# D15 — Tiny retrospective with incompatible estimates

Mara: 2–4 hours of repeated exchanges per week might be avoidable; based on a two-week message count, not timing. Ivo: 2–3 hours per month may be removable from his and Mara's combined reconstruction; excludes useful reassurance. Oren: 2–3 hours per field visit might go away after a one-hour common handover; Niall: probably 30 minutes. Suri: 30–60 minutes on a genuinely repeated repair; sample of two remembered cases. Cato: five of nine order exchanges may be repetition or correction, but only three were copying mistakes.

These estimates use different denominators and overlapping work. No sum or average is authorised as an agreed benefit. Some may turn out wrong. The first useful change may be discovering what should be measured or whose problem is actually being solved.

People nevertheless agree that a role nobody accepted and a form nobody can correct will make matters worse. That shared observation does not select an organisational form. A less ambitious booking change, keeping the present organisation, declining additional business, or ending an experiment can each be competent responses.

Suggested observation, not an imposed score: record a small number of actual clarification episodes, including time spent entering and reviewing the record, and allow the person who supposedly saved time to disagree.