The Data Mesh approach goes further in generalizing the concept of data products. It distributes product creation across domains and considers data as a product itself! Data is no longer a component of a digital product. Data Mesh gives consumers of domains the responsibility and freedom to analyze and restitute data.
Data as a product: what do we mean?
In the Data Mesh approach, we have to distinguish between considering data as a product and the very concept of data as a product (which we saw in our first article!).
We can define a data product as a product that facilitates a goal thanks to data. Whereas when we talk about Data as a product: we imply that data becomes not an end but a means. Data becomes a product when it meets a set of characteristics that place the data between usability, feasibility and value.
6 key criterias for data to be a product
5 main types of data product
- Raw data
- Derived data
- Data resulting from the processing of source data by an algorithm
- Decision support data
- Automated decision support data
But to put it exactly, a data product is inspired by a DevOps approach, it is the combination of a dataset, the associated governance, the means (processes) needed to build it, its destination (analysis, communication, etc.) and its distribution packaging. The Data Product can also act as a data science algorithm, which, made available as an API, can be searched by domains.
Two prerequisites are needed to make the data product usable:
- the establishment of governance rules and processes
- standardization to facilitate its use
How to build & implement a data product approach?
The creation of data products is based on operational activities. It is important to select data sources, to document them, to detail the technical chain of data provision and its dissemination.
To implement one of the Data Mesh pillars, it is necessary to have a transformation within the organization with greater agility (in Spotify mode for example!).
In addition, it is necessary to involve employees in this vision of changing methods. A gradual adoption is necessary for companies of course, and they will be able to rely on agile teams in dedicated departments in the implementation of this new approach.
It will be necessary to be supported by a new key function: the data product manager to maintain and monitor the products. His role is to coordinate the activities related to the products for which he is responsible.
And to get started you will need to do with a first pilot product using the principles of the product roadmap and MVP. The creation of the product will allow to acquire these methodological and organizational skills.
To go further it is necessary to have an IT platform, which is the challenge of the third pillar: the self-service data infrastructure as a platform!
The design of a first product is the key stage of the initiation. It contributes to the transformation by introducing the principles of the product roadmap and MVP (Minimum Viable Product), while promoting agility and its benefits. It encourages producers to prioritize and therefore identify the functions and products that create the most value.
The pilot product will ideally focus on a relevant use case, which will require access to multiple data sources, close to the business and considered complex to access in the enterprise.
The creation of the product is an opportunity to acquire methodological and organizational skills. But to pursue agility, domains also need an IT platform and services that make it possible. This is the challenge of the third pillar of Data Mesh: self-service data infrastructure as a platform.