The first time I watched this happen, I did not have a name for it. A general manager I had promoted from a single Bay Area unit into a three-unit oversight role was working harder than anyone in the region. Twelve hour days. Weekends. She was the first person on every group text and the last person to leave every unit visit. Her numbers, taken as an average across the three restaurants, were fine. Not great. Fine.
What she was not doing was thinking. Or hiring. Or coaching the next layer of shift leads under her. All the work she had done brilliantly as a single-unit operator, she was still doing, just spread across three kitchens. The three units were quietly being run by whoever was on shift, because the person meant to be running them was answering the last message that had come in on her phone.
Six months in, I moved her back to one unit at a slightly larger scope. She loved me for it. Ten months in, that unit was the highest performing in the region. She is still there. That is not a failure story. That is a match story.
What "does not scale" actually means
Scale is not a size problem. It is an operating model problem. The reason certain leaders stall as their scope grows is that they built their success on personal touch, and personal touch does not divide.
At one unit, a great general manager can carry the operating rhythm inside their head. They know every line cook's schedule, every regular guest's preference, every piece of equipment's history. That intimacy is the whole product. Guests can feel it. Teams can feel it. The P&L reflects it.
At three units, that same intimacy has to be transferred to systems, or it has to be transferred to other people. If it is transferred to neither, then three units get one third of the operator's attention each, and none of them get the intimacy that was the source of the original result.
The operator did not get worse. The seat got bigger. The question is whether their model can grow with the seat, or whether the model is the ceiling.
The early signals are calendar-shaped
Nobody's numbers get bad in the first two months of a stretched scope. Numbers lag. The signals that a report is not scaling show up in their calendar and in their reactions, not on the P&L, and they show up two full quarters before the numbers do.
Fig. 1 · Calendar signals lead the P&L by roughly a quarter.
Their calendar fills with rework
Look at their last two weeks of meetings. If more than a third are titled something like "follow up on" or "circle back to" or "re-review," you are watching a leader who cannot close a loop in a single pass. That is not laziness. That is a person operating past their capacity, trying to catch what falls through by repeating the pass.
Their peers stop asking them for help
The other general managers in the region used to ping this person for advice on a hairy schedule question or a tricky guest situation. Now they do not. That is a signal. Peers stop asking when they no longer expect an answer in time to matter. It is quieter than a complaint. It is more accurate.
Their unit-level numbers stay strong, their portfolio numbers do not
The individual unit they visit most often looks great. The other two drift. Six months in, the average is fine and the variance is enormous. That is the shape of a leader trying to be the general manager of every unit instead of the leader of every general manager.
The 90 day coaching contract
If you see the signals early, you have room to run a real intervention. Ninety days is the right window. Anything less is not enough to reset a habit. Anything longer teaches the person and the team that the standard is negotiable.
The contract is written. Three pages, not thirty. It has three parts.
Two behaviors to stop
Specific ones. Not "delegate more." Something like: "You will not personally sign off on any schedule at any of the three units. The general manager on shift signs their own schedule and you review the weekly variance on Monday." That kind of specific.
Two behaviors to start
Also specific. "You will run a 30 minute one-on-one with each of your three general managers every week, same time, same agenda. You will not cancel or reschedule these for any reason other than a family emergency." Again, specific.
One clear success measure
Not "improve performance." Something the person can look at and know if they are winning. In this case usually: "By day 60, the three general managers under you will each be running their own weekly P&L review with their shift leads without you present, and you will attend one of those reviews per month as an observer."
Sign it. Both of you. It is not legally meaningful. It is psychologically meaningful. A written contract you both signed is much harder to drift away from than a verbal understanding.
Fig. 2 · Three parts. Written. Signed. Reviewed every two weeks.
The three outcomes at day 90
At day 90, one of three things is true. You have to be honest about which one.
The person changed their operating model. This does happen. Not always, not most of the time, but it does. When it does, you will see the shift in the peer signals first. Other general managers start pinging them again. Their calendar has whitespace where it did not before. Their one-on-ones are on the books and hold. When you see that, the contract worked. You extend the runway, take the guardrails off, and let them run.
The person is trying and it is not landing. This is the hardest case. They are doing the behaviors on the contract. They are running the one-on-ones. They stopped signing every schedule. And their portfolio is still not working, because the underlying operating instinct is still to be the doer, not the leader. In this case, the seat is bigger than the person, and the honest read is to move them, not stretch them further.
The person did not really try. This is actually the easy case, even though it feels bad. If you can point to the specific behaviors they did not do, the specific meetings they cancelled, the specific rhythm they broke, the conversation is easy. The person either commits in a different way for the next 30 days, or you make the move.
Move, do not demote
When the seat is bigger than the person, a demotion in place is almost always the wrong move. It humiliates the person, confuses the rest of the team, and gives you the worst of both worlds: they still have the title long enough to lose the room, and you still have the problem long enough to lose the year.
The right move is a lateral. Sometimes back to a single unit at a slightly larger scope. Sometimes into a specialist role, like leading a catering channel or running new-unit openings, where their operating style is a fit and their strength is a real asset. In one case at Hana Group I moved a strong general manager into a training and openings role across the region. She has since led six unit openings, all of them cleaner than the openings I ran myself two years earlier. That is not a consolation prize. That is a match.
A great one-unit operator running one unit at a very high standard is worth more to the region than the same operator pretending to run three at a mediocre one.
The mirror the conversation holds up to you
There is a version of this conversation that is really about the reporting leader, not the report. If you have three or four direct reports at once who are not scaling, the pattern is not them. It is you.
The two most common versions of that mirror:
- You are hiring for your old operating model. The people you keep promoting are versions of the operator you used to be at their level. That worked when your scope was theirs. It does not work now that your scope is bigger. You need people whose instincts match the seat you are trying to build, not the seat you already grew out of.
- You did not resize your own calendar. Your reports cannot scale into decisions you are still making. If you are still signing every schedule, approving every hire, and reviewing every guest complaint personally, you have no way to know whether your direct reports could scale into those decisions, because you have not let them try.
Every regional director I know, myself included, has learned this the hard way at least once. The first time a report of mine did not scale, I thought it was about the report. The second time, I started asking whether it was about me. It was mostly about me.
The conversation the person deserves
When the day 90 read is that the seat is bigger than the person, the conversation you owe them is a real one. Not softened. Not corporate. Not framed as "let us think about your development path." A real one, in your own voice, in a room with a door.
Two things to prepare before you sit down. First, the specific behaviors on the contract they did do and the ones they did not. That way the conversation rests on evidence, not impression. Second, the specific alternative role you have in mind, if you have one. Do not walk into the conversation with "we should figure something out." Walk in with "here is the seat I think fits you." The person can push back on the specific idea. They cannot push back on nothing.
End the conversation with a clear timeline. Not "let us talk again soon." Something like: "I want to make the move official in three weeks. In the meantime, you and I are going to tell three people together, in this order." That specificity turns what would otherwise be an ambiguous, painful stretch into a real handoff with a defined shape.
The point
Direct reports do not scale for one of two reasons: their operating model has a real ceiling, or your operating model is not making room for a bigger one under you. Both are fixable. Neither is fixable by pretending.
Watch the calendar signals early. Run the ninety day contract with three parts and a signature. Move, do not demote. Look in the mirror every time this happens more than once in twelve months. That is how a region gets bigger without breaking the people you promoted to grow it.