With a lean team, a single bad incident handled badly can consume a week and shake customer trust. A runbook turns the moment of failure from improvisation into procedure.

The lean runbook

  1. Define severity levels. Not everything is a fire. Clear sev tiers decide who gets woken and how fast you respond.
  2. One owner per incident. Even on a tiny team, somebody runs the incident; everyone else supports. Diffuse ownership means slow response.
  3. Communicate before you've fixed it. A status update buys patience. Silence breeds support tickets and churn.
  4. Write the postmortem, blameless. Capture the cause and the prevention so the same incident never costs you twice.

Bottom line: a lean team survives incidents on procedure, not heroics — define severity, assign an owner, communicate early, and learn every time.