At the heart of the Data Mesh architectural approach (simplified)

In the era of big data and AI, many organizations struggle to manage and govern the volumes of data they produce. For a long time, the default solution was straightforward: create a centralized data team, responsible for the entire company's data assets — defining standards, enforcing common rules, and taking ownership across all domains.

On paper, it made sense. In practice, it created something far more problematic: a bottleneck.

As companies grew and volumes exploded, this central team became the single point of contact for every data request, every pipeline fix, every governance question. Business teams sometimes waited weeks to get a response.

What is Data Mesh?

Conceived by Zhamak Dehghani in 2019, Data Mesh is neither a technology nor a tool you install. It is a new way of thinking about how organisations own, manage, and share their data at scale.

The core idea is simple: rather than concentrating all data responsibilities within a single central team, data ownership is distributed to the domains that know it best.

In this way, the marketing team manages marketing data. The logistics team manages logistics data. Each domain becomes responsible for the quality, availability, and reliability of its own data.

The four principles of the Data Mesh approach

• Domain ownership: data is owned and managed by the business domain that generates it. The people closest to the data are best positioned to understand, maintain, and extract value from it.
• Data as a product: each domain does not merely store data — it publishes it as a product. This means treating datasets with the same rigor as a software product: clear documentation, reliable quality, and identified consumers defined upfront.
• Self-serve platform: to enable domains to work autonomously, the organisation provides a shared platform that allows any domain to build, publish, and consume data products without advanced technical expertise.
• Federated governance: a common foundation of standards and rules applies across all domains — particularly around security and privacy — while leaving each domain the freedom to organise itself as it sees fit.

Data Mesh is also a cultural and organizational shift

Data Mesh is first and foremost a shift in mindset before it is a shift in architecture. It asks organisations to stop treating data as a byproduct of their operations, and to start viewing it as something that has owners, consumers, and a level of quality to uphold.

It also requires a strong commitment from all employees in terms of adoption, as the Data Mesh architecture involves supporting users so they progressively gain greater autonomy in managing their data scope and the associated tools.