Examining change: Practice as the central unit of enquiry

This article was first published by Chris McDermott on Medium, 23rd Sept 2019

The book Dynamics of Social Practice puts forward the idea that practice is the integration of meaningmaterial and competence and that by using practice as the central unit of enquiry, we can better understand adoption, evolution and continued enactment of practices.

One of the examples used, when looking at the adoption and circulation of practice, is Nordic Walking, “a form of walking involving the use of two specially designed poles to increase the intensity of the exercise”. The book illustrates, in this example by considering the meaning of practice, how to help a practice being adopted in an environment. For Nordic Walking to take hold, walking with sticks had to be dissociated from the idea of frailty. To do this manufactures and others interested looked to ensure that the understood meaning of the practice was “one of personal health, the other of fresh air, nature and outdoor life”.

In this post I’m looking to put forward the idea that by using Social Practice Theory (SPT) as a lens to understand an organisation or a team we can be better informed as to the likelihood of success when we look to make change. I’ll also show how Maturity Maps can help us by representing the environment we are looking to change.

While I’ll focus on the element of meaning both material and competence also are key considerations, topics for future posts.

A word on theories of change

Two suggested approaches to change that I’ve found challenging are the ideas that “by changing the thinking the change in behaviours will follow” and the suggestion that we can “act ourselves into a new way of thinking”. These two distinct change approaches that are often advocated have frustrated me for quite some time. Both approaches suggest that the desired behaviour is already understood by someone in the context and by either thinking or acting the way they do, the desired change will occur.

What I’ve found more helpful when considering change in a complex environment is the idea of introducing, reshaping or removing constraints. The change in constraints acting on the system changes the system’s dispositionality which in turn allows new behaviour to emerge. By using practice as the central unit of enquiry we can be better informed as to what constraint (potentially in the form of a practice) to introduce, reshape or remove. That is, SPT could help us understand how well positioned an environment is to accept and adopt the change in constraints.

Remind me again, what is Social Practice Theory?

As I mentioned above and in a previous post, Elizabeth Shove sees practice as the integration of meaningmaterial and competence. Eg driving as the integration of:

  • Meaning: your need for transportation, to move things or for leisure in the form of racing, a Sunday drive etc
  • Material: the vehicle you are driving, the road on which you‘re driving etc
  • Competence: your ability to observe, steer, make decisions based on your observations etc

Examining Change

The classic organistional and culture change story is that of NUMMI. This is the story of a collaboration between Toyota and General Motors. The collaboration resulted in a workforce, that previously produced, arguably, the worst quality cars in America, turning around to a position where what they produced was among the best quality in America.

GM Fremont

As the story goes, in 1982 GM shut the Fremont plant, laying off all of the workforce, due to the terrible production quality and unreliable production workers. At Fremont workers would drink alcohol, take drugs and frequently so few of them turned up that the production line could not be started. When the plant closed there were 700 outstanding grievances, 25% absenteeism and the fewest number of cars per worker across the whole of GM. If we were to use a Maturity Map to examine the practice within GM Fremont it might look something like this.

The map, which is purely my unvalidated assumptions of both a chain of ‘work to be done’* and the maturity of those items on the chain, suggests that the underlying practices in place were at a low level of maturity. It also suggests that improvement practices may well have been conceptual, in that the meaningmaterial and competence for improvement may have existed, but they were not being integrated to create the necessary practice. What we can also see is that the underlying meaning, and therefore influence, that would encourage or discourage improvement of practices was of quantity over quality and a low trust in the workforce.


When Toyota and GM were looking for a home for the NUMMI venture they decided to use the empty plant at Fremont. It also turned out that 85% of the workforce they hired were former employees at GM Fremont. In stark contrast to GM Fremont, two years after NUMMI was created, there were 30 grievances, 2.5% absenteeism, productivity double the previous high, 90% of employees reporting high job satisfaction as well as quality that was comparable to Toyota factories in Japan.

How was this so? What change had taken place between GM Fremont and NUMMI that enabled these conditions to emerge?

There were two considerable differences, two notable constraints introduced that weren’t present in GM Fremont. The first was that the workforce were asked by management, “how could you make this better?”. In their time at GM this was not a question that they were asked. Improvement wasn’t the job of those manufacturing the vehicles.

