The measure-change-measure cycle improves the workflows, feedback loops, standards, and tools around code—not just the code itself. It turns recurring friction into small changes whose effects can be checked. section 10.0
Run the cycle on a specific friction point:
- Observe the actual workflow rather than relying on guesses or memory. section 10.1.4
- Measure the problem with a baseline metric that fits the friction point. section 10.1.4
- Make one small, specific change so its effect can be isolated. section 10.1.4
- After one or two sprints, repeat the measurement using the same metric and method. section 10.1.4
- Keep and document the change if it helps without degrading something else; otherwise revert it and test another approach. section 10.1.4
Baseline a release process
Instead of calling releases painful, record the 11 manual steps, average two-hour duration, and three failed releases in six months. That baseline gives a later comparison a specific measure. section 10.1.4
Table 10.1 Mindset versus method
|
Characteristic
|
Mindset team
|
Method team
|
|
Frequency
|
Improves when something breaks
|
Improves every sprint
|
|
Scope
|
Reacts to the loudest pain
|
Selects based on measurement
|
|
Tracking
|
“We fixed some stuff.”
|
Before/after data for each change
|
|
Sustainability
|
Bursts of effort, then drift
|
Steady compounding
|
|
Knowledge
|
Improvements live in one person's head
|
Changes documented in standard work
|
Compare a measured, documented improvement method with an improvement mindset that lacks consistent tracking. section 10.1.4

