Group & Meeting Decisions · Module 4 of 4
The Decision That Didn't Stick
The meeting ran forty minutes and it was one of the good ones. Retire Cranfield, the eleven-year-old provisioning system, and move everything onto the new platform by end of quarter. Corinne argued for keeping it another year and argued it well, was heard properly at length, and did not get her way. Oksana said her team could take it. At 3:38 you said “so we're agreed,” eight people said yes, and everyone stood up. That is the last thing that happened. Twelve weeks later nobody has reversed the decision and nobody has objected to it — it simply stopped being true. Six things can be attached to a decision in the minutes after it is made, and three of them exist only while the room is still in the room: an owner, a date, and the condition that would re-open it. Making all three firmly costs 7 and you have 5; the other three cost 12 firmly and you have 9. Neither window can hold its own contents, so there was never a version of this where you did all six properly. And the scoring is a chain rather than a sum: five threats arrive in order, the decision dies at the first one it cannot survive, and everything positioned behind that gap is worth nothing at all. A named owner, a specific date, a visible first step and a full record is the complete defence against drift, costs 13 of your 14 points, and scores 67 — because it never bought the three-point revisit condition, and week 8 came first. The best plan on the board costs 9 and scores 151, and it dies too, in week 12. No plan of the 529 survives all five threats. You were never choosing whether to leave something undefended; you were choosing which, and the only real failure available was doing it without noticing. Nobody in this module behaves badly, and that is what makes it hard.