Practice : Data Mesh
Purpose and Strategic Importance
Data Mesh is a decentralised approach to data architecture that treats data as a product and assigns domain-oriented teams with full ownership of their data pipelines, governance, and quality. It shifts away from centralised monolithic platforms to scalable, self-serve data platforms and federated governance.
By aligning data ownership with domain expertise, Data Mesh helps organisations democratise access to high-quality, real-time, and trustworthy data - enabling better decision-making, faster insights, and data-driven innovation at scale.
Description of the Practice
- Domain teams own, produce, and serve data products (e.g. datasets, APIs, event streams) within their contexts.
- A self-serve data platform provides tooling, infrastructure, and standards for data discovery, access, and quality.
- Governance is federated - shared accountability for data contracts, quality, lineage, and access policies.
- Data products are discoverable, versioned, and built with observability and trustworthiness in mind.
- Emphasises interoperability, decentralised ownership, and data-as-a-product thinking.
How to Practise It (Playbook)
1. Getting Started
- Identify a candidate domain with rich operational data and an engaged engineering team.
- Define the first data product - what problem it solves, who consumes it, and its expected SLAs.
- Establish baseline observability, quality metrics, and ownership within the team.
- Build and publish the data product with clear contracts and metadata.
2. Scaling and Maturing
- Enable teams with a self-service platform for ingestion, transformation, storage, and publishing.
- Introduce federated governance policies (naming, tagging, access, retention, quality).
- Catalogue and index all data products in a shared discovery portal.
- Evolve contracts with versioning, validation, and consumer collaboration.
- Integrate lineage tracking and usage analytics to drive improvements.
3. Team Behaviours to Encourage
- Think in terms of data products - well-documented, trusted, and usable assets.
- Collaborate with consumers and stakeholders to meet analytical and operational needs.
- Monitor data quality and trust signals - proactively address issues.
- Treat data production as a first-class part of software delivery.
4. Watch Out For…
- Inconsistent standards or fragmentation without a strong platform and governance.
- Overhead for teams without clear value or support.
- Silos emerging if domains avoid shared protocols and interoperability.
- Data sprawl from poor lifecycle or ownership hygiene.
5. Signals of Success
- Data products are discoverable, usable, and aligned to domain ownership.
- Consumers trust and use domain data to drive product and business decisions.
- New data products are published and iterated quickly.
- Data quality and observability are embedded into delivery pipelines.
- Decentralised teams confidently produce, share, and evolve data.