What I Learned from Chapter 4 of '失敗の科学'.
By breaking down a big goal into smaller parts and improving each one step at a time, you can make significant progress.
RCTs (Randomized Controlled Trials) are effective for evaluating objective effects, but their implementation requires high costs and can sometimes be difficult. In such cases, breaking the problem into smaller parts and accumulating marginal gains can lead to better outcomes. Even if each step is small, the cumulative effect can be surprisingly significant.
In GitHub Flow, it is recommended to make many small Pull Requests and deployments. This allows for identifying and improving issues while they are still small. Additionally, by limiting the scope of impact, it can also lower the barrier to deployment. This means that, by adopting GitHub Flow, we are accumulating marginal gains on a daily basis!