Every executive in every industry knows data is important. It is central to digital transformation and the key to beating out competitors. Without data, there are no analytics, so no ability to find new sources of revenue. It’s not even possible to do the basics of running a business without data. However, for data to fuel organizational initiatives it must be readily available, of high quality, relevant, secure, and operate within ethical guidelines. Good data governance ensures data has these attributes, and it is only with these attributes that data can create value.
Most governance programs today are ineffective or, worse, non-existent. Originally, data governance was largely focused on regulations and standards and addressed issues such as the definition of data, internal data ownership, quality control, and the establishment of internal rules for common use, which were often oriented toward informational use cases. The rise of data product management means a new approach is required to govern data through its entire life cycle and in a distributed, federated fashion. We call this approach data product governance. Whether you are new to data governance or want to build on an existing practice you need to incorporate data products into your governance program. Here are 7 practical steps for attaining data product governance.
Step 1: Select a domain
Don’t try to cover everything at once—start small by selecting a domain. A big-bang approach can eliminate some of the reworking that occurs, but it’s often not aligned with business use cases and therefore fails to support end users’ specific needs. End users often struggle to confirm that the data products provide the necessary level of governance and quality and this can lead to time wastage. Keep in mind that introducing data products has the goal of federating data governance, giving domains more freedom while providing general, applicable rules and regulations. Choosing the right domain to begin with is critical. It also requires senior leadership support from the respective domain.
Step 2: Derive a use case
Be very specific by focusing on a business use case. In this practical example, define key terms such as “data product lines” and “data products.” Derive the use case from the business and data strategy. A successful data product governance initiative starts when organizations leverage a well-crafted data and analytics strategy that reflects broader strategic corporate goals. This requires identifying the desired business outcomes and their relative priorities. Effective stakeholder engagement requires data and analytics leaders to deconstruct the business problem and decisions underpinning the data and supporting analytics.
Step 3: Link the use case to data product management steps
Instituting a model that describes how various data products relate to specific business purposes is vital. Without a model there is no ability to ascertain the relationship between products, and without this ability the data is of limited use. In short, we are moving from a use case to the semantics of it. By mapping the information supply chain, data and analytics, leaders can better communicate the utility and value of data and who is accountable for it. Here, it is important to consider the entire data product life cycle—start with “identifying data sources” and end with “retirement of data products.”
Step 4: Assign roles and responsibilities
Define, along with the data product management steps of your use case, who will be responsible, accountable, consulted, and informed. Defining roles and the decision-making rights associated with these roles is essential to the success of your data product governance program. It ensures dedicated management and the establishment of standards and best practices, performance tracking, and quality assurance. Consider the best practice data governance components of: a central office; governance roles distributed by domains; and governance mechanisms (data council).
Step 5: Define critical technology capabilities
Select data product governance critical capabilities to optimize technology investment. Assessing your current data governance capabilities against your target picture for data product governance allows the identification of gaps, overlaps, and insufficient support. For critical data product governance, in particular, there should be investment in automation for repetitive tasks and use of AI/ML for recommendations and an improved self-service experience. The application of AI and ML is key to augmented data governance as it aids operation of data management tasks and optimization of configuration, security, and performance.
Step 6: Leverage policies, standards, and best practices
Determine essential principles, policies, and practices to drive continuous improvement as your organization executes its data and analytics strategy. It is important to find the right balance between general applicable regulations for the most important aspects and decentralized and federated management of governance by data domains. The policies that are most applicable to you depend on the business driver or the outcomes you are aiming for. Make a clear commitment to moving to a governance model for the entire data product life cycle. Such models require a distribution of data domains governed in a decentralized manner, a product-centric approach to both data domains and data solutions, and governance principles implemented on a common data platform or infrastructure used throughout the organization.
Step 7: Roll out and expand
Develop a road map for the transition to holistic data product governance. Address the points above with a holistic view as you add domains and use cases. Embrace the fact that data product governance is an ongoing task and that your concepts need to evolve. Expand your model as new use cases add consumption types like digital applications, advanced analytics, reporting, external data sharing, and discovery sandboxes. Implementing an agile mindset will facilitate the transition, and especially the creation, of the figure of the product owner. Prioritizing data-driven transformation means ensuring a move beyond the proof-of-concept phase to going operational at scale.
