On March 15 we conducted another workshop in our Community of Practice series. The topic of this workshop was data governance, and the goal was to provide an introduction to more expansive thinking on data governance and unpack how foundational this issue is to a more socially equitable, inclusive, and sustainable digital strategy. Through our work with local governments and their digital transformation projects over the last years, Open North has developed a stance on data governance that places it at the core of such a strategy.
This more systems-focused and ethics-based approach does not come from Open North alone, and to underscore the diversity of arguments and champions for this way of thinking, we invited our European partners Cities Coalition for Digital Rights (CC4DR) and Open and Agile Smart Cities (OASC) to keynote the workshop. Both organizations are doing vitally important work to, respectively, establish digital human rights as a common basis for digital governance, and build common standards to enable easy interoperability and thus scaling across transformation projects. Although quite different in emphasis, the aim is the same: a digital transformation that is inclusive and beneficial to all.
To their important approaches we added an introduction to our own contribution to better data governance: a data governance development framework that guides the user through an iterative and modular process. The iterative and modular structure is key because our experience shows a gulf between abstract principles and practical realities of local government. Our framework enables users to adapt the process to their specific contextual needs and work through this context to establish a comprehensive data governance framework.
After these talks, we moved to the usual community of practice discussions in which all participants discussed their progress, goals, and barriers around data governance. The rest of this report summarises the points we heard made, the common themes and trends, and then provides a set of takeaways. Once we have completed our in-depth, one-on-one data governance assessment sessions we will publish a longer analysis report on our findings and conclusion from the whole project!
As always, if you work for a local government and would like to take a 1:1 data governance assessment deep dive, please don’t hesitate to reach out to us at csn@opennorth.ca.
Community of Practice Insights
Articles about data governance tend, on the whole, to be either wholly theoretical and normative — i.e. they develop a set of criteria and recommendations for what constitutes ‘good’ data governance — or they are empirical and evaluative, and assess a series of examples or case studies against a normative framework and draw a set of conclusions about how data governance is progressing. Both types are very important, and there are excellent examples in the literature, to which Open North has also contributed.
This blog’s discussion includes components of both types as Open North has a strong normative stance on data governance, and we are talking about the examples of the participants in the workshop. However, the aim here is less to judge against a set of best practices, but rather to tease out a few less-written about but nonetheless important issues in the uptake and development of good data governance. That focus on practical issues is, we think, a missing component in the current data governance discourse, and one whose absence is markedly impeding the adoption of better frameworks and practices. This blog is a first step to sketch out some of the main issues we’re seeing, and a component in a larger project to identify and help solve the problem.
Strategies
The first rounds of conversation centered on establishing a picture of participants’ self-assessed data governance maturity. Participants described the steps they had undertaken so far and what they hoped to achieve. The loose definitions and freedom to frame the issue as they pleased gave space to look broadly at the kinds of developmental trajectories and policy conceptualizations with which local governments operate.
What we heard was a frequent articulation of their progress in terms of a linear development through stages. This is not particularly surprising — journey metaphors are common tropes — but it does raise interesting questions for how we think about developing data governance. For example, several participants said their government had developed a relatively robust data maturity state in that they had established architecture around data standards, quality, storage, security, etc and were now looking for “the next step” into thinking about roles and responsibilities like data owners and data stewards, or designing official data sharing agreements. This makes a great deal of sense, and is a complex undertaking as it involves engaging and aligning a wide range of internal and potentially external stakeholders around socio-political and organizational issues of governance rather than more technical data management procedures.
However, we also heard from a number of smaller governments who had not yet developed such a data maturity level, but were also motivated to start thinking about similar broader governance issues. As the conversation unfolded an unspoken assumption began to manifest that they needed to first build out their data maturity, and then consider issues of governance — but is this correct? Is it a necessarily linear, sequential progression through stages? Or do these ‘less developed’ organisations have a unique opportunity to begin embedding ‘more advanced’ governance ideas in the data maturity steps? This is a core digital strategy challenge, and it lies as the heart of Open North’s work and thinking. Questions about data standards or quality, for example, actually involve far-reaching issues of governance once considered from our more comprehensive view on data governance. You needn’t prescribe a fully developed data governance framework before taking data maturity steps, but such steps are an excellent practical opportunity to start building alignment and buy-in around components of such a framework.
Across all levels of development, participants expressed difficulty strategizing the implementation of public engagement around data governance. This too is a ‘strategy’ problem, albeit of a slightly different nature. The issue wasn’t whether or not there needed to be public engagement — this was largely undisputed — but rather a question of where in the development and the ongoing management process it should be established, and with what goals. Public engagement is expensive, particularly for smaller municipalities, so being able to demonstrate its value as well as the principle is important. Interestingly, it is exactly on the question of values and principles that the public engagement component of data governance development can be particularly impactful: participatory democratic debates are excellent tools for establishing consensus-based digital charters and building social license for data usage.