The second was the practice, that John Shook argues in his paper “How to change a culture: Lessons from NUMMI” changed the culture, called “stop the line”. Stop the line was implemented using a tool called the Andon Cord. This was a cord that workers could pull if they couldn’t complete their task in the allotted time, or if they became aware of a fault with the car they were building. The result of pulling the cord would mean, initially, support from a manager, and if the issue could still not be resolved the stopping of the production line. This practice was unheard of at GM. GM believed that if workers were given the opportunity to stop the line they would do it all the time in order to break from work.

The Maturity Map for NUMMI might well have looked like this.

So why did this practice take hold in NUMMI? Of course people were trained to act differently, but I believe this doesn’t fully answer why. By examining “stop the line” through a SPT lens we can see:

  • Meaning: quality.
  • Material: the Andon Cord.
  • Competence: ability to identify an issue or know when you were unlikely to complete your task in the allotted time.

The underlying systemic meanings in the way in which Toyota carried out work both in their own plants and at NUMMI was of ‘building quality in’ and ‘respect for people’. Workers were trusted, encouraged and even obliged to pull the cord when things were going wrong.

While in Shooks terms, the Andon Cord helped workers “act their way into a new way of thinking” and thus create a culture with high engagement etc, it was a thinking that was embedded within the design of the work. The Andon Cord and the underlying meaning of quality were, in complexity terms, enabling constraints introduced by the NUMMI management to help create the conditions in which a desirable culture could emerge.

In contrast when Andon Cords were subsequently introduced to other GM plants in an attempt to replicate the success of NUMMI, they were not pulled. The underlying meaning of the system did not change, in contrast to the change in meaning which existed at NUMMI. They still operated under conditions where meeting the production targets took precedence over building a quality car.

In a software context…

With practice as the central unit of enquiry we can examine the practices we want to introduce and the interlinked meanings that exist in other embedded practices or systemically within a system. For example, think about introducing the practice of Test Driven Development into a product team environment where testing is the responsibility of the tester. The meaning for product quality is held by a different discipline to the discipline that is being asked to adopt and enact the practice. Even the practice of Unit Testing, integral to TDD, at the time of introduction was understood to be something different by each discipline. The pushback on adoption of this practice is systemic.

Now think how it might differ in an environment where product teams are encouraged and supported to do Continuous Deployment. Given this approach encourages teams to make software available quickly to end users, the meaning of ‘quality’ is then more likely to be systemic. It’s is, IME, more likely to be spread across the team and the adoption of practices like TDD are potentially easier.

Another example to consider is that of introducing trunk based development to a collocated product development team when the version control tool of choice is Git. Git was designed to make branching easy for distributed teams working at different times. In this instance it’s the meaning embedded in the toolset that makes adoption more challenging. How might introducing a constraint that removes the ability for teams to branch change practices used to manage integration?

Practice as the central unit of enquiry

Therefore using the framing of SPT as the central unit of enquiry we can ask ourselves: is the meaning required for a practice to be successfully integrated into an environment present? If not then what action can we take?

  • Do we seek another practice to reach the same outcome?
  • Do we look to create the conditions that could establish meaning that is coherent with the practice we are looking to introduce?
  • Do we consider positioning the meaning of the practice in line with meanings already present within the system?
  • Or, like NUMMI do we create a trophic cascade by replacing senior management? 😉

What’s next?

Take time to draw a Maturity Map that reflects the practices in your team or organisation. What improvement in practice do you think you need to make to better serve your customers needs? While doing that consider the underlying meanings of each practice, how they are interrelated and what might encourage or discourage improvement of these practices. How does this inform your thinking about what constraints to introduce, reshape or remove?

Hopefully this has given you some food for thought.

Thank you

*The term ‘work to be done’ was suggested by Marc Burgauer as an improved way of referencing meaning in practice as articulated in the post Mapping Meaning.

Many thanks to Valerie McLeanChris Downey and Marc Burgauer for review and feedback. Thanks also to Jabe Bloom for pointing me towards the Social Practice Theory rabbit hole.