Think Systematically

We build scalable processes with prompt feedback mechanisms. We take measures not only to fix defects, but also to prevent them in the future.

Back to the universal principles

Calibration

Under, just right, over.

Situation	Under	Just Right	Over
The same defect twice	Patches it again and does not ask why the first patch failed.	Fixes the customer impact, then adds the one check that would have caught this cause.	Opens a program to prevent every class of defect and the current one stays open.
A fix with no check	Ships the fix and relies on someone remembering next time.	Adds a small check, a test, or a step in the process that runs without heroics.	Adds a review board for a problem a single automated check would catch.
A process nobody can run twice	Depends on one person who knows the trick.	Writes the process so a second person can run it, and watches them do it once.	Documents every branch until nobody will read it, and the hero remains the process.
Feedback that arrives late	Learns about the defect from a customer weeks after the next batch has shipped.	Puts a signal where the work happens, so the next run can change.	Builds a dashboard nobody checks and calls the mechanism done.
Prevention versus the fix	Only prevents, and the people already hit by the defect are still stuck.	Fixes the people who are stuck now, then spends the next slice on the cause.	Only fixes, and schedules the same incident for next month.
A mechanism that becomes the job	Has no mechanism, only a promise to be more careful.	Adds the smallest mechanism that changes the next occurrence, and leaves it alone if it works.	Keeps adding steps after the defect has stopped, and the process is now the delay.

Further reading

Sources and essays on this principle.

Coupang Leadership Principles
The published wording of this principle.
coupang.jobs