I was in a leadership forum this week when the conversation turned to span of control. Not in the abstract. In numbers.

There is a new floor forming: a minimum expectation that managers carry double-digit direct reports, with a future state being sketched where something closer to twenty becomes normal. Not fifty. Not the headline-grabbing ratio some of us have heard banded about. But unmistakably in that direction - and moving faster than my gut thought it would.

I felt two things at once. First, a kind of grim validation. Second, the uncomfortable realization that I still do not know how to think about it cleanly.

The Prediction vs the Minimum

In early March I published a blog arguing that we should expect span to stretch as AI eats coordination overhead and as organizations look for leverage in how they staff leadership layers. If you want the full argument, it is here: The 1:50 Ratio: Is the “Pod” Dead and the “Horde” Rising.

I was not predicting that every company would land on one manager for fifty engineers tomorrow. I was predicting pressure: fewer managers per engineer, wider spans, and a leadership model that rewards orchestration over intimate coverage of every report.

That pressure is now showing up as policy and expectation, not just as speculation. The minimum is no longer “seven plus or minus two” as a soft norm. It is becoming a line in the operating model. The trajectory points toward spans that would have looked reckless to me five years ago.

So I was directionally right. What I got wrong was the speed. This did not arrive on a comfortable glide path. It arrived like a schedule compression in a program plan - it’s the same destination, half the time budget.

Arbitrary on the ground, obvious in the spreadsheet

Here is the part I struggle with.

On the ground, a new minimum can feel arbitrary. Why twelve-ish instead of nine? Why march toward twenty? The human work of management is not linear. Trust does not scale like headcount. Coaching, conflict, performance risk, and moral distress do not neatly divide across a larger denominator.

And yet the spreadsheet story is brutally simple. If leadership layers are a cost center and the business is optimizing for efficiency, widening span is one of the cleanest levers available. You reduce management payroll, you reduce coordination nodes, you flatten hierarchy. In a world where boards and CFOs are hunting for structural savings, that logic is not mysterious. I also have experienced some real benefits of really flat organizations.

Can both things be true - that the number feels blunt as a management principle, and sensible as a business lever?

I think they can. That is what makes this era disorienting. The organization is not asking for a philosophical proof that ten or twenty is the “correct” cognitive load. It is asking for a staffing model that matches economic pressure and a bet that tooling and automation will absorb what used to require more human managers.

Whether that bet pays out is a different question than whether the decision is understandable.

Who pays for fewer managers?

If spans widen and layers thin, someone pays.

Some of that cost will be paid by engineers who get less consistent access to a manager who knows them well. Some will be paid by quality and culture drift that only becomes visible later. And some will be paid directly by managers who are no longer needed in the old shape of the org.

I want to name that plainly because it is easy to talk about “efficiency” in the aggregate and skip the human sentence underneath. There are skilled people, many of them friends and peers, who did everything we asked of them in the old model. They built trust, ran fair processes, carried incidents, and protected teams. In a compression cycle, they are the most exposed to redundancy.

I also want to name something uncomfortable for myself. I am not writing this from a balcony. I am writing it from inside the same risk pool. If the definition of “manager” shifts toward fewer roles with wider spans and more executive-style orchestration, then nobody should pretend their seat is automatically safe because they were good at the last version of the job.

Fear is not the same as fatalism. But dismissing fear as weakness is how leaders lose empathy, and how they blind themselves to personal stakes that matter.

The early march pause I am still glad about

There is one decision I am still pleased with, even as the ground shifts faster than I expected.

Around the time I published the March blog, I had a senior manager who wanted to hire another manager to relieve workload. It was a reasonable instinct in the old model: when coverage feels tight, add a layer or split a tree.

I pushed pause. Not because I enjoy saying no, but because I could see the vector forming. Adding managers is a long-lead commitment. If the organization was about to redefine what “normal” span means, I did not want to hire people into roles that might be structurally redundant before they could even prove value.

That call looks better today than it did even a few weeks ago. It does not make me clairvoyant. It makes me cautious in the right dimension.

What surprises me is not that the direction arrived. It is how little time we got between “signal” and “expectation.” That is the part that feels new. The AI era is not only changing what teams can build. It is compressing how long leaders get to adapt staffing and career bets before the model updates underneath them.

What am I doing with this uncertainty?

I am not going to pretend I have a tidy framework yet. I am doing a few practical things while I think harder.

I am revisiting what I believe a manager owes individuals when the span is wide: clarity of expectations, fair process, timely escalation paths, and honest conversation about limits. I am also revisiting what must be delegated to systems and peers so the job remains survivable.

And I am holding myself accountable to one standard: not confusing my discomfort with the number for a denial of the economic logic, and not confusing the economic logic for an excuse to stop caring about people.

Questions I am stewing on

If your organization is widening span, what are you explicitly stopping doing so the role remains honest?

When you look at your own career, are you betting on orchestration at scale, or on technical leadership with a smaller span - and have you chosen, or are you drifting? I’m not sure I have fully chosen yet.

And for those of us who saw this coming: what responsibility do we have to peers who will not be given as much time to adjust?

I do not have crisp answers yet. I suspect many of you are in the same place. If nothing else, we can stop treating the stretch as hypothetical. It is here. The only question is what kind of leaders we become on the other side of it.