Good testability can reduce the development time, and shorten service windows during deployments.
I have learned that good observability can contribute to increased testability in that the live streaming of metrics provides immediate feedback during testing on whether steps in flows succeed or fail.
This in turn speeds up locating the payloads from failed steps, reducing the time to find root causes of failures.
Certain design techniques can also make testing in live systems easier, but this requires added code, which means battling stakeholders who are eager to see initial versions running, and who are less interested in you spending extra time making the solution test-friendly.