Why Middle Management Fails in Promoter-Led Businesses
It is 9:40 at night and a supervisor is calling you directly about a dispatch that will not go out. He reports to your operations manager, who is at home, has not been called, and will hear about the problem tomorrow morning — from you.
You have corrected this before, in a review meeting, in front of everyone: route things through your manager. It held about two weeks.
The explanation you have probably settled on is that you appointed the wrong man. This article argues something narrower and less comfortable. The seat is hollow, and the occupant is almost irrelevant. Your middle manager has the title, the headcount reporting to him and the responsibility for the outcome. He has no decision rights, no budget, and no real say over who joins his team or who leaves it. Everyone knows this, including the people who report to him, which is exactly why they route around him to you.
So he does the only job the structure leaves available. He relays — your instructions down, their problems up, nothing added in transit, because he is not permitted to add anything. Relaying is not managing, and a company built of relayers cannot produce ownership at any level, because nobody between you and the work is allowed to decide. This is rung two of the five conditions ownership requires, failing where it does the most damage.
Search this question and you get Harvard Business Review, Gartner and McKinsey treating it as a support problem — the squeezed middle, manager overload, the accidental manager nobody trained — answered with coaching and development programmes. That assumes the seat is sound and the person in it under-resourced. Here the reverse is true, and teaching a man to occupy an empty chair more confidently does not help him.
The Hollow Layer
Nobody designs a management tier with no authority in it. It forms, invisibly.
How the layer gets hollowed — one exception at a time
A customer calls you directly with an urgent problem. Your manager is in a meeting, the customer matters, the answer takes you ninety seconds. You take it. On its own, that is a correct decision.
Two weeks later the same customer calls you again, because last time it took ninety seconds. Your manager, hearing about it afterwards, says nothing. What would he say.
A month later one of his own team members needs an answer today, sees the manager's calendar, and calls you instead. Four minutes.
Nobody decided to bypass the manager. There was no meeting about it. The bypass is simply faster, and under pressure speed wins every time. Within a year the shortest route to a decision runs to you for everything that matters. The manager's role was never withdrawn; it was out-competed.
Each exception was defensible, urgent and small, which is why the hollowing is invisible from your chair. The sum of them is a layer that exists on the org chart and nowhere else.
What the hollow manager actually does all day
Watch the job honestly and it is three activities, none in his KRA sheet.
- Relaying. Your instructions down, their questions up. Asked what he thinks, he answers with what you are likely to think, because his own view is not load-bearing.
- Buffering. He absorbs the friction between what you have asked for and what is possible this week, managing your reaction rather than the work, and becoming skilled at it, because it is the part of the job he controls.
- Absorbing blame with no authority attached. When the dispatch misses, it is his failure. He could not have set the credit terms, chosen the vendor, replaced the underperformer or approved the overtime that would have prevented it. The outcome is filed under his name anyway.
The third makes this the most stressful job in an Indian mid-market business. Not the workload — the ratio. Full accountability, no control, and a team that has worked out he cannot help them.
Why the good ones leave first
The capable manager notices the hollowness within about four months. He is the one who most wanted to decide, which is why you hired him, and therefore the one for whom the seat is least tolerable. He will not say this on his way out. He will say he got a better opportunity, more scope, a role closer to home.
The manager who stays is the one for whom relaying is an acceptable job. He is not lazy and not stupid. He has optimised, correctly, for the role as it actually exists. Run that filter for four years and your middle layer consists entirely of people comfortable not deciding. Then you conclude mid-market talent is weak, when you are looking at the residue of your own selection.
The Two-Boss Test
You do not need a consultant for this. Four questions and an afternoon.
Pick a junior person two levels below you — not the manager, someone who reports to him. Ask in normal conversation, and take the first answer rather than the corrected second one.
- Who decides your increment?
- Who approves your leave for next Thursday?
- If a customer escalated tonight at nine, who would you call?
- If you disagreed with an instruction from your manager, who would you go to?
Do it with five or six people across two or three departments, without telling their managers.
If your name appears in more than one answer, that manager is decorative. If it appears in all four, he is not a manager at all, but a senior coordinator carrying a manager's designation and a manager's stress.
Watch the hesitation too. When someone pauses, names the manager, then checks your face to see whether that was right, the pause is the finding. Most promoters run this expecting one weak manager and find the same pattern in every department.
What the evidence says about this layer
Gallup's State of the American Manager, built on 27 million employees across 2.5 million work units, found that managers account for at least 70% of the variance in engagement scores across business units. Be precise, because this is misquoted constantly. It does not say the manager causes 70% of engagement. It says that when you compare one team with another inside the same company — same pay bands, same brand, same policies, same promoter — around seven-tenths of the difference between them tracks to the manager.
