In a recent article, my colleague Kelsey Beyer pointed out that transforming your operating model is just as important as the technology when introducing data mesh in an organization. Although, as Kelsey notes, organizations often sideline making the necessary changes to their operating model because of a perception that doing so is too complex, such changes are, in fact, critical to the success of data mesh. If done correctly, changing the operating model in parallel to the technology can even make it easier to move to the decentralized way of working data mesh requires.
That said, while the aim of data mesh is decentralization, some degree of centralization is still necessary if data mesh is to deliver for an organization. While the delivery of data and data products will sit within the respective domains closest to the source of data, the platform delivery remains centralized — this means the organization can leverage economies of scale through the centralized platform, while also making it easier to create value by decentralizing the delivery of specific use cases.
Striking a balance between decentralization and centralization inevitably comes with a number of challenges. Some of the most typical include changes in ownership and end-to-end responsibilities of data product teams, setting priorities to limit the amount of work-in-progress, and the new client-like relationship data product teams begin to have with internal consumers.
One of the mechanisms that can really help overcome these challenges is a transformation office. A transformation office can help to manage the change program and plays a crucial role in guiding the organization in tackling the challenges mentioned above. It’s a centralized body, set up to coordinate relevant stakeholders. In this blog post I’ll explain some of the things it does and how it helps tackle some of the most common issues that come from a data mesh implementation.
Strategic planning
Every change initiative faces a degree of ambiguity when they begin: what are we trying to achieve? Is data mesh the right solution for our problem? What will our new operating model look like? What kind of governance bodies do we need to apply the principles of domain ownership, data as a product, federated governance and the use of a self-serve platform?
These and more questions need to be dealt with from the first idea to the final implementation. The transformation office can help by facilitating the development of a clear vision and roadmap for the transformation journey with the top management of the organization; it creates alignment with the organization's overall business goals and objectives across all relevant stakeholders. Note that a “goal” describes a business outcome, not a task — so, “Roll out Data Mesh” isn’t a strategic goal; it’s a hypothesis on how to improve a certain aspect of the business and create customer value.
It’s important to note that the vision and roadmap need constant re-alignment and iterative reviews. While goals are stable for a longer time (e.g. more than one year), specific initiatives and priorities may change over the course of the transformation. A transformation office can really help here.
Imagine, for instance, the focus of a given data use case changing from cost savings to revenue generation, based on internal consideration and feedback. The transformation office can update strategic plans as necessary, avoiding the risk — that such strategic shifts are not captured.
Change management
One of the biggest changes when rolling out data mesh is the introduction of long-standing product teams and domain ownership. This means significant change for the way people in an organization work together — those affected have to reorient how they should behave in the new environment and say goodbye to old habits. This is rarely easy.
The transformation office has an important part to play here. It can — and should — design and implement strategies to communicate changes and help the people adapt to them. Just as the shift from projects to products requires you to work in an iterative manner, the same approach should be used in the context of managing change.
Every organization is unique; over time the organization will learn how to best support those involved.
It’s important to note that this isn’t about managing resistance; it’s about providing guidance and removing ambiguity to ultimately make people's lives easier and preparing them for a life of constant iterative change.
Stakeholder engagement
Introducing new ways of working and a change in organizational structure towards data domains and data product teams is only possible if key stakeholders are aligned. However, a change in the operating model will have changes in the realm of control for top and middle management. For instance, domain owners may have increased responsibilities with newly created data product teams in their organization, while a central data team may become smaller. This is not uncommon when data product teams move to the domains and the focus of the central function will move to the platform only. Getting key stakeholders — at all levels of management — involved in this process is therefore just as important as classical change management.
The transformation office engages with key stakeholders at all relevant levels of the organization from the very start to build support, gather feedback and ensure any concerns are addressed. Due to the fact that the transformation touches a large part of the organization, the transformation office needs to engage all layers of management; it can, after all, only be successful if fully supported by the top-management.
One way of tackling this is to find ambassadors for the transformation among stakeholders and build an active community that supports the change from a management perspective. To support this, additional community events and a future governance body should be established to create a formal structure supporting change in the organization. A Steering Committee, involving top management sponsors, will ensure compliance with the organization’s processes and establish a formal setting for decisions.
