Principal Resolution

He was not a difficult leader to work for. That is worth saying at the start, because the story that follows can read like an indictment, and it is not one. He was thoughtful, genuinely invested in the people around him, and capable of articulating where the organization was going in a way that made people want to help it get there. In his first year, he had taken a mid-sized professional services firm through a strategic repositioning that most people in the building believed in. The direction was clear. The reasoning was sound. The team had heard it enough times to repeat it back without prompting.

What they could not have told you, by the end of the second year, was whether any of it still applied.

The first exception was reasonable.

A legacy client — one of the firm's oldest relationships, outside the new strategic focus, the kind of account that predated the repositioning by a decade — came back with a significant renewal. The revenue was real. The relationship was real. The argument for taking it was straightforward: this is one client, this is one year, the direction hasn't changed, we are simply being pragmatic about the transition.

He approved it. His leadership team accepted the reasoning. A few of them noted, privately, that the account sat outside the repositioning. None of them said so in the room.

The second exception had different logic but the same shape. A new opportunity — adjacent to the strategy, not quite inside it, lucrative enough to warrant a conversation. He took it to the leadership team. The discussion was substantive. The conclusion was that this was the kind of work the firm would be doing in three years anyway, and moving toward it now was acceleration, not deviation.

He approved it. Two members of the leadership team exchanged a look. Neither said anything.

By the fourth exception, no one was calling them exceptions anymore.

The strategy did not disappear. It lived in the deck that got refreshed each year, in the language of the all-hands, in the answer he gave when a prospective client asked where the firm was headed. He believed it when he said it. That is the part that made it hard — he was not describing something he had abandoned. He was describing something he was slowly making impossible to reach, one reasonable decision at a time.

The people who worked for him learned what the organization had taught them to learn. The stated direction was a reference point, not a constraint. The real filter for whether a piece of work got approved was whether the leader found it compelling in the moment — and he found many things compelling. He was curious, energetic, responsive to a good argument. Those were genuine strengths. They were also, in the absence of a structural check, the mechanism of the drift.

The senior staff stopped building three-year plans. Not formally — they still produced them, still presented them, still filed them in the places plans get filed. They stopped believing in them. The planning process became a ritual that everyone performed and no one relied on. When a new initiative arrived, the question the team asked each other — not in meetings, in the conversations after meetings — was not "does this fit the strategy" but "is he excited about it." Those two questions had once had the same answer. They no longer did, and everyone knew it except him.

He found out the way leaders usually find out about drift — not through a conversation that named it directly, but through an accumulation of signals that eventually became impossible to read as anything else. A key hire declined the offer. The recruiter's debrief was diplomatic. His COO, who had been with him since before the repositioning, handed him a resignation letter and said she needed to tell him something first.

She told him.

She was careful about it, precise, and she had clearly decided that this was the conversation she owed him before she left. She named the exceptions. She named what the team had stopped believing. She named the look that had passed between two leadership team members in the room where the second exception got approved — she had been one of them — and what it meant, and what it meant that neither of them had said anything. She told him that the organization had spent eighteen months learning that his stated priorities were negotiable, and that once that lesson is learned it is very difficult to unlearn, because the learning is not in anyone's head. It is in the behavior the organization has already locked in.

He listened. He did not defend himself. That is also worth saying — he heard it, and he understood it, and the understanding was visible on his face in a way that was hard to watch.

Then he asked her what he should do.

She had thought about this. She told him that the direction itself was not the problem — it was still right, still worth pursuing, still something the organization could orient around if it believed the orientation was real. The problem was that he had no mechanism for making his commitments binding on himself. Every exception had been approved by the same person who set the priority the exception was overriding. There was no structure in the gap between the two — nothing that made the cost of the exception visible before the approval happened, nothing that required him to account for what the exception would teach the organization about whether the direction was real.

She left. He stayed. And he did something that leaders in this situation rarely do: he told his leadership team what she had said.

Not an edited version. Not a reframe. He told them what she had told him, including the part about the look that passed between two of them in the room, including the part about the planning process becoming a ritual no one relied on. He said he wanted to know if it was true. He asked them to tell him.

They told him.

What followed was not a transformation. It was a series of smaller decisions that, taken together, changed what the organization was able to rely on.

He built a filter into the approval process for any work outside the stated strategy. Not a veto — he was still the decision-maker — but a required step: any exception had to be named as an exception, documented as one, and reviewed against the cumulative pattern of exceptions before approval. The first time the filter surfaced how many exceptions had been approved in eighteen months, he sat with the list for a long time before he said anything. The number was not what he had thought it was.

He reinstated the COO role and hired into it someone whose explicit mandate included pushing back on him. He told the new COO, in the room where they negotiated the terms of the role, that the most important thing she could do was tell him when his behavior was contradicting his stated direction — and that he was aware he would not always receive that well, and that he was asking her to do it anyway. He put that in writing. Not because he didn't mean it, but because he knew that verbal commitments from leaders about their own openness to challenge have a short half-life when the challenge arrives in a real moment with real stakes.

He stopped treating the annual planning process as a ritual. He started treating it as the mechanism by which the organization held him accountable to his own direction — which meant it had to have teeth, which meant the exceptions had to show up in it, which meant people had to be able to point to the plan and say "this is what we said we were doing, and this is what we actually did."

The senior staff did not immediately believe any of this. Belief is rebuilt the same way it is destroyed — through accumulated evidence, over time, one decision at a time. The first time an exception came to the table that met the old criteria — compelling, adjacent, lucrative — and he declined it, citing the filter, citing the pattern, citing what he had been told about what the approvals had taught the organization, a few people in the room recalibrated. They did not say anything in the meeting. They said something to each other afterward.

That is how it starts. Not with a declaration. With a decision that holds.

Broken Compass does not always look like a leader who doesn't know where they're going. Sometimes it looks like a leader who knows exactly where they're going and cannot stop redirecting the organization away from it — one reasonable exception at a time, each one defensible in isolation, none of them survivable in accumulation.

The direction was real. The exceptions were reasonable. The drift was the product of both — and it was invisible to the person producing it until someone who was leaving had nothing left to lose by saying so.

The COO told him what she saw because she was walking out the door. A different organization is built so that someone can say it while they're staying — and so the leader has already built the structures that make hearing it survivable, and acting on it possible, before the resignation letter arrives.

Every organization has that window. Most organizations are in it right now.

One Exception at a Time

Seeing this pattern in your organization?