Get the essential data observability guide
Download this guide to learn:
What is data observability?
4 pillars of data observability
How to evaluate platforms
Common mistakes to avoid
The ROI of data observability
Unlock now
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Sign up for a free data observability workshop today.
Assess your company's data health and learn how to start monitoring your entire data stack.
Book free workshop
Sign up for news, updates, and events
Subscribe for free
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Getting started with Data Observability Guide

Make a plan to implement data observability across your company’s entire data stack

Download for free
Book a data observability workshop with an expert.

Assess your company's data health and learn how to start monitoring your entire data stack.

Book free workshop

What is a Data Mesh? Definition, Examples, and Best Practices

As data management continues to evolve in complexity with vast amounts of data being produced daily, there is a growing need for a more decentralized and scalable data management approach. Enter "Data Mesh", a term coined by Zhamak Dehghani of Thoughtworks. In this blog post, we will explore the Data Mesh architecture and its benefits, challenges, and best-practices. Additionally, we will look at real-life use cases to demonstrate the practical application of a Data Mesh.

May 23, 2023


May 23, 2023
What is a Data Mesh? Definition, Examples, and Best Practices

What is a Data Mesh?

Data Mesh is about shifting from a centralized data platform managed by a centralized team to a federation of domain-oriented, decentralized data management. This approach breaks down data bottlenecks and silos within an organization, allowing each domain team to take full ownership of their domain data. This results in increased scalability, faster insights, and a more optimized data-driven decision-making process.

Dehghani's vision of a Data Mesh is simple – self-serve data platforms that enable data consumers to obtain the data they need without involving a centralized data team. This leads to federated data sources and federated analytical data products where domain teams and data product owners are responsible for their own data products. While there are clear benefits to a Data Mesh architecture, there are also challenges – particularly around data ownership, governance, and trust, which we will explore further in the next section.

Key Concepts of a Data Mesh

Let's take a closer look at some of the key concepts that make up a Data Mesh architecture.

Domain-oriented Decentralized Data Management
In a Data Mesh architecture, decentralized domain-oriented data teams take ownership of their data. This allows for increased autonomy and innovation in creating data products while also breaking down data silos within the organization.

Self-serve Data Platforms and APIs
Data consummation teams require self-service data platforms that enable them to obtain the data they need without involving the centralized data team. This leads to federated analytical data products where domain teams and data product owners are responsible for their own data products.

Discoverability and Accessibility of Datasets
To ensure that the data is not siloed within teams, data sources need to be discoverable and accessible throughout the organization. This requires a change in the way data sources and datasets are created, managed, and accessed.

Data Mesh Best Practices

Implementing a Data Mesh effectively requires adherence to certain best practices. These best practices include:

  1. Ensure Clear and Uniform Interfaces and Standards for Data Sources and Pipelines: Establish a consistent framework for how data should be structured, accessed, and delivered. This involves creating well-defined interfaces that allow teams to interact with the data uniformly, regardless of the underlying data source. This standardization ensures data interoperability, reduces the complexity of data consumption, and enhances the usability of data across different domains within the organization.
  2. Establish and Enforce Data Governance and Domain Ownership: Data governance is crucial to ensure that the right practices are in place to manage and protect data assets. This includes clarifying data ownership roles, establishing clear policies around data use, privacy, and security, and ensuring that teams adhere to these policies. Additionally, establishing and respecting domain ownership encourages teams to take responsibility for their data, promoting better data quality and management.
  3. Pay Close Attention to Data Quality and Metadata: Data quality is crucial for generating reliable insights, and this quality is largely dependent on the accuracy, consistency, and completeness of the metadata. Metadata, the data that describes other data, should be consistently managed and accessible, providing critical context for data users. Regularly conducting data audits can help identify and correct any issues, ensuring that the data remains trustworthy and useful.
  4. Implement Access Controls and Automated Processes: To ensure data security and privacy, it's important to implement robust access controls. This means defining who can access certain data and the level of their permissions. Also, leveraging automation can streamline data processes, reduce manual errors, and free up teams to focus on higher-level tasks. This could involve automating data pipelines, validation processes, or other repetitive tasks.
  5. Encourage Domain-Driven Design and Involving Business Users in Data Mesh Development: A domain-driven approach encourages teams to design their data architecture based on their specific needs, which enhances relevance and usability. Inviting business users to contribute to Data Mesh development ensures that the system meets their needs, improving acceptance and adoption across the organization.
  6. Promote a Culture of Collaboration and Communication: Implementing a Data Mesh requires a shift in mindset, moving from a centralized approach to a distributed one. This change can be facilitated by fostering a culture of collaboration, where teams work together to manage data, share insights, and overcome challenges. Regular communication about data usage, issues, and updates can also help keep all stakeholders aligned.
  7. Invest in Proper Training and Skill Development: As with any new framework, ensuring your teams are equipped with the necessary knowledge and skills to use a Data Mesh effectively is critical. Regular training and upskilling sessions can help them understand the architecture, the technologies involved, and their responsibilities regarding data management.
  8. Ensure Data Discoverability: A critical aspect of a successful Data Mesh implementation is making sure that data is discoverable. Implement tools or procedures that help teams understand what data is available, where it is, and how to access it. This will facilitate effective data utilization across different domains.
  9. Plan for Scalability: Given the distributed nature of a Data Mesh, it's essential to consider scalability from the beginning. Design your Data Mesh in such a way that it can easily accommodate increasing amounts of data and a growing number of teams without a drop in performance or manageability.
  10. Iterate and Evolve: Understand that your initial implementation of a Data Mesh may not be perfect, and be prepared to iterate and evolve. Encourage feedback from teams, learn from the challenges faced, and continually optimize your architecture to better suit your organization's changing needs and objectives.

