Agile Software Development As Action Research

The heyday of Action Research may have come and gone.  Add it to the list of 60s enthusiasms that faded with time—alongside psychedelics, self-actualization movement, and free love.  Action research has had its lasting effects but it’s faded from our contemporary moment.  Indeed, intellectual history hasn’t quite known what to do with Kurt Lewin, action research’s founding father.  Like Timothy Leary, Margaret Mead, Gregory Bateson, Milton Erickson and others, he doesn’t fit neatly into category, and these figures defined more by their intellectual adventurism than the content of their breakthroughs.

Yet these 60s phenomena live on in a variety of forms: research is demonstrating the power of psychedelics to treat PTSD and it’s hard to imagine gay marriage without the sexual revolution. Indeed, action research has always enjoyed its Bethel, Maine home at the National Training Labs, but it’s also found a home in Latin American liberation theology (Paulo Freire added the “participatory”), and education schools have added “Youth” and YPAR informs groundbreaking learning.

But places like the Tavistock Institute that pioneered action research have largely lost the large scale ambition that shaped its early years.  Looking back on his groundbreaking work in coal mines, Eric Trist describes his action research methods quite simply, it wasn’t “a matter of questionnaires or things like that.  It was being there with them, while they were doing their work.”   He reflects that, ““I often used to be told that I invented socio-technical systems.  I said that was not so.  I had found them, down a coal mine.  Being developed by the people who were doing them.”  This captures the essence of action research: rather than a researcher as objective agent, making discoveries by observing the object of his [sic] research, he works in systems to learn about it, embedded in its context.

As Gianpiero Petrolieri writes in Harvard Business Review (Jan 24, 2017), action researchers “followed Kurt Lewin’s admonition that the only way to truly understand social systems is to try to change them. They gained their insights from personally engaging with such systems, not from maintaining a scholarly or clinical remove. They proudly called this “action research” to distinguish it from the more contemplative kind.”

It’s hard to imagine today’s researchers proclaiming their engagement with the systems they study.  Indeed, our double-blind standards move in the opposite direction.  And our evidence-based, data-driven approaches to things like medicine and education treat systems as if they can be divorced from context of time and place, and decontextualized research can make decisions for us, rendering doctors and teachers as algorithmically-driven deliverers of content or treatments, not human learners and healers, in here-and-now relationship with their students and patients.

Strangely, while evidence-based disciplines have moved away from action research methods, one field seems to have made a sharp turn towards action research in the past decade or two: software development, in the form of agile development.  Where old development philosophies emphasized hierarchical conformity to how things should be (precisely designed, perfectly documented, thoroughly negotiated projects), agile development recognizes that this isn’t how processes work in the real world, and develops a method that  embeds learning and adapting into the process.  Not only does this result in better products that can be responsive to user testing and rapid iteration, but it empowers the individual producers to experiment, learn and alter their course along the way.

A 2017 Atlantic article offers a good introduction to Agile, as well as a fascinating account of its origin.  Like action research, which uses “a spiral of steps, each of which is composed of a circle of planning, action, and fact-finding about the result of the action”, Agile works through a series of embedded cycles: daily scrums, biweekly sprints, quarterly project intervals, etc.  Like action research, agile “happens” embedded in communities of practice, where practitioners learn and change based on experiences in their system, with regular opportunities to reflect and fractal structures for the smallest subsystems and the largest organizational units to inter-relate at appropriate scales and intervals, as they change and learn.

The Agile Manifesto—which has ironically remained unchanged for a decade and a half—is beginning to look as dated as any of the 1960’s manifestos, but their relevance and intellectual kinship remain, and agile development is how most every piece of software we currently use is developed.  There’s been a proliferation of the word “agile” to market novel approaches to things like management, sales, HR, etc.  Perhaps the ultimate determination of whether these actually improve functions like management, sales, HR, etc. will be whether they resurrect action research techniques to allow practitioners to learn and improve their processes in real time, and for organizations to adapt to the fractal model of change this engenders.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s