Instead of a concept paper written once and never touched again: a PDCA cycle that sharpens with every turn. That happens to line up exactly with Rule 2 of our own PM Rulebook — a plan without a feedback loop drifts. A rollout plan as PDCA is the feedback loop.
Click a quarter. Each phase has a clear task and a clear handoff criterion to the next — no phase is "done" until it delivers that.
Each wave runs the wheel once, fully, before the next starts. Small scope per cycle — so "Check" stays honest instead of becoming a formality (Rule 3).
One cycle has already run: rulebook formulated, landing page built, design tested and refined (the Mentalist restyle was "Do→Check→Act" in action — your feedback directly drove the correction).
Plan: select 1–2 real, ongoing projects; fix the success criterion in advance — "does the tool find at least one risk the regular project review would have missed?" Do: run a guided inversion pass with real stakeholders, document the result.
Starts only once Wave 1 has reached at least the "Act" gate — capacity and credibility are limited, don't spread thin in parallel (that would itself be a case of Rule 1: goldplating the rollout plan).
Higher stakes (org-level diagnosis), needs the reference story from Wave 1. Deliberately held back rather than pulled forward — impatience here would be the most expensive mistake in the whole plan.
At the end of each wave, the result decides, not the calendar. Basic rule: if only the metric "tool was used" is measured instead of the real effect, the gate is worthless (Rule 3 — Lem/Goodhart).
Clearly demonstrated value (the criterion fixed in advance is met). Becomes a standard tool, the next wave starts.
Approach right, execution not. Same cycle, tighter focus, a different lever — no new concept paper, just a new plan step.
No demonstrated value despite a fair chance. Documented openly as "tested, discarded" — that too is a result, not a failure.
Not talking more — talking more regularly, in smaller, more honest pieces. Exactly the inversion of the isolation tank (Rule 2 — Lem's isolation-tank test from The Conditioned Reflex, the Pirx cycle): whoever stays silent for a long time and then delivers one giant package triggers overreaction. Every message follows the same three-part structure: fact → context → next step.
| Stakeholder | Medium | Cadence | Purpose | Mode |
|---|---|---|---|---|
| Pilot project team | Short sync (15 min) or voice memo | 2×/week, during the Do phase | Collect friction & raw data immediately, not only at the end | Interactive |
| Sponsor / decision maker | Three-sentence traffic-light update | 1×/week, fixed weekday | Keep progress and blockers visible — honest color even at yellow/red (Rule 5) | Push |
| Early-access list | Short "work in progress" post + status page in the portal | Every 2–3 weeks | Let expectations grow along with reality — real progress and setbacks, no overselling | Push Pull |
| Gate decision (Act/Re-Plan/Kill) | Personal conversation first, then a written summary | Once, at the end of each wave | Opening a kill decision by email is the comfort trap from Rule 5 in its purest form | Interactive |
Why this split and not "everyone gets everything": a sponsor pulled into the details twice a week eventually stops listening — exactly the feedback saturation Rule 2 is meant to prevent. A pilot team that only gets the weekly report can't raise its friction in time. The right medium per role isn't a nice-to-have, it's the difference between "positively present" and "annoying."
Three coordinates form the vector for each stakeholder: influence on rollout success (P), self-interest in the outcome (I), and the actual stature this person carries (S) — on the same scale as P. The dashed arrow on the P/I field isn't a look back, it's the work instruction for the next round of communication. S itself is one-dimensional — a single comparative value, not its own direction — and is made visible in the diagram as a third dimension via circle size: large circle = high stature, small circle = low stature. If a gap opens up between a high position (P) and a small circle (S) — stature that doesn't keep pace with the assigned role — that's a risk in its own right, not a footnote (see the warning below). The precondition for all of this: actually knowing the stakeholders correctly in the first place — see outdated org charts as a time trap.
Circle size = S (status), the third vector dimension alongside P/I · small circle = low stature, large circle = high stature
The vector is what matters. A point on the grid says where a stakeholder stands today. The dashed arrow says where the strategy is meant to steer them — and makes the communication measures from the table above directly checkable: does the measure match the arrow, or are they working against each other?