Skip to content

Knowledge Retention

Every artifact in the chain is a hedge against the departure of the person who created it. A team that keeps the chain intact has institutional memory. A team that doesn't has individual memory — and individual memory walks out the door.

Owners: Whole team (each role keeps its own artefacts honest) Phase it lives in: Continuous The corpus principle this enacts: The artefacts the chain produces survive the conversation.

Where it lives in the chain

How to do this

The three artefacts that carry the load when someone leaves:

  1. Briefs and ADRs — the why. The decisions, with their rejected options. A new owner reading the ADR understands why this shape — not just what this shape.
  2. Runbooks — the what to do when. The operational memory. Tested as part of the release gate, not buried in Confluence.
  3. The model file — the assumptions, witnessed and not. Status, dates, evidence. The next cycle's discovery starts from this.

Plus: commit messages carry the why into git history; scenario names carry the what was intended into test failures; postmortem records carry the what we learned the hard way into the next postmortem.

What good practice looks like

SymptomChain status
New dev reads the codebase and asks "why is the idempotency key structured this way?" — and the ADR answersHealthy. The understanding survived.
New dev asks the team and the team has to guessGap. The decision was made but not recorded.
Slack thread holds the why and nobody can find itFailure. Knowledge that depends on search is knowledge that has already evaporated.

Slack threads are not knowledge retention. Email is not knowledge retention. The corpus is. When someone leaves, the question is not "who knew this?" The question is "which artefact has it?" If the answer is "nobody wrote it down," that is a chain gap, and the next retrospective owns it.

200apps · How We Work · NWIRE