There is business urgency to deliver data-driven results today. The Data Mesh journey will require a transformational change program (people & process), but getting started is not difficult. With the right strategy, it is not labor-intensive and there is a low cost of entry. If you start small with the right use case, you can gain some early wins and build momentum. I had the pleasure of delivering a session on Data Mesh readiness at Datanova: The Data Mesh Summit.
Data Mesh offers an agile design that can be delivered as an MVP to meet your immediate needs. Most pathfinders — Data Mesh early adopters — start small and move fast by establishing basic integration across selected high-value data sources. A couple of domains and some sharp data products can make an immediate impact on an analytics use case and it demonstrates the effectiveness of the Data Mesh concept. These early wins create the executive support that’s required to take on bigger challenges. Let’s take a closer look at some key considerations to jump-start your Data Mesh Strategy.
Start with the business strategy
You should start with your organization’s data strategy or the digital transformation strategy. These documents are linked to your business strategy and they set the direction and priorities for your data initiatives. These are the key items you should focus on:
- Business Outcomes: The Data Mesh is designed for the business, you need to identify the strategic business objectives that are dependent on data analytics. These objectives can be broken down into smaller, prioritized use cases.
- Data Capabilities: There are three elements to this; technology capabilities, process integration, and data culture. Your domains and data products will be curated to support the business, so you may need to dig into the business processes as well.
- Priorities: Your pathfinder should focus on one use case. Select an area where the Data Mesh approach can create immediate value, within 90-days. With a Data Mesh, you don’t waste time migrating or moving data, you can start designing and building domains and data products within a few weeks.
Review your architecture design
Many organizations have a diverse data architecture, with multi-cloud environments and legacy on-premise solutions. And there are new data environments being built or acquired every year. The more diverse and decentralized your environment, the greater your business case for a Data Mesh.
Usually, the first question from the architecture team is, “How does it work?” It’s a valid question. We have been chasing the “single source of truth” vision for years. Migration has been slow, expensive and disruptive to the business. The vision for accessing enterprise data from a single point is Right. The approach, moving all your data to a big lake, is Wrong.
Starburst offers a proven alternative to your migration projects. We can connect to the data wherever it’s stored with Trino, a query technology that has been tested by the biggest digital organizations for ten years. Now, you marry Starburst with a Data Mesh design and you have a perfect match. Your Data Mesh pathfinder can quickly connect data sources and organize the data into products that can be easily consumed for your selected use case.
Kick-off a Data Mesh literacy program
Beyond the architecture, getting the most value out of enterprise data integration will require a digitally skilled culture that is ready to drive insight, action, and automation. Data-driven organizations are succeeding with self-service programs that empower business teams to build their own analytics solutions. I have seen it happen, I know it works. When your business teams can drive their own analytics, they will move fast and with greater agility. Data Mesh is focused on the business, and it promotes sharing data and self-service delivery. Literacy takes time and it certainly shouldn’t be a goal for your 90-day pathfinder, but getting started early will increase your ROI.
Governance is a game-changer
There are many considerations when it comes to governance. However, a common theme in this post you might have picked up by now is that with Data Mesh, you can work with what you already have. Leverage the data governance processes you already have in place. Use the pathfinder to learn, then you can start thinking about where you can improve. You don’t have to reinvent the wheel. Consider the following as you look for ways to improve:
- Federation: In a Data Mesh, you want to move towards a federated governance model where your Domain Owners can work with some autonomy.
- Value-based approach: Each domain should manage a governance model that balances risk vs value based on the use cases it supports. If you are dealing with low-risk data that can generate big value, then follow an open access model. If you are dealing with higher-risk data, then adjust accordingly. Domain Owners should understand the governance levers they can control.
- Master data: In a Starburst design, your data is not moved or copied. This will greatly simplify discussions about data management and duplication. It’s a game-changing advantage.
An invitation to have a dialogue on what Data Mesh means to you
If you would like to engage in a deeper discussion on how to get started, take the Data Mesh readiness assessment? It will help you assess where you can get started, how fast you should go, and where you will need to focus your efforts. The outcome is to create a design that fits your unique digital value journey. We’re certainly not going to build a complete Data Mesh overnight. However, we can rely on what you already have in place today, connect to your data sources, and then build a roadmap for what we want to build over time.