Data Mesh, conceptualized by Zhamak Dehghani in 2019, proposes treating data as a product, managed by the business teams who know it best. After several years of adoption in large enterprises, it's time for an assessment.
The 4 fundamental principles
1. Domain ownership: each business domain is responsible for its data, quality, and availability. 2. Data as a Product: data is treated with the same standards as a software product — documentation, SLAs, versioning. 3. Self-serve data platform: a technical platform enabling business teams to publish and consume data without strong dependency on a central data team. 4. Federated computational governance: quality, security, and interoperability standards defined globally but applied locally.
Our recommendation: pragmatic Data Mesh
Rather than an organizational big bang, we recommend progressive adoption: start with 2-3 pilot domains, invest first in the self-serve platform (data catalog, data contracts, observability), and measure data product quality before scaling. Data Mesh is not a goal in itself — it's a means to accelerate data-driven decision making.
DataData MeshGovernanceArchitecture
About the author
CR
Camille ROUSSEAU
Data & AI Consultant, ACCENSEO
Data engineer and AI consultant, Camille helps enterprises structure their data strategy, from governance to production-grade AI model deployment.