That is good news, because it makes the problem local. Two units in your business with identical policies and wildly different results are telling you about one seat, not about Indian workers or your industry — far smaller and cheaper to fix than a culture.
The second figure is the warning. Gallup's India data shows engagement among managers falling from 39% to 30% in a single year, while individual contributors fell from 24% to 19% — almost twice as fast. The layer you rely on to fix everything else is deteriorating faster than everything else.
Jim Harter, Gallup's chief workplace scientist, put it plainly: "Manager engagement affects team engagement, which affects productivity. Business performance — and ultimately GDP growth — is at risk if executive leaders do not address manager breakdown."
The two complications nobody writes about
Neither appears on the Western pages. Both are ordinary here.
The family member in or above the layer
Sometimes the hollow seat holds a relative: a brother-in-law running operations, a nephew on his way up, a cousin who has been in purchase a decade.
The difficulty is rarely capability; plenty of family members are the most capable people in the room. It is that no consequence can attach to the role, and everyone in the building knows it. A manager who cannot be performance-managed is not a manager in any operational sense, because the layer below him calibrates to that fact. Place him above a professional manager and he hollows that manager out in one stroke, without doing anything wrong.
The honest handling is not removal. It is to stop asking a family role to carry a management function it structurally cannot. Either he takes a defined domain with a real number and is reviewed in the same forum as everyone else, or he sits outside the reporting line of non-family staff. The middle position, where most of these arrangements live, is what does not work — a governance question, taken apart in family business governance in India.
Seniority by tenure
The other common occupant is the fifteen-year man. He joined when you were doing ₹6 crore, has never taken a day of unexplained leave, and was made General Manager because there was no other way to recognise fifteen years.
He is not the villain here. He is usually the most loyal person you have and knows things nobody else knows about your customers and your history. The company promoted him for service, a decision about fairness, then judged him on capability, which he was never developed for.
Contempt is both wrong and useless. Honesty means separating two questions: what is he genuinely excellent at, and is that the job the title now describes? Often the answer is a senior specialist role with real money and standing attached, rather than a management seat he never asked for. The related pattern is in why good employees underperform.
Why hiring a better manager does not fix it
You may have lived through this one. You decided the layer was weak and hired properly: a strong professional from a larger, more organised company, at a salary that made you slightly uncomfortable.
Then he did what a real manager does. He made a decision — a vendor change, a pricing exception, a person he was moving out — without checking, because in his last company that was his job. It collided with you. Once, in public.
From that moment he was finished, and the organisation knew it before he did. The team had been waiting to see whether the new authority was real, and got their answer in one meeting. Strong hires into hollow seats fail faster than weak ones, precisely because they try to use authority the seat never had.
Replacing the occupant is the most expensive way to discover that the seat is the problem. Landing a senior outside hire is covered in why executive hires fail; sequencing the layer, in how to build your first leadership team.
The repair: narrow but real beats broad but notional
The instinct here is to announce decentralisation. Do not. Announced authority is notional, and the first time it is overturned it is worth less than nothing, because the layer has now been told twice.
Take one manager. Give him three genuine decisions, each with a stated boundary. Not principles — decisions. Discounts up to 8% on orders below ₹5 lakh. Up to ₹50,000 a month on his department's operating problems. Final say on confirming anyone in his team after probation.
The boundaries matter as much as the grants. A right with no stated edge is an invitation to be overruled at an unknown point, and people feel that from a distance. Writing them down is covered in the decision bottleneck.
Then the part that is the actual intervention. You refuse to take those three decisions, even when asked, and especially when asking is easier for everyone. When the supervisor calls at 9:40: that is his call, call him. When the customer calls you personally for a better discount: that is his call. When a family member asks you to overrule it, the same sentence, in front of whoever is present.
That refusal is the whole repair; everything else is paperwork. The layer becomes real the first time you say that is his call in public and mean it, and stays real only if you keep saying it in the case where his answer was worse than yours would have been. That case arrives within a month and costs something real: a discount you would not have given, a person you would not have confirmed. Pay it. You are not buying that decision, you are buying every decision the layer makes for years.
People will keep calling you for weeks while they test whether this is a mood or a rule. Then watch two things: whether the layer holds weight sideways between departments and not only downward, the failure examined in why execution breaks down at scale, and whether consequence follows the authority both ways — otherwise you have moved the blame without the control, the asymmetry in the accountability gap.
Where to start
Run the Two-Boss Test this week. If your name comes up in more than one answer, do not start recruiting. Pick your best manager, write down three decisions with boundaries, and spend the next month refusing to make them.
If you would rather see the pattern across the whole organisation first, the Business Pulse diagnostic maps where decisions are actually made in your business, as against where the reporting structure says they are. In a hollow layer, the distance between those two maps is the whole problem.