I recently came across the Einstellung Effect in a Tim Harford Cautionary Tales podcast episode. The main point of that episode is about AI, and why AI shouldn’t replace human thinking. However, it seemed to me to be equally relevant to the growth and popularity of Agile, and the more recent reports of its death. I’m using a very broad definition of Agile to cover a wide variety of approaches. A very large parasol, rather than an umbrella if you like.
The Einstellung Effect
The above-mentioned podcast gives a good overview of the Einstellung Effect. The Wikipedia page provides a good description as well. From that:
The Einstellung effect occurs when a person is presented with a problem or situation that is similar to problems they have worked through in the past. If the solution (or appropriate behaviour) to the problem/situation has been the same in each past experience, the person will likely provide that same response, without giving the problem too much thought, even though a more appropriate response might be available.
The classic experiment to demonstrate this is the water jar test. Participants were given three jars of different volumes and asked to measure a specific amount of water using those jars. The solution is always the same pattern. For example, given three jars with capacities of 3 units (S), 21 units (M) and 127 units (L), and a required amount of 100 units, the solution is 127 – 21 – (2 x 3). Thus the solution is always the pattern L – M – 2S.
However, after several problems with this pattern, participants were asked to measure an amount for which there was a simpler solution (S + M). For example, given jars with capacities of 3 units (S), 15 units (M) and 39 units (L), the required amount is 18 units. Most participants continued to use the more elaborate solution because they were predisposed to that approach. They would not consider or discover the simpler solution.
Cynefin Dynamics
The Einstellung Effect can help explain one of the common dynamics described by Cynefin. Problems start in the complex domain, where solutions are explored through probing and experimentation. The problem then shifts into the Complicated domain, where the solutions become codified through expertise. Finally, the problem shifts into the Clear domain, where the solution becomes “Best Practice”. At this point, people no longer consider whether the solution is still appropriate for the problem at hand, and the Einstellung Effect kicks in. The Cynefin cliff from Clear into Chaos occurs when the problem changes, there is a better solution, but the predisposition to the initial solution means it continues to be used and fails catastrophically.
Agile Methods
I believe that the above dynamic, and the Einstellung Effect, can be seen in the evolution of Agile methods. Let’s pick on Scrum, although this can be applied to all approaches. The first Scrum teams use the approach to solve a problem in their context. It works, so they try it again. And it continues to produce successful results. Then it gets codified and published, training courses are created and certifications are given out. More success comes. Eventually, people are using Scrum with success all over the world.
But the Einstellung Effect means that often, people are using Scrum – and getting good results – even though there may be better solutions to the same problem. And sometimes people are using Scrum – and getting poor results – because they have lost the ability to come up with new solutions when they need to.
This helps explain why I favour Strategy Deployment. I know I’m equally prone to the Einstellung Effect, but for me, Strategy Deployment can be a mitigation. Treating change as a series of nested hypotheses, with Tactics as experiments, can be a way of minimising the Einstellung Effect, and recognising that there may always be better ways, and we should always consider exploring new approaches.