Data Mesh Use Cases and Examples

Let's take a look at some use cases for Data Mesh and how Rainforest, as a fictional company, might use this framework to improve their data processes.

Activated Users Table for Marketing Emails
To illustrate, let's say that you're VP of Data at Rainforest, and you want to automate marketing emails to customers that show the highest level of engagement with the company's products. You have an `activated_users` table that stores user engagement data, and you want to segment this data for targeted marketing campaigns.

In a traditional data setup, this process would involve your data engineers querying the data warehouse, building the data pipeline, and creating the data models. However, with a Data Mesh approach, the responsibility of the `activated_users` table's ownership passes to the domain team. Now, marketers can access this data through self-service data platforms and APIs, reducing the time spent by the data engineering team on these tasks. This leads to a faster turnaround time for marketing campaigns, resulting in better ROI overall.

Daily Revenue Table for Sales Decisions
To take another use case, let's assume that the VP of Sales at Rainforest is making critical decisions based on daily revenue data. In a traditional setup, the data team would produce a daily revenue table, which the VP of Sales would use to make crucial business decisions. However, with a Data Mesh approach, the VP of Sales can access the `daily_revenue` table through self-serve data platforms and APIs. This leads to more accurate and timely data analysis, giving the VP of Sales the ability to make quicker, data-driven decisions.

Sale Leader's Hiring Decisions and Machine Learning Features
Sales teams often use data to make better hiring decisions. To take another use case, let's assume that the Sales Leader at Rainforest wants to review the payroll budget and revenue data to determine how much to budget for hiring. Similarly, the Data Science team is developing machine learning features that will be embedded into the product, which need to be built with different data sets.

In a Data Mesh setup, the Sales Leader and Data Science team have access to microservices and tools necessary to collect and analyze data. This leads to more accurate data analysis, better decisions, and more cost-effective and beneficial features.

Real-Life Examples of Data Mesh

Some real-life examples of companies successfully implementing this approach:

  • ING: This Dutch multinational banking and financial services corporation has been embracing a decentralized, domain-oriented approach to their data management. Their efforts have been focused on promoting self-serve data infrastructure, which is a key principle of the Data Mesh concept.
  • Zalando: This e-commerce company implemented a domain-driven data platform that enabled domain teams to develop, test, and deploy their data products independently. This approach aligns with the Data Mesh's principles and has helped Zalando handle huge amounts of customer and product data effectively.
  • Intuit: Introduced a domain-oriented team structure to manage their data operations, enabling independent product development teams to own and manage their own data platforms.


In summary, Data Mesh is a decentralized approach to data management that empowers domain teams to own their own data and data products. By providing self-serve data platforms and APIs, Data Mesh enables domain users to access and integrate data into their products, leading to federated data sources and faster insights. While implementing a Data Mesh framework can have challenges, such as data ownership and governance, best practices, including data quality and metadata, domain-driven design, and access controls, can help ensure success.

Metaplane is a data observability platform for modern data stacks, providing monitoring and troubleshooting tools to ensure data quality. Our platform is user-friendly, transparent, and designed to help data teams focus on what's crucial: delivering data products with confidence.

Sign up for free or book a demo to learn more.

Table of contents

    We’re hard at work helping you improve trust in your data in less time than ever. We promise to send a maximum of 1 update email per week.

    Your email
    No items found.
    Ensure trust in data

    Start monitoring your data in minutes.

    Connect your warehouse and start generating a baseline in less than 10 minutes. Start for free, no credit-card required.