Every transformation is decided in a moment nobody plans for.
Six weeks after a leadership team announces it will focus on fewer priorities, a senior leader brings forward a pet project that doesn't meet the new criteria. If it gets approved, the program is over. Everyone learns the prioritization was optional, and no town hall reverses that.
Why does one exception carry that much weight? Because employees know that announcing a change costs leadership nothing. A vision deck is cheap. The first costly no is the only evidence that the change is real, so people watch for it and discount everything that comes before it. Harvard Business Review recently ran a piece arguing the same thing about values: a value means nothing until someone pays a real cost for it.
That raises a harder question, though. Who should the no apply to?
Most rules point down
Look at the average values statement or change charter and count who the rules constrain. The guardrails govern frontline behavior. The compliance training goes to the people with the least discretion. The new ways of working describe what teams three levels down will do differently. The people with the most power to kill the change face the fewest written limits on how they use it.
Employees read that imbalance correctly. A rule that only binds the people who were already going to comply is a poster. A rule that binds the person who could override it is a commitment.
I looked for cases where change held, and a pattern emerged across three very different settings. In each one, the rules that made the change believable pointed to one thing. Leaders limited their own power first.
Three leaders who gave up power on purpose
A submarine captain stopped giving orders. David Marquet took command of the USS Santa Fe, a ship he hadn't trained on. During an early drill he ordered a speed setting that didn't exist on that vessel, and the officer on deck passed the order along anyway. When Marquet asked why, the answer was that the captain had told him to. Marquet realized he had a crew trained for compliance and a captain trained for the wrong ship, a combination he describes as deadly. His fix was a limit he placed on himself: he stopped giving orders. He kept only one order, to fire weapons, because he decided responsibility shouldn't pass to anyone else. Officers began stating what they intended to do instead of waiting to be told. The payoff came later, on the surface at night, while waiting to recover a SEAL team, when Marquet gave the wrong command and a junior quartermaster told him directly he was wrong. He was. The crew member who could do that only existed because the captain had given up control first.
A CEO whose "no rules" company wrote its own leadership rules. Ricardo Semler's company, Semco, is famous for having almost no rules, but that description is misleading. The rules existed. They just governed the people at the top. Nobody could become a leader without being interviewed and approved by the people they would directly manage. Every leader got an anonymous evaluation from their people every six months, and leaders who scored below roughly 70 to 80 percent lost the role. Semler says that's probably why he stopped being CEO after ten years. The company refused to track hours, review expense reports, or monitor where people worked. Read that list again and notice what each rule does: it strips managers of power they normally hold. Semler is also blunt about why the model rarely spreads. Giving up control takes a leap of faith, and almost nobody in control is ready to take it, especially on a 90-day earnings clock.
A founder who dissolved her own job. Helen Sanderson ran a UK health and care consultancy and stepped out of the CEO seat to join a self-managed team. She admits she went in assuming her organization already had this figured out, and found that the shift replaced the whole operating system rather than adding to it. Her home care teams, people supporting older and disabled clients in a regulated, low-margin sector, now run without managers. What makes her case the most useful of the three is that the limits don't depend on her restraint. They're structural. Teams write their own working agreements, record them, and review them on a schedule. Every member has a named buddy whose role includes challenge, not just support. Meetings explicitly expect tension to surface early. Teams choose their own new members. The founder can't quietly take the power back, because the power now sits in mechanisms other people run.
A submarine, a Brazilian industrial company, a home care team. The settings share almost nothing except the move: the person with the most authority bound their own hands, visibly, and the organization believed the change because of it.
Why this works: belief follows authorship
Bain published a statistic this month that explains the mechanism. In organizations undergoing structural change, 88% of senior leaders believe the new structure will achieve its aims. Everyone else, however, agrees that only 36% do.
Most commentary on that gap explains the 36%. The more revealing number is the 88. Leaders believe in the design because they spent months inside the reasoning, argued the trade-offs, and put their names on it. Their confidence is a byproduct of authorship, and authorship doesn't transfer by announcement. Everyone else gets a conclusion and judges it by what changes in their week.
The three cases above close the gap from both ends. When leaders give up some control, they also give up some authorship, and the people doing the work pick it up. Marquet's officers stated their intentions, which made every decision partly theirs. Semler's employees set salaries with open information and graded their own leaders. Sanderson's teams wrote the agreements they're held to, which is exactly why they hold each other to them. People defend rules they wrote. They comply with rules they received, until complying gets expensive.
One more reason the self-limiting move carries so much weight: A refusal that costs the business money is credible. A refusal that costs the leader personally, in control, in exceptions, in having the final say, is even more credible, because everyone can see the leader gave up what leaders usually protect most.
What a leadership team can do before launch
None of this requires running your company like Semco. It requires three decisions, made before the pressure arrives.
Write the refusals before the vision. Name the two or three things leadership will say no to in the first 90 days, including the exception the team itself will want most. A leadership team that commits to fewer priorities should decide in advance what happens when one of its own members brings a pet project. Deciding at the moment of temptation is deciding under the influence.
Put one limit on leadership's own power, in writing, stated publicly. For example: decisions pushed down in the new model won't be overridden by the executive team for six months, and any override requires a written explanation the affected team sees. Publicly stating it matters, because a commitment with a visible reversal cost matters more than a private intention.
Hand the people affected the pen. Let the teams living the change write their own working agreements, then build enforcement into the structure rather than into anyone's courage: named people whose role includes challenge, standing meetings where raising tension is expected, scheduled reviews against the agreements. Shared responsibility works when you assign it. "Everyone owns the values" reliably means no one does.
The third decision is what makes the first two durable. A limit that lives in one leader's self-discipline leaves with that leader. A limit that lives in mechanisms other people run survives the next reorg and the next CEO.
The diagnostic
Pull up your values page or transformation charter and count how many rules constrain people. If every rule points down, you've written a document for the people who were never the risk, and the organization already knows it.
Then ask the sharper version: when did leadership last give something up to keep a commitment? A deal, an exception, a decision it wanted to make itself? If nobody can name one, the change hasn't been tested yet. The first test is coming either way. The only question is whether leadership decides how to handle that moment now, or in the meeting where the pet project is on the table and the person who brought it signs the budget.
***
Sources: Nick Hobson and Margie Warrell, "Will Your Organization Stand By Its Values—Even When It's Hard?" (HBR, Sept 2026); Bain & Company, "Winning Operating Models That Convert Strategy to Results" (Oct 2026); TED talks by L. David Marquet, Ricardo Semler, and Helen Sanderson.
