By: Emily Walsh, Director
If you recently read my first post in this series about data governance, you might have realized that your organization needs data governance, but you’re not quite sure how to implement it. In the data governance literature, there are three common approaches for implementing a data governance program: 1) command and control, 2) traditional, or 3) non-invasive. Although this post will briefly cover all three, I’m going to start with the punchline: my favorite approach for nonprofits is the “non-invasive” approach, although some mixing and matching can be helpful as all three approaches have their pros and cons. Read on to learn more!
Command and Control: You Will Do This
In the command and control model of implementing data governance, roles are usually assigned to people in a top-down manner, often as something new or above and beyond their current roles. Processes also tend to be new and pre-determined by leadership. In this model, data governance is all about taking control of business processes and redefining those processes so that they better govern the data. The tone of rollout communications is directive - people are told exactly what to do in an authoritative and clear manner. Finally, the measures of success are often focused on immediate ROI (measuring value), with the expectation that a data governance program will improve efficiencies and save the organization money as a result.
Traditional: You Should Do This
In the traditional model of implementing data governance, team members are identified into governance roles based on their seniority and oversight of systems and data. Their responsibilities are often spelled out in formal data governance policies. There is often an overarching process for how to govern data - a process that is applied to every data-related activity. The tone of rollout communication is that data governance is something that you should do for the benefit of your team and the organization more broadly. Finally, the measure of success is focused on the quality of the data. Organizations measure the quality of their data by benchmarking their data definitions, maintenance, and usage processes before they launch data governance - and then put metrics in place to measure improvements on these activities.
Non-invasive: You Already Do This!
In the non-invasive model of implementing data governance, you recognize that your data is already being governed, but through informal channels, processes, and approaches. Your team members are positioned into data governance roles based on their existing relationship to the data they work with. These users are then educated in data governance practices. Data governance is an approach that is applied to both new and existing processes, although the processes may not be formally labeled as “data governance” processes. For example, processes for requests for access, creation of a new field, or data maintenance audits may retain those names, which creates familiarity and recognition that you don’t have to call something “governance” for it to be “governance”.
The tone of rollout communication in the “non invasive” approach is that “data governance” is something the team is already doing, but can do better. People are recognized for their existing relationship in governing the data - it’s simply a matter of formalizing the activities that people are already doing in informal ways. Finally, the measure of success is focused on the value achieved through existing and new information-based resources. ROI is typically measured from improved operational efficiency and effectiveness of analytical capabilities brought forth from the governance investments in information technology.
So, which approach is right for you?
The differences between the three approaches are perhaps fairly obvious by now. Under the “command and control” model, implementing data governance can often be done quickly and efficiently, and some organizations may benefit from the clear and authoritative direction that comes with this approach. A downside to the “command and control” approach is that it does not leave time for leaders to build consensus or achieve buy-in, which can put your team’s overall adoption of data governance at risk.
On the flip side, in the “non-invasive” approach the governance roles are defined from within the existing organizational structure - your team members are recognized for already doing some form of governance, which can elicit positive feelings of recognition and value in work that has been done (rather than a sense of criticism over what has not been done). This approach, however, takes time. It is the “slow evolution” into more formalized data governance - it builds goodwill and eases the challenges with change-aversion, but it can take time to implement and see changes that feel significant within the organization. Additionally, this approach might feel too informal for some organizations and could leave too much control in the hands of people who (although already governing the data), have perhaps not governed it well.
Finally, the “traditional”model (perhaps unsurprisingly), falls somewhere between “command and control” and “non-invasive”. In part, this is why the “traditional” model tends to be one of the most popular approaches to data governance.
In my experience, it’s easier to bring people along and “normalize'' data governance within your organization when it’s woven throughout your existing people, teams, and structures via the non-invasive approach. But each organization is different. Ultimately your organizational culture, business norms, and the urgency of existing data governance challenges can help determine which approach is best for you and your team.
So, which of these three models do you think will work best in your organization? Please contact us if you need advice or would like to learn more about setting up a data governance program!