Metrics · RGTM · R

Reading

What share of the kept record actually gets read?

Reading is the share of a window's sessions that were read after the fact — reopened by their author, cited in a coaching conversation, or opened by a teammate the author shared them with. Keeping a record and reading it are different things; Reading measures whether the second codebase actually gets used.

What it reads

  • Session records (PSF) for the team and window
  • Read events: an author reopening a session, a shared session opened by its audience, a coach conversation citing one
  • Sharing state, so only sanctioned reads count — the four safeguards decide who has the right

How it’s computed

Take a team's sessions for a trailing window and check, for each, whether anyone with the right to read it actually did within fourteen days of its close.

// window: trailing 30 days · scope: one team const sessions = psf.load({ team, window }); const readBack = sessions.filter(s => reads.of(s, { within: days(14) }) // author reopen · shared-audience open .some(r => r.sanctioned) // the four safeguards decide who counts ).length; const reading = readBack / sessions.length; // report with n and the trend

How to read it

  • A low Reading with healthy volume means the record is write-only: custody is working, the loop is not. Coaching and retros that cite sessions are what move it.
  • Look at the mix. A Reading made entirely of author-reopens is custody without conversation; shared-audience reads are the sign craft is traveling.

How it gets misused

  • Reading counts sanctioned reads only. A read that bypasses the author's consent isn't a metric, it's an incident.
  • Computed per person it becomes a surveillance score, which the house rules prohibit. Reading describes a team's habit.