Failure as research development in action

Within the phases of an action research cycle, there are multiple points at which failure happens, and where we have opportunities to treat these points as generative learning experiences. To illustrate this, I will discuss an example from the recent ‘Congruence Engine’ project, which sought to use systemic action research to explore approaches to develop a united digital collection as part of the larger ‘Towards a National Collection’ funding programme.¹ From the beginning, Helen Graham, my action research co-facilitator, and I sought to build structures for the project that would allow for a degree of self-organising – recognising that everyone will have a different starting place on the project, have different forms of knowledge and have different motivations for being involved. In doing this we embraced a complexity worldview in the project: one that ‘reminds us of the limits to certainty, [that] emphasises that things are in a continual process of “becoming” and that there is potential for startlingly new futures where what emerges can be unexpected and astonishing’ (Boulton et al. 2015). In understanding that things are in a ‘continual process of becoming’ we immediately open ourselves to seeing everything we do, whether it works or not, as part of that process.

It is difficult to begin from a place of collective understanding in an interdisciplinary project the size of Congruence Engine. The project itself is large and complex, with over fifteen organisations and around sixty people involved to various degrees. We started out by developing a working group model, taking insights from literature on the positives and critiques of Holacracy’s forms of organising (Bernstein et al. 2016; Schell and Bischof 2021), on the role of distributed leadership and open boundaries between inquiry streams in systemic action research (Burns 2007) and on collaborating in cultural heritage (Knudsen and Olesen 2018). We specified that each of these needed a facilitator who was responsible for feeding the group’s work back into central team meetings which, in theory, would work to cohere a sense of direction for the whole project. We dedicated time at the beginning of the project to making space for divergence, as well as congruence, and veered away from articulating collective agreement or understanding as the only way forward for the project. In doing this we were seeking to enable parallel and distributed action that would really help unleash people’s potential to innovate – both practically and conceptually.

From the onset we were upfront that this approach was not static and would probably need to change as the project evolved. This is an admission and an allowance to fail; we knew that things wouldn’t work, but that we had to start somewhere. This approach to iterative development, knowing something won’t work properly at first, is central to an action research approach. It builds failure in as part of the learning process.

Nine months in, after observing and reflecting on how the working groups were developing, we noticed that they were not as dynamic as we had hoped they might be. It is difficult to pinpoint exact reasons for this; there were no doubt multiple factors that inhibited their adoption, from the unfamiliarity of working in less directed and hierarchically structured ways, to the strength of disciplinary silos, but I want to focus on the hybrid nature of our project. Congruence Engine was hybrid from the onset – starting in November 2021, the opening project conference in February 2022 was, for some, the first in-person event of that ilk since the COVID-19 pandemic began. Pre-pandemic, both the project’s action research facilitators had been facilitating in-person inquiry groups, being physically present, noticing the atmosphere of a room, paying attention to body language and to slight intonations in a person’s voice. These were important factors for us in facilitating, improvising and cultivating relationships in action research. Attempting to do this type of work, to manifest energy and dynamism through working groups, is much harder to do digitally, and this is potentially one of the stumbling blocks we encountered. For Congruence Engine, whilst there are many pre-existing connections between some colleagues, the majority of us had never worked together, and come from different disciplines and working cultures. So, whilst we knew parallel work was happening, we realised we needed more active facilitation methods that acknowledged the differing environmental and communicative dynamics that are possible through digital and hybrid working. This was, in a more traditionalist sense, a failure of our first attempt.

Keen to adapt our approach, maintaining the core principles of parallel, distributed action, led by those who the work resonates with, we devised a new structure for the project that built in a directed digitally led engagement with each other, and more facilitated cross-project meetings. Within the context of the action research’s cycle structure, this was a planning forward, based on what we’d learnt during our first phase of the project. Utilising the language of iterative, incremental and adaptive change, the action research facilitators actively sought to communicate that nothing was fixed and that structures could be changed throughout the course of the project without it needing to be framed as a failure of one approach.

The project had been using Basecamp to try and facilitate the working groups in their original instantiation (Figure 9.2), but the platform had not enabled the open and generative form of collaboration we had hoped for.² Research on undertaking action research in online environments points to benefits of rapid and open communication services whilst acknowledging the incompleteness of communicating online, but also to the potential to work both synchronously and asynchronously across digital platforms (Embury 2015). Knowing we had groups of researchers who worked on different days, on different fractional contracts, and with Congruence Engine being a larger and smaller proportion of their focus, encouraged us to focus on developing synchronous and asynchronous modes, opting to introduce Notion to support our work.³ For some, this added yet another digital system into the mix; something new to learn, something else to keep an eye on, but for others, it opened up the project and enabled far more openly and collectively developed ideas to be shared. Alongside the introduction of Notion, we formed a new Action Research and Project Management group and designed a new central meeting to cohere all the investigations, embedding collaborative and live documentation as part of those meetings. One of the the intentions here was to provide a new digital space for synchronous and asynchronous documentation and research development, alongside a more actively facilitated set of meetings to more explicitly encourage sharing across strands of investigation.

This second cycle of the project’s structure and management was no doubt disorientating for those who had adopted and got used to the working groups model, but there was a lot to be learnt about the dynamics of a new interdisciplinary project and the types of structure that were needed to support a wide variety of researchers and practitioners. Whilst the structure of using Notion, supported by two ongoing meetings to actively facilitate the movement of the research inquiries continued throughout the project, the approach continued to be tweaked, acknowledging when a meeting, a form of organising or an approach wasn’t working as intended. Instead of seeing these as a series of failures, we saw them as a step closer to understanding what is necessary for the context of our research.

