When something goes wrong in production, the first question is usually "who skipped the step?" I have learned to ask a different one: why was skipping it easier than doing it?
People rarely skip the safe way of working because they are careless. They skip it because it is slower, and on a busy day slower loses.
Defaults beat rules
A rule says "always do X". A default makes X happen unless someone works to avoid it. Rules depend on memory and discipline. Defaults just work.
- Tests run on every change without anyone remembering to start them.
- New services start from a template that already has logging, alerts and a health check.
- Access is time-limited by default, so nobody has to remember to remove it later.
Count the steps
A useful exercise: write down every step needed to do the right thing, then every step needed to cut the corner. If the safe path is longer, you already know which one people will take at the end of a long week.
The fix is rarely another checklist. It is usually removing steps until the safe path is the short one.
Make the wrong thing visible
Some shortcuts can't be removed. Those should at least be loud. A deploy that bypasses review can still be allowed, but it should show up somewhere everyone sees, with a name next to it.
Why this matters more as you grow
A small team can hold its rules in its head. A growing one can't. Every new engineer is a fresh chance for a rule to be forgotten, and every good default is one less thing to teach.
I'm Pratyush Shivam, CTO at Konfidio in Berlin, where we build TrustedTrucks. More about my work at pratyushshivam.com.
Top comments (0)