Data Mesh and Change Management

In the evolving landscape of data management, the concept of a data mesh introduces a significant shift by assigning ownership of data domains to individual stakeholders. This decentralization means business owners are now directly accountable for their respective domains’ data and related assets. This transition marks a dramatic departure from the traditional model where responsibility rested solely with centralized data teams. This centralized approach often leads to fragmented knowledge and siloed information, whereas data mesh encourages closer alignment between business needs and data management.

Data mesh is not just an architectural shift but a complete overhaul that matches data products to business assets, making it easier for business users to find, understand, and use trusted data. However, common challenges include fostering business adoption, establishing ownership, and ensuring self-sufficiency in data stewardship. 

For these reasons, implementing data mesh requires more than just a technical restructuring—it involves change management: A structured approach to developing new processes, engaging stakeholders, and providing ongoing support to ensure a smooth transition. Whether an organization opts for data mesh or another approach like data fabric, the focus on effective change management remains the same to drive successful adoption and long-term value.

ADKAR for data mesh implementation

Let’s say your organization has decided to implement data mesh. To create a successful onboarding and engagement process, it's essential to integrate key elements that drive change and foster involvement. 

Start by applying Prosci’s ADKAR model, which involves:

• Creating Awareness of the need for change

• Fostering Desire to support the change

• Providing Knowledge on how to implement it

• Developing the Ability to carry out the change activities

• Ensuring Reinforcement to sustain the change

Creating awareness and knowledge can start by building out use cases, focusing on demonstrating the benefits of the change, such as time savings, increased revenue, risk reduction, and/or end-user stories. Clearly communicate how these changes will ultimately help business users and stakeholders. A detailed change management use case should include a description of the current state, the reasons for the change, the benefits it brings, and who is responsible for managing the change. 

To learn how Fifth Third Bank leveraged ADKAR to scale data governance and literacy across the enterprise, watch this webinar. 

Data mesh change management use case example

For example, let’s imagine a scenario where a retail enterprise just moved to a data mesh architecture, but the supply chain team (a team of 100) still asks the data team (a team of five) questions about the data. This indicates the supply chain team is struggling to take full ownership of the data. Leaders overseeing the data mesh implementation would illustrate the use case and catalyze change by laying out the business problem in the following way:

Current state: After converting to data mesh architecture, the data team still receives ad hoc questions regarding data products and reports, even though the business group owns that data.  When the supply chain team has a question, they have to submit a ticket, and it takes hours or days to get a resolution. If each person on the supply chain team had 1 question a week: That's 100 questions divided by five people on the data team, which is 20 questions per analyst to find, understand, and model a report. If each question takes three hours to address, that’s 60 hours the data team would spend on just ad-hoc questions alone each week.   

Why change: By enabling self-service, leadership can transfer responsibility for knowing, updating, and sharing the data to the supply chain team. This will reduce the time the team spends waiting for data objects, reports, and business context.

Benefits: Leveraging the business expertise within the supply chain can significantly reduce time, increase productivity, and build greater trust in the data as the supply chain team takes ownership of their data. This distributed ownership model also enhances data governance, as the governance team can focus on enabling better policies and oversight across various domains.

For example, consider the impact on time savings: if 50 individuals on the supply chain team use self-service tools to access and manage their data, it would reduce ad-hoc questions by 50%. This shift would free up analysts, who typically spend 30 hours a week finding, understanding, and modeling reports, allowing them to focus more on high-value tasks like predictive analytics or other critical areas of their roles. This not only increases efficiency but also empowers teams to drive more informed, data-driven decisions.

Who it affects and will be managed by: This change impacts the supply chain team and the analytics team. The supply chain data will be managed by three named supply chain business owners and ten supply chain stewards

Best practices for implementing data mesh

Clarify roles and responsibilities

Clear role definitions are crucial for an effective data mesh rollout. It's essential to clearly identify who is responsible for each domain and ensure their roles and responsibilities are well-articulated. This includes outlining the specific tasks, time commitments, and level of involvement required for successful data ownership.

To identify domain owners, consider a few approaches: ask leadership to nominate candidates, seek volunteers from within business units, or leverage the expertise of top users of key data assets. Defining ownership in this way helps ensure that those closest to the data are accountable and engaged, creating a stronger foundation for data governance.

Establish a data mesh council

Additionally, establishing a committee or stewardship council can enhance and reinforce the decision-making process and manage change more effectively. This group should communicate the changes, address concerns, and oversee the change management process, focusing on behavioral adjustments and active participation.

Engage leadership support

Another critical element in reinforcing the implementation of a data mesh is involving leadership in the process and communication. Leaders play a pivotal role in driving team engagement and ensuring the success of the transformation. Reflect on past projects: how often have initiatives faltered due to a lack of leadership communication and support? Engaging leaders from the outset can provide the necessary momentum. It’s essential to share detailed information about the data mesh use cases, explain its significance, and outline what is required from leadership—such as their support, effective communication, and identifying suitable data owners. This phase should occur between presenting the use cases and enabling the teams.

Re-organize data into domains

Managing data can become overwhelming, particularly when data is scattered and unaligned with business needs. To address this, place irrelevant or misaligned data into sub-domains or separate files, making it accessible without cluttering the central search interface. 

Another strategy is to periodically review and remove data objects that haven’t been accessed or used for a set period, such as 180 days. Ensuring that data remains readily available and easily searchable is crucial for maintaining efficiency and effectiveness in your data management strategy.

Conclusion

In conclusion, implementing a data mesh involves significant changes to managing and accessing data, moving from a centralized model to one where individual stakeholders take ownership of data domains. This shift enhances data accessibility and usability but requires careful change management to ensure success. Emphasizing Prosci’s ADKAR model for managing change—Awareness, Desire, Knowledge, Ability, and Reinforcement—along with clear role definitions and leadership involvement is crucial for a smooth transition. As you navigate this transformation, integrate these strategies to foster effective adoption and engagement. 

For further discussion or questions, share your thoughts in the Alation Community, read how Kroger successfully implemented data mesh using Alation, or if you are in the middle of a data mesh discussion, consider talking to an Alation Customer Engagement Manager!