Prof. Jan Bosch, 07 December 2020
Last week, we organized the sprint 19 reporting workshop of Software Center. The opening keynote by Frances Paulisch (Siemens Healthineers) was about the transition from a transactional business model to continuous value delivery to customers. The closing keynote of Aleksander Fabijan (Microsoft) discussed starting and scaling A/B testing. Though the keynotes focused on very different topics, they shared a common theme.
This theme was concerned with driving change and the challenge of successfully implementing the change. Both Frances and Aleksander raised the point that most changes in organizations touch many individuals, functions and departments, as well as numerous processes and ways of working. As I discussed in earlier posts, in most organizations, to change anything, you have to change everything. However, it’s impossible to change everything completely at the same time. And this is where I see many organizations get stuck.
The way out of this is to accept that all change will need to be gradual and that rather than accomplishing the change in one fell swoop, it will have to be a continuous process for an extended time. The analogy is that of a flywheel. Getting a flywheel from a complete stop to at least some rotation requires quite a bit of energy. Once there’s some motion, you need to keep exerting power to speed up. However, once it achieves more speed, it becomes easier and easier to maintain that speed. The question is how we achieve that in organizations. In my experience, there are three main aspects: showcase the value, increase engagement and build infrastructure.
First, showcasing the value requires finding a small scope where success is virtually guaranteed but where the value of the change you’re looking to accomplish is clearly demonstrated. As you initially have to realize all of this with a small team, it’s important to keep what you’re looking to accomplish as much as possible in your scope of control to minimize the risk of others torpedoing your efforts. For example, when running your first A/B experiment, pick a topic where opinions are highly diverse in the organization, ensure data quality from the A/B experiment and use this to engage with relevant stakeholders to show how the data from the experiment benefits the organization.
Second, once you have a successful case, engage the stakeholders that you need to convince to increase the scope of the change and show the real, concrete, tangible benefits you created in the first loop. Use this to increase engagement with the people that you need in the next iteration of the flywheel to create the next showcase. For example, when upping the release frequency of software on your way to DevOps, initially often the ability to rapidly resolve defects in the field can easily be used as a means to increase awareness and buy-in with relevant stakeholders.
Third, look for ways to automate some of the activities you’ve so far conducted through manual effort so that over time the cost of running through the iterations of the flywheel becomes lower. This is concerned with building the infrastructure for the change you’re looking to realize. For A/B testing, this may mean automating parts of the data collection pipeline and for adopting DevOps, this typically requires automating the CI/CD pipeline, as well as the test infrastructure.
Once you’ve gone through the first iteration of the flywheel, it’s basically rinse and repeat to take the next step and try to accelerate. It’s easy to get discouraged when trying this, but remember that flywheels accelerate very slowly and require a lot of energy to get moving at all. And initially, as everything needs to be done manually, the flywheel has a lot of resistance. With more and more of the infrastructure in place, it rotates easier and easier.
Many of the companies I work with struggle with realizing the changes required in their organization. Some oscillate between trying to realize a big-bang change and a complete deadlock where nothing happens. The most effective way to realize change is the persistent, perhaps slow, but continuous accelerating of the flywheel through showcasing value, engaging stakeholders and building infrastructure. Accept that it takes time and, because of that, start yesterday instead of tomorrow. Build your flywheel and get it spinning!