Writing about action research as a mode for talking about failure

Action research is an inherently pragmatic approach to managing a research process, grounded in learning through practical experimentation. It sees any form of action as deeply contingent on the individual contexts you are operating within, and this means that the strategies put in place for Congruence Engine were a product of the people, institutions and wider political, social and economic realities we found ourselves in. This will be different each time, and therefore the approach will need to be different each time. This orientation towards learning by doing, seeing and observing what works and what doesn’t is an alternative to the negative connotations of failure. It is a generative reframing of failure enabled by the iterative and complexity-embracing nature of action research.

Difficulties arise when we are asked to write about the project methodology and give overviews or accounts of how the project developed. Judi Marshall and Peter Reason argue that it is easy for accounts of action research to become bland and prosaic and urge quality action research to respect the aliveness of the method – its wildness at times (Marshall and Reason 2008). With that in mind, how can we use the opportunity of an action research framing of failure to write productively about what didn’t go to plan?

As a mode of doing research, it does not necessarily align with the same models of success and significance found in traditional forms of academic inquiry. Rather than looking for ‘truth’ and replicability, action research is concerned with engagement, dialogue, pragmatic and locally useful outcomes, and an ‘emergent, reflexive sense of what is important’ (Bradbury and Reason 2001). Here we begin to step away from reportage of what was successful to what the learning was from moments of tension, from changes of process, from what was difficult and impossible to get done. Marshall and Reason sum up the potential of not knowing, or of failing to know, when they note that ‘if you think you know what you are doing as an action researcher, have it comfortably in hand, you are really not doing it, are not on a learning edge’ (Marshall and Reason 2008). Capturing the liveliness of the research that pushes you and your collaborators to the edge of your learning and comfort zone is the sign that significant research and learning experiences are taking place.

In talking about, and writing about Congruence Engine, I have always attempted to be (painfully, at times) self-reflexive and critical about the urge to tell a good story – of glossing over the complexity, the tensions and the failures. There is always an underlying fear that surfacing the challenges and the improvisations you made along the way will eclipse the traditional success stories we wish to tell our funders. However, in using action research as method, in demonstrating the challenges through defining what didn’t work, we have the opportunity to release ourselves from that urge.

Action research is an orientation towards research that undermines the traditionalist sense of failure. Instead of defining what didn’t work as a failure, action researchers see it as a development in their research, a finding that will help us move towards a better understanding of how to create a more effective research intervention during the next cycle of work. Knowing what doesn’t work is as useful as knowing what does work in a complex and changeable environment.

Acknowledgements

My thanks to Helen Graham, my co-facilitator in the action research of the Congruence Engine project, whose insights on the action research methodology and the forms of investigation this chapter is based on, are forever helping me learn.

Notes

1. 1 Towards a National Collections (TaNC) was a five-year, £18.9 million investment by the UK Arts and Humanities Research Council, looking to understand the requirements of developing a united digital collection made up of the cultural heritage materials held in the UK’s museums, libraries, archives and heritage organisations. https://www.nationalcollection.org.uk/Discovery_Projects (accessed 24 November 2024).

2. 2 Basecamp is a project management tool by 37signals. It was used by the Congruence Engine team from March 2022 until the end of the project in January 2025.

3. 3 Notion is a documentation tool and wiki developed by Notion Labs Inc. It was used by the Congruence Engine project from January 2023 until the end of the project in January 2025.

References

• Bernstein, E., J. Bunch, N. Canner and M. Lee. ‘Beyond the Holocracy Hype’, Harvard Business Review (July–August 2016): 38–49.
• Boulton, J.G., P.M. Allen and C. Bowman. Embracing Complexity: Strategic Perspectives for an Age of Turbulence. Oxford University Press, 2015.
• Bradbury, H. and P. Reason. ‘Conclusion: Broadening the Bandwidth of Validity: Issues and Choice-Points for Improving the Quality of Action Research’. In The Handbook of Action Research: Participative Inquiry and Practice, edited by P. Reason and H. Bradbury, 695–707. SAGE Publications, 2001.
• Brydon-Miller, M.. and D. Coghlan. ‘First-, Second- and Third-Person Values-Based Ethics in Educational Action Research: Personal Resonance, Mutual Regard and Social Responsibility’, Educational Action Research 27 (2019): 303–17.
• Burns, D. Systemic Action Research: A Strategy for Whole System Change. Policy, 2007.
• Embury, D. ‘Action Research in an Online World’. In The SAGE Handbook of Action Research, edited by H. Bradbury, 529–35. SAGE Publications Ltd 2015. Accessed 24 November 2024. https://methods.sagepub.com/book/the-sage-handbook-of-action-research-3e.
• Grewcock, D. Doing Museology Differently. Routledge, 2014.
• Knudsen, L.V. and A.R. Olesen. ‘Complexities of Collaborating: Understanding and Managing Difference in Collaborative Design of Museum Communications’. In The Routledge Handbook of Museums, Media and Communication, edited by Kirsten Drotner, Vince Dziekan, Ross Parry and Kim Christian Schrøder, 205–18. Routledge, 2018.
• Marshall, J. and P. Reason. ‘Taking an Attitude of Inquiry’. In Towards Quality Improvement of Action Research: Developing Ethics and Standards, edited by B. Boog et al. Sense Publishers, 2008.
• McCarthy, C. ‘Theorising Museum Practice Through Practice Theory: Museum Studies as Intercultural Practice’. In The Routledge International Handbook of Intercultural Arts Research, edited by P. Burnard, E. Mackinlay and K. Powell, 24–34. Routledge, 2016.
• Reason, P. ‘Pragmatist Philosophy and Action Research: Readings and Conversation with Richard Rorty’, Action Research 1 (2003): 103–23.
• Schell, S. and N. Bischof. ‘Change the Way of Working. Ways into Self-Organization with the Use of Holacracy: An Empirical Investigation’, European Management Review 19 (2021): 123–37.