Alignment
Alignment as a barrier to progress then rapidly dominated the conversations, but it appeared in a number of different forms. It was frequently used to frame the struggle to establish buy-in around different data governance initiatives, and initially generally described lacking cooperation laterally between different units or departments in local governments, or vertically between senior management and other staff. This echoes what Benfeldt, Persson, and Madsen called the data governance “collective action problem.”1 Participants described difficulties with units with independent management and specialized data governance mandates, like the legal department or records management, who were quite advanced in their area but resistant to scaling beyond their remit. They also described struggling to convince senior management of the importance of data governance, and struggling to get other departments or units to prioritize collaboration.
In the conversation we identified several specific causes of misalignment around trust and incentives. Particularly with regards to issues around data architecture, data standards, and data sharing there was insufficient incentive to work to create more alignment and thus functionality. In addition, they also described a lack of trust in the viability, or sustainability, of the development process. Alignment was thus also frequently seen as a buy-in problem: as much a systemic problem of incentives as a more cultural issue of awareness and understanding. As many of the participants said, both can be tackled through better literacy around data governance, and several said that most of their work was actually around educating and developing a common understanding of the issues and goals.
Increasing data governance literacy may, however, prove to be a necessary by insufficient solution in some cases. It can help recognize incentives that already exist, but it cannot create them. This is where the issue of ownership and leadership arose.
Ownership & Leadership
Many participants said that an issue that exacerbated their efforts to create alignment was a lack of leadership. Leadership was seen as lacking on several levels: a number of participants mentioned insufficient regulatory guidance from provincial or federal levels, while others highlighted a lack of an internal position endowed with sufficient mandate or authority to push the development of data governance forward. In its absence, there were people who were unofficially ‘owning’ or championing the project, but this they were undertaking “off the side of their desks” and were thus limited both in terms of time and authority as a result.
In others the lack of leadership was also seen as a lack of guidance around goals, values, or direction. One participant said that although their organisation was quite large and worked with significant amounts of data in a wide range of products, their ability to increase their data maturity and develop more effective governance was severely hamstruck by a lack of leadership around principles and values. Immediate, technical questions around data management could be tackled, but broader issues around governance and direction were stymied. As with alignment above, regulatory compliance is frequently an incentivizing factor, but more comprehensive data governance thinking goes beyond these questions to discuss the values and principles of data usage. On this, perhaps more than anywhere else, strong leadership is key to develop a data charter or similar document outlining an organization’s ethical stance.
This issue of leadership is, as implied above, deeply interwoven with the issues of alignment and strategy. Participants described the difficulty of impressing the importance of leadership on issues like developing a digital charter, in part because it is seen as intangible and thus not obviously connected to more apparent projects and outputs, but in part also because of the strategy problem. Developing values and principles can appear as a nice-to-have, a charter to crown an established digital strategy, but perhaps not a necessary starting point when weighed against more impactful-seeming options. Normative data governance reports tend to provide a checklist of best practices, but often don’t consider the time and resource constraints many local governments have. In the face of those, alignment with leadership on prioritizing developing such a project can be more difficult.
From this intersection of issues presented in this workshop we have developed a few initial takeaways:
Takeaways
- New thinking on digital strategy: Flexible, stepwise development: responsible data governance is not a monolithic checklist, but should instead be approached as a set of separate yet related components. The goal is eventually to check them all, but the order in which they are tackled, and the way in which they are implemented, should be the product of considerations that include the particularities of a project’s real world context. This is why Open North’s ‘Data Governance Self-Evaluation Tool’ and its accompanying frameworks take an iterative and flexible approach to the development and implementation process.
- Tangible focal projects: Data governance, particularly from this comprehensive stance, can easily appear as too large and abstract to garner effective buy-in across an organization’s many units. However, in the same vein that data governance is not a monolith, it also doesn’t have to be the explicit project output in order to be implemented. Smaller, more tangible projects like developing an open data portal, building a data warehouse, or establishing a data sharing project can serve as focal projects in the development of which good data governance objectives can also be achieved. In this strategy, the value of developing a strong open data portal serves as a measure and example for the value of implementing the kinds of responsible data governance measures needed for such a portal. The overall data governance implementation process not only makes progress but also receives concrete buy-in and alignment.
- Leading with public engagement: local governments know that their residents are anxious about the use of data, and they often struggle to develop a clear stance on their principles around data usage. This intersection is an opportunity to empower a citizen assembly-style participatory debate to produce a digital charter, involve the residents in developing a data governance framework, provide democratic leadership and social license on values and principles, and establish a project around which leadership in government can focus.
- However, in all of these points the initial champions still often lack the resources to get the ball rolling, develop the strategy, and undertake the initial steps. This is where an organization like Open North operates to hold those multiple project strands, and implement them effectively in the particular context of the project. We can provide a one on one, in-depth assessment session to map out those first steps and begin the process. If you are interested and/or have questions, please contact us at info@opennnorth.ca
- Olivia Benfeldt, John Stouby Persson, and Sabine Madsen, “Data Governance as a Collective Action Problem,” Information Systems Frontiers 22, no. 2 (April 2020): 299–313, https://doi.org/10.1007/s10796-019-09923-z. ↩︎