Facilitating Stakeholder Engagement#
When working with stakeholders beyond your own team, there are many tools and best practices for improving the quality of engagement through activities like mapping relevant project stakeholders, organising stakeholder workshops, establishing shared language and understanding, and creating the foundation for effective collaboration.
This chapter will share guidance and resources for facilitating different forms of stakeholder engagement for data science collaborations drawing upon expertise from the Turing Research Application Management (RAM) team.
Key question: Who is involved in a collaboration, either as a user, impacted group, part of the community of practice, or in some other capacity?#
The goal of stakeholder mapping is to understand the people and organisations involved in a collaboration, and to collect all this information in one place. This helps establish a shared understanding of who is involved in what capacity, which is a useful resource for onboarding new teammates and to make sure everyone is on the same page. This activity is also often a precursor to other engagement activities such as impact assessments, user experience workshops, collaboration cafes and more, as stakeholder maps can help identify who should be centred in the research process. Stakeholder mapping is a core activity of many Research Infrastructure Roles and may serve different goals, for example:
Research Community Managers may use this to establish a project governance structure and decide the most effective methods and channels for communication with each stakeholder.
Research Application Managers may use this to categorise and prioritise different kinds of users of research outputs, such as local government vs. industry partners.
Research Project Managers may use this for project reporting and evaluation purposes and identifying who to invite for a certain meeting.
Stakeholder mapping is not meant to be a single source of truth. It is a tool to capture a team’s understanding of the stakeholder landscape, clarify relationships, and surface groups that may share common characteristics or challenges. A well thought-out, detailed stakeholder map will facilitate bringing a team onto the same page, which can help with developing engagement pipelines.
The Alan Turing Institute, Tools, Practices, and Systems programme stakeholder mapping template
Teams should be encouraged to define their own metrics for mapping stakeholders on a matrix for prioritisation. Choosing the right metrics will help identify the most important stakeholders to engage.
These metrics should be carefully chosen to represent the key stakeholder considerations and project team goals. For example, as part of the mapping, the team may try to:
Establish a community of practice for project governance.
Prioritise communication methods and cadence for different stakeholder groups.
Identify which groups may be negatively impacted by a research output.
Identify which stakeholders are most prepared or willing to adopt a particular research output.
Requirements gathering & alignment#
Turing RAM team requirements gathering questionnaire example for understanding different team approaches to creating case studies.
Often times different people will use the same term to mean completely different things or for different purposes. This exercise can help detangle the nuances of commonly used terms and whether or not there is overlap in the collaborators’ understandings of it, and thus provides an opportunity to build shared understanding and goals.
A good measure of success for requirements gathering & alignment is a glossary of terms and concepts that the project team and additional stakeholders have created together, and can refer back to when confusion or misunderstanding arises. This glossary should be a living and breathing document, that can be updated, and terms redefined, as the team’s knowledge and expertise grows.
It is hard to overstate the importance of having shared language and understanding in a project team - not embedding this at an early stage can allow miscommunications and misinterpretations to snowball into highly incompatible outputs that don’t represent the team, community or userbase!
Organising stakeholder workshops#
Key question: What work before, during, and after a workshop should be done to run an effective stakeholder workshop?#
Workshops are best suited for engagement activites where there is a need to bring people together to deeply engage on an area of interest with the purpose of working on a task, as opposed to gatherings which are purely driven by conversation. This ensures participants are taking active steps towards achieving a shared goal.
Organising an effective workshop is a multi stage process which can be segmented into three stages; pre-workshop, workshop, and post-workshop. We think it’s important to prepare for each stage with equal weight and consideration to ensure the best process for participants throughout.
RAM Stakeholder Workshop Checklist with key steps and considerations the pre-workshop, workshop, and post-workshop stages
Be prepared to rein in conversations which labour a point with no apparent solution
Read around the workshop topic to be at the very least prepared enough to engage with participants at a high or abstract level
Offer opportunities to network and make sure you make every effort to speak to all participants
Strategy alignment workshops#
Using the basic “Why/How/What” format as a starting point for the conversation helps provide focus and clarity and can lead to a better outcome. As with all frameworks, this starting point should be adapted to match the understanding and ways of working of the specific project team.
By not overstructuring the conversation, you will also allow people the space to express their thoughts freely. We feel this is an important approach as it allows open conversations to happen and things to emerge that otherwise might not.
Strong facilitation skills are crucial, including reading the room and identifying patterns in the conversation.
