We started the project by conducting a thorough analysis of the current situation. The main goals were to get a proper understanding of the existing challenges and to align on the client’s preferred way of working to ensure optimal adoption.

Once the scope was defined, a Datashift team, consisting of both data engineers and data governance experts, was formed to take on this task. We established the general structure of the new cloud platform in Microsoft Azure to ensure high-level security and optimal performance. In parallel, we designed a Data Governance Roadmap, commencing with the first step of gaining a better understanding of the data landscape.

Introducing a data product methodology

In view of the many custom applications in use, we decided to adopt a phased approach. The main aim was to avoid a ‘big bang’ that would hinder the client’s operations.

Therefore, the client pinpointed the most crucial applications for us to concentrate on. We analyzed these priority applications, gathering information from both business users and data owners. Within each application, we identified one or more data products.

To structure the optimization of these data products, we employed the Data Product Operating Model. This model provides recommendations for the development, deployment and maintenance of data products within an organization.

To further facilitate the adoption of the operating model, we integrated Microsoft Purview with our data platform. Microsoft Purview streamlines data governance through features like data catalog and business glossary. It enables data discovery, data classification, data lineage and facilitates collaboration among teams in a user-friendly manner.

Shaping each data product

Drawing upon the insights gained through this exercise, we were able to design an optimized data model for each data product in Microsoft Azure. At the same time we updated our data catalog and business glossary, and defined clear roles and responsibilities within the organisation to ensure effective data governance.

Adoption of the new data products

We concluded the process with a comprehensive knowledge transfer session to maximize adoption of the new data product approach. As a final step, the solution was rolled-out successfully to the business, backed by change management and training efforts.

Impact

Optimized Data Landscape

The methodical, phased approach enhanced the understanding of the data landscape, benefiting both the team and the client. This improved understanding facilitated the optimization of the data models on the new cloud platform. The combination of optimized data models and an updated framework prepared our client to embrace new technologies, maximizing the use of their data. This readiness enables the development of new and improved applications, allowing the client to create innovative products and services for its customers.

Robust Data Governance and Increased Collaboration

The implementation of data governance, guided by our Data Governance Roadmap, introduced a more structured way of working in relation to data management. A good example is the use of Data Product Templates, that contain the key information from our analyses of the applications. The template serves multiple purposes. It informs data engineers on the key technical details of the applications and provides business users with insights into the origin of their data. It also serves as an effective summary for other stakeholders who need to be informed. This means that these templates are used at various stages, including analysis, development and roll-out.

The introduction of data governance has also improved the information flow throughout the organization. This improvement was achieved through the implementation of a data catalogue, business glossary and data lineage mapping in Microsoft Purview. The integration of data classification has played a pivotal role in the identification of sensitive data and organization-specific information. This aspect is crucial for compliance with data protection regulations, especially when handling PII data. Additionally, roles and responsibilities were assigned at the data asset level, resulting in increased accountability and ownership for individuals and teams within the organisation.

Want to exchange ideas on data productization?  Reach out to learn how Datashift can support you on this journey.