Your company has plateaued while you keep hiring: the bottleneck is not where you are looking

The sentence comes back almost word for word from one first meeting to the next. “We hired, we restructured, we invested. And still it will not take off.” Then usually a list: a tightening market, a more aggressive competitor, two badly handled departures, an IT project that drags.

Rarely does that list include the person speaking.

This is not blindness. It is that a company’s growth ceiling never presents itself as a decision problem. It presents itself as ten separate operational problems, all of which happen to pass through the same office.

The ceiling exists, and it can be counted

Let us start with what cannot be argued. In 2023 France had 174,614 small and mid-sized companies and 7,442 mid-caps, according to Insee. The first group accounts for 29 % of full-time equivalent jobs, the second for 26 %.

How many cross the step each year? The French directorate general for enterprise puts it in writing in its Thémas no. 29 of May 2025: 825 companies moved up to mid-cap status in 2022, 575 of them crossing the threshold for the first time, that is 8 % of the mid-caps that year.

825 out of 174,614. The ratio speaks, and one should resist drawing a defeatist conclusion from it.

Because the same publication adds a figure few commentators pick up: of the 600 companies that became mid-caps in 2021, 85 % still were in 2022. In other words, the threshold is hard to cross, but not hard to hold. It is not an altitude the company fails to maintain. It is a wall it fails to break through.

The difference matters, because it changes the diagnosis. A problem of holding would be a problem of resources. A problem of crossing is a problem of how the company operates.

The only rigorously documented ceiling is regulatory

Before going further, the opposing case deserves its full strength, otherwise the discussion is worthless.

The best established French ceiling has nothing to do with the leader’s psychology. It is administrative. In work published in the American Economic Review in 2016, Luis Garicano, Claire Lelarge and John Van Reenen show that the obligations triggered at the 50-employee threshold amount to a variable tax of 2.3 % on labour, with a welfare cost estimated at 3.4 % of GDP. They observe in France a clear bunching of companies just below 50 employees, which does not exist in the United States.

That result is solid and should be taken seriously. It says that some French companies do not grow because growing is expensive.

But it does not say everything. It explains why some companies stop at a precise threshold. It does not explain why two companies in the same sector, of the same size, under exactly the same obligations, take opposite paths.

That is where the subject becomes interesting.

At equal size, what remains is how the leader works

The most serious work on this question is CEO Behavior and Firm Performance, published in the Journal of Political Economy in 2020 by Oriana Bandiera, Andrea Prat, Stephen Hansen and Raffaella Sadun.

The method deserves describing, because it avoids the usual trap of self-reporting. The researchers followed 1,114 chief executives in six countries, including France, and recorded 42,233 activities covering an average of fifty working hours per executive, quarter-hour by quarter-hour. Not what they say they do: what they do.

From that data they extract an axis. At one end, manager-type behaviour: a great deal of face time with operational staff, subjects handled one at a time. At the other, leader-type behaviour: time with the leadership team, meetings bringing several functions together, decisions taken in the presence of those who will have to execute them.

The result: a one standard deviation rise on that index is associated with 7 % more revenue, at comparable capital, labour and sector. The authors also note that it takes about three years for a change of chief executive to produce a measurable effect.

Seven per cent. Not for working more. For allocating time differently.

A second result completes the picture. In a study published in the American Economic Journal: Applied Economics in 2021, Philippe Aghion, Nicholas Bloom, Brian Lucking, Raffaella Sadun and John Van Reenen compare the behaviour of centralised and decentralised firms during an economic shock. The first saw revenue fall by 11.8 %, the second by 8.2 %. Three and a half points of difference, across a sample of 1,330 companies in ten OECD countries.

Honesty about its scope: that study measures resistance in turbulence, not growth in normal conditions. But it documents something rare, namely that the distribution of decision power has a measurable effect on results, independently of the quality of the people.

The blind spot

One question remains: if the mode of decision weighs that much, why do so few leaders look at it?

The Bpifrance Le Lab study on governance, covering 1,452 leaders, gives the most troubling answer in this whole file. Two figures, from the same survey.

25 % of the leaders surveyed consider themselves the sole decision-maker. And 92 % believe they are surrounded by people able to contradict them.

Both cannot be true at once in the same companies. What the contradiction reveals is not a lie, it is a blind spot: being surrounded by people who could contradict you is not the same as having an arrangement where someone actually does, on the subjects that matter, before the decision is taken.

Sixty-six per cent of those companies have a leadership team. So the question is not whether one exists. It is what is genuinely decided there, and what is merely announced.

One note of honesty: that survey was conducted in 2019. I found no more recent replication on French data. The figures should be read as an order of magnitude, not as a photograph of 2026.

What the day does to your judgement

There is one last factor, more physiological, and it must be handled carefully because it is often misquoted.

The cleanest study on the subject was published in JAMA Internal Medicine in 2014. It covers 21,867 consultations by 204 physicians. The probability of prescribing an antibiotic, including when it is not indicated, rises through the session: the adjusted odds ratio goes from 1.01 in the second hour to 1.26 in the fourth.

A quarter more risk of taking the easy decision rather than the right one, simply because you are at the end of a run.

Two safeguards are needed. First, these are physicians, not company leaders: the transposition is reasonable, it is not demonstrated. Second, a systematic review of 82 studies published in 2025 in Health Psychology Review concludes that only about 45 % of quantitative studies show a clear decision fatigue effect, and that the concept remains poorly defined in the literature. The phenomenon exists. It is not the law of nature it is sometimes made out to be.

What to retain is more modest, and sufficient: the order in which you place your decisions through the day is not neutral. The nine o’clock meeting and the seven o’clock meeting do not produce the same judgements.

Why hiring makes the problem worse before it solves it

Let us put it together.

A company reaches a plateau. The leader observes that everything is saturated. They hire, which is the most natural response in the world.

But every arrival immediately produces a series of new decisions: what exact scope, what boundaries with the existing team, what priorities for the first six months, what happens in a conflict with a peer. All of those climb to the same place. The execution benefit arrives later.

The result: for several months, the company has more resources and less throughput. The leader often concludes they hired badly. Sometimes that is true. More often it is a sign that the post was created without its decision perimeter being defined.

The figure that troubles this analysis also deserves facing. In the Bpifrance Le Lab and Rexecode barometer for the fourth quarter of 2025, 66 % of leaders cite weak demand as the main brake on their activity, against 33 % for recruitment difficulties. Almost none cite their internal organisation.

I am not claiming they are wrong about their market. I note that the internal factor never appears spontaneously in a questionnaire, for a simple reason: it is the only brake whose origin is oneself.

The three levers, in order

There is no recipe, but there is a sequence that avoids false starts.

One: inventory the decisions, not the tasks. For two weeks, write down every call that climbs to you, and sort it into two columns: those that belong to your role, and those that are habit. The second column is almost always the longer one. It is also the easier to deal with.

Two: move where the decision is taken, not only where it is executed. Delegating a task without delegating the judgement that goes with it does not reduce your load, it adds a validation step. That is exactly what Aghion and his co-authors measure when they compare centralised and decentralised firms.

Three: protect the order of the day. High-stakes decisions do not belong at the end of a run. It is the cheapest lever on the list, and the only one that produces an effect the following week.

This work replaces neither a commercial strategy nor a financing plan. It conditions their execution. A company does not cross a step because its leader works more. It crosses when decisions stop having a single point of passage.