Doing this effectively takes a long time - at least 2 hours per session (we recommend a minimum of 6 hours, if not longer). These sessions should be separated and standalone (for instance, 2 hours slots on different days), but close enough together to ensure continuity- schedule all session well ahead of time so that all team members can attend all sessions.
A synthesized output from these sessions, such as a comprehensive, professional slide deck, can be used both internally and externally by any team member to communicate the project to stakeholders.
Key question: How can research outputs benefit from a user-centric design process?#
Although many research tools are created for a very narrow purpose, we encourage researchers to consider the user experience (UX) throughout the process and to incorporate UX design principles during development to make the tool more accessible and usable for all potential users. Activities such as user journey mapping and user testing can greatly enhance your understanding of who is using the research tool and how they employ the tool in a real-world scenario. By considering potential users of the tool from the early stages of a project, you can include features that will make the tool easier to use and more relevant to potential stakeholders, and therefore more likely to be used and applied in practice.
Existing resources can also benefit greatly from making improvements to the UX by establishing heuristics to define criteria for usability and to identify areas in need of improvement. By utilising exercises and heuristics designed specifically to consider and evaluate the user experience, researchers will examine the assumptions built into the tool and gain a deeper understanding of how it can serve different users types. These insights can help a team to develop a roadmap for the tool to plan out future developments with varied stakeholders in mind, leading to more meaningful stakeholder engagements based around a deeper understanding of how those stakeholders are interacting with and repurposing a research output.
UX and UI design are specialised skills that researchers are not expected to be familiar with, so consult available resources and bring in a professional team when possible!
Look for design firms that focus on coaching and prioritise user accessibility and an open, collaborative process. Working with designers on a fixed project will improve the UX of your tool in the immediate term, but learning these skills as a team will allow you to consider UX and apply design tools at each stage of your project’s lifecycle.
If in doubt or short of time, have a user try out your tool and ask them open ended questions about their experience. You’ll be surprised at how much you can learn from doing this a few times!
Stakeholder impact assessments#
Key question: How can we develop projects with continual sensitivity to the ways in which these impact stakeholders?#
Stakeholder impact assessments (SIAs) are a tool used to gain anticipatory insight into the project’s likely impacts, defined as the possible harms and benefits. SIAs take the form of questionnaires addressing the ethical permissibility, transparency, accountability, and equity of projects. SIAs are used to identify and document the full range of potential impacts and provide project teams with a contextual awareness of the social environments impacted by their projects. This activity can help teams steer the direction of their projects preemptively to ensure that they support the wellbeing and sustainability of the individuals and communities they impact and mitigate any harms.
Evaluations of project impacts should ideally be done in collaboration with stakeholders. In the context of SIAs, stakeholders are defined as individuals and communities that are impacted or may impact projects. Stakeholders that are most vulnerable to potential project harms are considered the most salient. Engaging stakeholders in conducting SIAs helps secure the accuracy and integrity of SIA outcomes. This is because gaps in identity and experience between project teams and impacted individuals may cause there to be differences between how team members view project impacts and how stakeholders experience them. Facilitating proportionate stakeholder engagement and input throughout the AI lifecycle is a way to gain a richer understanding of the impacts that projects may have on stakeholders.
Conducting a SIA is not a one-off activity but an iterative one that occurs at key points throughout the design, development, and deployment stages of the AI project lifecycle, each time informing the direction of the project so that it continually adapts to changing contexts. After each iteration is conducted, project teams are asked to re-assess questions addressed within the Stakeholder Engagement Process. These questions motivate project teams to iteratively re-assess the extent to which their analysis of project stakeholders and reflection of team positionality continues to be accurate and relevant. These re-assessments are used to determine engagement objectives and methods for each following SIA.
A practical template for conducting SIAs can be found in Leslie, David (2019). Understanding artificial intelligence ethics and safety: A guide for the responsible design and implementation of AI systems in the public sector. https://doi.org/10.5281/zenodo.3240529
Ensuring that SIA outcomes accurately reflect the experiences of impacted stakeholders is critical for this tool to achieve its purpose, for this reason, taking a methodological approach to determining proportional levels of engagement is very important. Impact-informed guidance for conducting a Stakeholder Engagement Process (SEP) can be found in the Turing Commons AI Ethics and Governance Skills Track. The SEP is a complementary process that precedes SIAs and provides reflection questions set to identify and analyse project stakeholders, and reflect on team positionality in relation to project stakeholders. These questions support teams in determining proportional engagement objectives and methods for conducting Stakeholder Impact Assessments.
The Turing Commons AI Ethics and Governance Skills Track also provides further guidance on skills for conducting SIAs.