I once sat in that chair. Eight years ago, at the close of a quarter, I let one detail aid through that should have been stopped. No one threatened me. The sales VP's seal simply sat at the top of my evaluation sheet. What I can do now, from the president's office, is move the machinery first, so that M—two floors below me—does not have to make the same call on the same night. Independence is not held by resolve. The only way is for me to take back the two things that matter: who sets his pay, and who can cut his budget.
Cutting two cords: appraisal and budget
Independence is not a posture. Whether a department head who adjudicates two hundred items a month can say no to the business comes down to two questions. Who decides his appraisal? Who can cut his budget? Eight years ago, both answers were the sales side. So the first thing I moved, on taking office, was those two cords.
I reassigned M's first-line reviewer from the sales VP to the officer for review and governance, and put his final appraisal directly under the president. The sales seal no longer appears anywhere on his standing. His budget I unhooked from revenue and fixed it on a prior-year base. If the eighteen-billion-yen quarterly detail aid falls short, not one yen comes off the salaries of his eighteen reviewers. Severing the price of the work that stops things from the performance of the work that sells them—it took me eight years to do just that.
| Before — held by revenue | Redesigned — placed under the president |
|---|---|
| Appraisal: the sales VP made the first-line review; the end-of-quarter seal set next year's promotion and bonus | First-line review to the governance officer, final review to the president; sales involvement reduced to zero |
| Budget: pegged at 0.3% of revenue; if the eighteen billion fell short, next year's headcount was cut | Fixed allocation on a prior-year base, unhooked from revenue; eighteen reviewers held regardless of performance |
| Reporting line: a single line to the sales officer; a stopped item could be quietly killed midway | Doubled to the president and the audit committee; serious findings go to both at once and cannot vanish midway |
Making the stopped work visible
Cutting is not enough. Stopped work, left alone, enters no one's field of vision. Approved material shows up as revenue; rejected material produces nothing, so it is treated as if it never existed. So I count the stops and put them on the board's table every quarter—not as an information item, but as an agenda item. Unless someone speaks to it and a director asks back, the numbers slide past and disappear.
Monthly stop-rate KPI
Of two hundred items a month: 24 rejected (12%), 60 returned for revision (30%). The first line of the scorecard counts not what was approved but what was stopped.
Standing board agenda
Each quarter, rejection, revision, and serious-finding counts become a standing agenda item. What eight years ago was not even an annual verbal report now sits in the minutes.
Dual reporting line
Serious findings go to the president and the audit committee at once. If one side buries it, the same sheet is already on the other's desk.
Escalation visibility
Three to four escalations reach the president per quarter. Too few suggests burial; too many, operational breakdown. The count itself becomes a health metric.
The morning M could, the night I could not
Last month, a revised detail aid for the same antihypertensive came to the review room at the same quarter's end—the last gate before it went to four hundred MRs. The same call from sales must have rung. Eight years ago I bent to that call. M did not. The difference is not his nerve. At the top of his evaluation sheet there is no longer a sales VP's seal. What sits there is the president's seal—mine, and only mine.
| The night I could not (8 years ago) | The morning M could (last month) |
|---|---|
| Situation: quarter's end, the antihypertensive DA at its eighteen-billion peak; the sales VP came to the review room in person | Situation: the revised DA for the same drug, same quarter's end, same eighteen billion; the same call from sales |
| Sheet: the first-line reviewer was the sales VP; his seal tied directly to my promotion | Sheet: the first-line reviewer is the president; the sales seal appears nowhere |
| Outcome: I softened the efficacy emphasis by one notch and let it through; it nearly drew a PMDA finding later | Outcome: returned for revision on two efficacy deviations; it reached my desk as the quarter's third escalation |
I stopped it. Two reasons, both expressions beyond the approved efficacy. There is no sales seal on my sheet—only the president's. ── M, head of material review and governance
Where the design breaks
Machinery is not omnipotent. By cutting the cords and making the stops visible, I gave M room to say no. But room is not substance. If the figure of twenty-four rejections takes on a life of its own, stopping becomes the goal itself. A department that learns it is judged by revision counts will return, in form only, material that should in truth have passed. As Michael Power wrote in The Audit Society, a verification apparatus slides easily toward producing the appearance of being verified.
So what I must keep in reserve, in the end, is not a number but a stance. On the morning M brought up his third escalation, had I made a sour face, the fourth would never have come. The psychological safety Edmondson describes cannot be written into a clause of the rules. It rests on whether M can believe, each time, that the president welcomes a no. Even after rebuilding the machinery that failed to save me eight years ago, what finally speaks is only this: the call he makes in the morning, and my not crushing it at night.
The Justice Disease IV ── Serving Two Masters ── Map of all 10 episodes
- Vol. 1: The President's Chair
- Vol. 2: The View from Above
- Vol. 3: The One in My Old Chair
- Vol. 4: Two Masters
- Vol. 5: Half the Picture
- Vol. 6: When Business Logic Swallows Review
- Vol. 7 (this one): Designing for "No"
- Vol. 8: Bringing the Unmeasurable into the Boardroom
- Vol. 9: The Machinery of Both
- Vol. 10 (finale): The Everyday Peace of One Who Serves Two Masters
This is as far as power reaches. I tore M's appraisal away from sales, cut his budget from revenue, doubled his reporting line, and put his stops on the board's table. With that, he said the no I could not say eight years ago—on the same drug, at the same quarter's end. But what the machinery protected was not independence itself, only the room to break a silence. What finally speaks is his judgment; all I could do was decide not to crush it at night. To serve two masters is to disconnect, with your own hand, the circuit by which one master silences the other.
- Independence rests on two cords, not resolve: who sets the appraisal and who cuts the budget. M's first-line review moved from the sales VP to the governance officer, with final review under the president.
- The budget shifted from a 0.3%-of-revenue peg to a fixed prior-year allocation, so a shortfall on the eighteen-billion DA costs none of the eighteen reviewers a single yen.
- Because stopped work is invisible, rejections (24 of 200/month, 12%) and revisions (60, 30%) became a standing board agenda item—not a report but an item that must be answered.
- The reporting line was doubled to the president and the audit committee, with serious findings sent to both; the escalation count itself (3–4 per quarter) serves as a health metric.
- Machinery only creates room; the rest depends on the president visibly welcoming a no. If the numbers take on a life of their own, the work decays into revisions in form only (Power; Edmondson).
- Robert Simons, Levers of Control (Harvard Business School Press, 1995) (The distinction between boundary systems and diagnostic control; routing stop-counts to the board as a standing item is a concrete way to wire diagnostic control into the boundary system.)
- Michael Power, The Audit Society: Rituals of Verification (Oxford University Press, 1997) (The risk that a department judged by revision counts decays into revisions in form only; how a verification apparatus drifts toward producing the appearance of verification.)
- Amy C. Edmondson, The Fearless Organization (Wiley, 2018) (Psychological safety that cannot be written into a clause; the theoretical ground for the scene in which a sour face from the president would stop the fourth escalation from ever surfacing.)
- Lynn S. Paine, Managing for Organizational Integrity (Harvard Business Review, 1994) (The compliance-versus-integrity distinction; the source for the conclusion that even when machinery creates room, the outcome rests on M's own judgment.)