Working together (pair/mob/ensemble) helps with fixing inefficiencies in the system

in a way that it amplifies the pain that every single individual felt when working in isolation, but because everyone was suffering alone, the pain was perceived not as high to address it.

When working together, the pain of that slow deployment pipeline, or flaky integration tests, or slow test suite, or waiting for a PR review, etc. becomes very hard to ignore. The reason is that you now have a group of people waiting, so the perceived cost of the inefficiencies in the process goes up, thus making it more likely to be addressed.

When you address the inefficiency, everyone gets to benefit from it from that point on.

Also, think about how much cumulative value you get because of the improvements in the cycle time not only because you’ve cut the wait times when working together (everyone needed available in the group), but also by cutting the processing/touch time by fixing inefficiencies.

All that translates to shorter feedback loops, less rework, accelerated learning cadence, and value sooner.

Corollary point: If you want to fix problems, get people to experience them as a group working together.