Tool 30 of 32 · about an hour

Checkpoint Review

The plan faces its numbers at three set points, on agendas fixed on day 10 so they cannot be dodged later.

At roughly day 40, 70 and 100, on agendas set in writing on day 10, in a fixed order that starts with numbers.

What it is for

To make the plan face its numbers at three set points, on an agenda that cannot be renegotiated when the numbers look bad, so anything off track is said out loud instead of quietly slipping.

The failure it catches

Plans stop facing their numbers the moment execution starts. Review meetings drift into progress narrative, agendas get renegotiated when the numbers look bad, and anything off track slips quietly instead of being said out loud.

Why it works

A written agenda, numbers before narrative, shield first then re-measure then read-back then the stuck branch, is the only structure that survives week six. Dates set on day 10 cannot be dodged later without the dodge itself becoming a finding.

How to run it

  1. Set the three agendas on day 10, in writing.
  2. Every checkpoint, same order. The shield numbers first, on screen, before anyone narrates. Then the Baseline Re-Measure. Then the Ledger Read-Back. Then anything off track, said out loud and handled in the room, including any change that arrives via the Trade-Off Price Tag.
  3. A checkpoint happened if and only if all three number segments ran against written numbers. A checkpoint that skipped a segment, or moved, or ran on narrative, is logged as missed. A missed checkpoint is itself a finding about the company.
  4. When the plan is stuck, type the stall, never narrate it. Four kinds. A depth problem: the fix was sized wrong and the trace conversation reopens. A signature problem: a committed decision sits unsigned or overridden, named plainly. An execution capacity problem: the plan is right and the hands are short. And a diagnosis problem, claimed only by elimination, when the commitments were honored, the moves were completed, and the numbers still did not respond past the shield's stated patience window. That last one means the diagnosis itself is the suspect, and the honest move is to re-examine it with the same evidence discipline that produced it, rather than defend it.
  5. Flat inside the patience window is the budget holding, not the plan failing.

How to read the result

The checkpoint record: numbers next to their baselines, the branch taken with its stall type where stuck, adjustments decided, tags accepted. At day 100 the standard is not whether something launched, but whether the numbers everyone agreed to moved, measured, including wherever they are flat.

How to run it well
  • Hold the order gently and say why. Context is coming. Numbers first. It keeps all of us honest.
  • One reschedule for cause is schedule. A pattern of them is the drift this tool exists to catch.
Where it breaks
  • Run as a status theatre it measures nothing. The fixed order, numbers before narrative, is the defense.
  • Rescheduling under pressure is how the drift starts. The day-10 written agenda is the anchor.
  • Softening a signature stall into a hands-ran-out stall to keep the room comfortable. Hands that ran out is only ever written when the contract was honored first.
  • Dodging a diagnosis stall to protect the original diagnosis. When the artifacts show the contract honored, the moves completed and the vitals flat past the patience window, type it.
What would prove it wrong

Checkpoints that all ran clean whose plans still drifted silently would mean the structure misses the drift channel.

What we watch

Whether companies whose checkpoints actually ran against written numbers keep the plan honest, or whether silent drift still finds a channel the agenda does not cover.

Ask yourself
  • Are the three checkpoint dates already written down, from the day the plan was made?
  • At the last review, did the numbers go on screen before anyone narrated?
  • If the plan is stuck, which of the four stalls is it: depth, signature, capacity, or the diagnosis itself?
  • Were any vitals omitted because they were flat?
  • At day 100, did the numbers everyone agreed to move, including wherever they were flat?

Fed by: Baseline Vitals, Savability Test, Trade-Off Price Tag, Leading-Indicator Shield, Baseline Re-Measure, Ledger Read-Back

Get the toolkit