Before a reorg, count what each person loses

Reorgs rarely fail because the new structure is wrong. They fail because nobody counted what each person would lose. Every line you move on the org chart takes something from someone: a reporting line, a budget, a seat in a meeting, a title, the next step up. Count that before you announce anything.

Where do reorganizations actually break?

In execution, mostly. McKinsey surveyed 2,063 executives in September 2013 and published the results in July 2014. Of the respondents personally involved in a redesign, only 21% called it a success. Among companies whose redesigns were completed but failed, 52% cited active employee resistance as a challenge. Among the successful ones, 19% did.

That's a 33-point gap. Resistance usually has a reason, and the reason is usually a loss.

Leaders look at a reorg and ask what the new structure does better. The people inside it ask what they no longer have. Those two ledgers don't balance by default. In Tversky and Kahneman's 1992 paper in the Journal of Risk and Uncertainty, the median loss aversion coefficient was 2.25: a loss weighed roughly twice as much as a gain of the same size. Those were lab gambles over money, and applying them to org charts is my analogy. But it matches how reorgs tend to feel from the inside. Give someone one new thing and take away one old thing, and the old thing is what they remember.

The loss ledger

Before you redraw anything, write one line for each affected person and mark which of these five they lose.

Power. Reporting line, approval rights, headcount. Merge two departments and two department heads become one head and one deputy. Pay may not change. What the deputy can decide just halved.

Information. The one everyone forgets. Change a reporting line and the person drops out of a weekly meeting, off a mailing list. They go from first to know to last to know about customers and HQ politics. A lot of reorg anger starts with "why was I the last to hear?"

Resources. Budget, people, tools, vendor relationships. Centralizing a function at HQ is the clearest case. The money and people leave together. The work stays.

Standing. Title, business card, role in front of clients and the team. Foreign executives tend to underweight this one in Japan. When "Bucho" on a business card becomes "Senior Manager," that person explains the change to dozens of clients, and every client notices.

Future. The visible path to the next role. Flatten one layer and the seat they were going to move into is gone. Nobody raises this at the town hall. It's also the loss most likely to show up six months later as a resignation.

Once the ledger is done, you'll usually find more losers than the names that disappear from the chart. The person who stays on as head of the merged department looks like the winner. Yesterday's peers are now their reports, and their favor ledger from the old department has to be rebuilt. I wrote about the relationships a chart doesn't show in The people who aren't on the org chart.

Japanese companies move people all the time. Why do reorgs still hurt?

Because what Japanese companies are used to is moving people, while the structure stays put. Many run regular transfers every April and October: an informal notice first, then the formal order. People switch departments and everyone treats it as routine. The seats, the reporting lines and the approval rights stay where they were. A reorg moves exactly those.

So resistance in Japan has its own shape. Almost nobody objects to your face. Everyone nods in the meeting. Then handovers slow down, information stops flowing, and replies to the new department head's emails come a beat late. A department is a "ba," a place people belong to. Dissolve the place and the sense of belonging scatters with it. It's the same root as Why global matrix structures break in Japan.

One legal point is enough here. Transfers in Japan have limits. On April 26, 2024, Japan's Supreme Court ruled that where employer and employee agreed to limit the employee's job type, the employer cannot move them to a different job type without their individual consent. Since April 2024, employment contracts must also state the scope of possible changes to work location and duties. If your reorg changes someone's job type or location, read their original contract first. More in Can You Refuse a Transfer in Japan?

The order: talk, announce, then switch the permissions

Get the order wrong and the same plan costs more.

Talk first. Start one on one with the people who lose the most on the ledger. Tell them plainly what they lose, what remains in the new role, and which losses the company plans to make up and which it can't. If it can't, say so. In Japan this conversation does one more job. It gives the person time to work out what to tell their own team and their clients. A department head who learns about their own demotion at the all-hands, at the same moment as their team, loses face in public. That's hard to win back.

Then announce. Spend more of the announcement on why the structure is changing than on how nice the new chart looks.

Then switch permissions. System access, approval routes, budget ownership, mailing lists, all on dates you set in advance, after the announcement. The worst order is the reverse. IT changes the approval flow first, and the person discovers the reorg when their approve button is gone. It sounds like a joke, but it's an easy accident in any company where approval workflows and personnel orders belong to different departments.

This order adds two or three weeks. I think that's cheap. Two or three weeks cost far less than six months of quiet resistance.

The ledger won't always balance. Sometimes the point of a reorg is to take power away from someone. You can't make that up, and you don't intend to. Write their line anyway, and have the conversation before the chart goes up on the intranet. My guess is that people forgive a loss they were told about face to face far more easily than one they discovered from a missing approve button.