Upskilling
With new ways of working come new teams and new roles in the organization — data domain owners, data product owners, platform owners and others in both data product teams and a data platform team. New roles may require new skills — not least because members of a data product team should be proficient in working with data and have an understanding of the work of the domain itself. It’s precisely the intersection of domain knowledge and data skills that allow teams to deliver high quality data products for the organization.
The transformation office has an important part to play in upskilling the organization. As the central body in the transformation process, it leverages its position to identify and hone new capabilities and skills required for the new operating model. Its focus should be on organizing training, coaching, community guilds and other development initiatives, such as pairing to match the required skill sets of the future operating model.
Supporting the creation of hiring plans can also be part of the transformation office’s work. The required skill sets may vary from one organization to the other. They depend on the technology the organization is building their data mesh on and the processes that are being established to drive value creation. Developing the right capabilities within the organization is crucial to run data mesh independently.
Performance tracking
With the introduction of long-standing data product teams and decentralizing the approach to data product delivery, teams become independent. This independence will create separate team cultures; in other words, every team will bring its own flavor to the new operating model.
Due to the nature of decentralization we must avoid comparing teams using unsuitable metrics. Use cases will differ a lot in scope and complexity, which means the delivery of use cases will differ, too. To measure success of a use case in the different domains of the data mesh requires the use of different KPIs and measures of success, avoiding the pitfall of comparing the progress of different domains.
Again, the transformation office needs to facilitate the necessary conversation to find common ground. How do we know if the transformation is successful? How do we know if the transformation is on the right path? Being placed at the core of the transformation, the transformation office leverages its overview across all work streams to help establish metrics and KPIs to track the progress and impact of transformation initiatives.
Regularly monitoring performance is key to being able to make adjustments as needed to stay on course. For a start, it’s required to establish a baseline for the KPIs and measures of success; without these it won’t be possible to identify the progress a work stream made. However, it’s not the responsibility of the transformation office to dictate the measures and KPIs to monitor — this responsibility lies with the data product teams themselves.
Project and risk management
At the very core of its tasks, the transformation office oversees the execution of various work streams and projects within the transformation program. It’s a key responsibility to ensure initiatives are delivered on time, aligned with strategic planning and according to agreed quality standards. In the context of data mesh, typical risks can be found in the timely provision of platform capabilities. This can create new dependencies for data product teams, especially in the first months of a data mesh implementation. Other risks may be found around internal processes of the organization for budgeting, compliance and regulatory affairs.
At the same time, there are a number of things a transformation office should NOT do. The project management does NOT intend to create a strict roll-out plan and to enforce operating strictly along this plan. It is important to guide the transformation in an iterative manner. Every step requires a clear review on what went well and what can be improved. This should influence the roadmap and immediate next steps. However, this does NOT mean planning or execution can be sloppy. It is important to encourage a mindset of continuous improvement and learning, fostering innovation and adapting to evolving circumstances to ensure that the organization remains agile and resilient. Another important task is to prevent analysis paralysis. Instead of staying with analytical work for a long time, the transformation office should encourage the sponsors of the transformation to take one step at the time to make things happen. The focus is to support work streams in measuring the outcome of their work, defining individual and iterative roadmaps and helping them to deliver the value they are planning to.
Start small and remember to coordinate
There are several ways this kind of work can be covered in a larger program. Setting up a central transformation office before even starting the transformation is a good practice we have experienced over the past years. You don’t need to start with a large team — start small, find out what the specific goals for all of these streams are and then decide on how many people are required to work on them.
Be aware, though, that running a transformation office requires significant coordination. That doesn’t mean it isn’t worth the effort, of course; it will benefit the data mesh program as a whole and reduce complexity for everyone.
A well run transformation office can communicate a vision, manage expectations and provide transparency. All three items are crucial in a change program. People want to know what to expect and why, they want to know what is happening to their surroundings and they want guidance on how to be a part of it.
If you want to know more about how a transformation office can help your transformation journey, get in touch with us!
