Process Library

Engineering Retrospective

A blameless retrospective that converts what went well and what hurt into a small number of owned improvements.

Product & Engineering 6 steps
When to use

At the end of each sprint, or after any significant release or incident.

Trigger
Runs when: End of a sprint, release, or incident

The procedure

6 steps

  1. Set the blameless frame: focus on the system and the process, not on people.
  2. Collect what went well, what did not, and what was confusing from everyone.
  3. Group the themes and pick the two or three with the most leverage.
  4. For each, agree on one concrete improvement with an owner.
  5. Check whether last retro's actions actually happened before adding new ones.
  6. Record the actions where they will be reviewed next time.
Outputs
  • A themed list of what went well and what hurt
  • Two or three owned improvements
  • A check on last retro's actions
Tools
Retro boardIssue tracker
Note: A retro that never checks whether last time's actions happened becomes a venting session. The follow-through is the whole value.
Download .md

Use this SOP in OrgTP

Don't just download it. Drop this SOP onto a seat in OrgTP and every AI agent under that seat inherits and runs it at runtime — with the trigger, steps, outputs, and tools already filled in.

Use in OTP → Browse more SOPs
Related searches
retrospective sop agile retro process sprint retrospective continuous improvement blameless postmortem

Related SOPs

Product & Engineering
Code Review
6 steps
Product & Engineering
Deploy
6 steps
Product & Engineering
Bug Triage
6 steps
Product & Engineering
Sprint Planning
6 steps
← Back to the Process Library