According to Terhi Aho, experimentation is a way of thinking that guides action. By experimenting we can develop ways of working without a major change process. It can help software teams to solve problems in small steps, relieve their workload, and foster self-management.
Terhi Aho gave a workshop about using experimentation to improve team performance at Agile Tampere.
We can try changing one small thing at a time, assess the benefits, and decide either to continue in a new way or to try another option, as Aho explained:
There is only a little resistance to change when the incremental changes to policies and ways of working are very small. In addition, the experimental process allows people to participate and give feedback. They themselves have a say in changing the way things are done.
Aho mentioned that this kind of progress requires patience on the part of management, as the initial results are small, even if they are regular. The process of experimentation fosters self-management, and gradually continuous renewal becomes the new normal, she added.
Important development actions may have been postponed when the effort required was seen as too great to fit into the current process. After people have realised the logic in breaking down work, they become able to identify small steps that can be taken to solve the problem without adding to the existing burden, Aho said. Solutions derived from those small steps can bring significant relief to the current workload.
Aho provided a guideline to make experiments small, quick and cheap:
Think about how you can try something:
- as small as possible
- in as little time as possible
- and to minimise the cost of the experiment
Aho mentioned that often experiments become bloated, and cease to be experiments. The idea of an experiment is to test an idea as quickly as possible to see if the hypothesis holds and to learn more. This learning can be used to modify the solution and try it again:
When you want to test the functionality of a new solution, start by thinking about what the most important part of the solution is. Then think about how to test that part.
Likely, the idea of experimentation is still quite broad at this stage. Aho mentioned that the next step is to start narrowing down the time frame. How can you test the solution if you have one month? What if you only have a week? What could be tried if only a day is available for experimentation? Limiting time forces you to reduce what you plan to do. A similar limitation can be made in terms of cost.
When coaching people to change their way of working, Aho suggested asking people what change they can try that will take them 15 minutes over the next week. An experiment this small is so small that it doesn’t get in the way of the rush. No one can say they don’t have 15 minutes a week to experiment, she concluded.
InfoQ interviewed Terhi Aho about experimenting to improve team performance.
InfoQ: What results can experiments bring, and why should we do them?
Terhi Aho: A few years ago, I was responsible for building a new function and team. We built the processes closely as a team, stopping once a week to assess our progress and agree on what to experiment with the following week. We also reflected weekly on our own work as a team.
During the first months, my role as manager was essential in driving the actions forward. Most of the suggestions for experiments came from me. However, after two months, the psychological safety became strong enough for the team members to have the courage to suggest changes to my activities as a line manager as well. Gradually, as the months went by and experience gained, the team members raised more and more ideas for improvement at the weekly meetings. They also actively offered to contribute.
InfoQ: How did the interaction between you and the team change over time and what was the cause?
Aho: After about six months, I found that thanks to their self-direction, I no longer had to spend so much time identifying opportunities and risks. The team members recognised them as part of their daily work and they actively shared them. Ideas for improving the process emerged from all members of the team, so the range of ideas was significantly more diverse than at the beginning.
The attitude towards the results of the experiments had also changed over the months to one of learning by doing. In the beginning, when the ideas came from me, there might be a comment in the evaluation that the idea had been a bad one. As ideas started to come from everyone in the team, the evaluation stopped focusing on the goodness or badness of the idea, but on what had been learned as a result of the experiment and how it could be used in the future.
InfoQ: What if an experiment fails?
Aho: In fact, most experiments do not lead to the expected outcome. This is because we are trying something new and we don’t really know what will happen. When the result is something different from what is expected, it is not a failure but an opportunity to learn something new. The main reason to experiment is to learn and through learning to improve and innovate.