The leverage curve: capacity is no longer a headcount question

Two companies can have the same number of people, similar technology and comparable demand — and still deliver very different levels of output. The difference is rarely effort, and rarely talent. It is how the work is designed.

For a century, the answer to "we have more work" has been the same: hire more people. That instinct is quietly becoming expensive — and often unnecessary.

Those three outcomes — hire, redesign, or cut too far — are the whole article. Here they are, one question at a time.

Why does growing usually mean hiring more people?

For most of business history, capacity has been a hiring problem: more demand, more people — budget a role, fill a seat, add some output. It is familiar and it feels safe. But it quietly assumes the process itself cannot be improved, so every rise in demand gets paid for in headcount.

The cost of that assumption shows up in the data. Studying firms between 2001 and 2013, the OECD found that the most productive — the top 5% in each sector — grew output per worker about 3.6% a year in services while everyone else managed just 0.4%; in manufacturing, 2.8% against 0.6% ^[1]. Year on year, that compounds into a chasm. The OECD points to several causes, and the one worth acting on is the most practical: the leading firms were quicker to put new tools and methods to work — to redesign around them — while the rest kept staffing the old way of working ^[1].

Add people to a process you never fixed and the maths turns against you: output goes up a little, headcount goes up more, and the output of each person actually drifts down. You have bought volume and called it capacity. (Later we will call this the under-leveraged trap — too many hands on a flow that was never redesigned.)

What changes when you redesign how the work gets done?

Redesign breaks the link between output and headcount. Instead of every extra unit of work needing an extra pair of hands, you reshape the flow — simplify steps, automate the routine, and point people at the parts that genuinely need judgment — so the same team handles more. That ratio, useful output per person, is what this article calls leverage: redesign is how you raise it without raising the payroll.

Make it concrete. Picture a team handling 1,000 customer requests a month with 10 people, and demand doubles to 2,000. There are two honest ways to get there — set your own numbers below and watch the trade-off.

The point is not that redesign is magic. It is a genuine choice with a genuine cost: hiring buys capacity you keep paying for; redesign asks each person to do more, which only works if you actually reshape the work. If a process is already sound and simply under-resourced, hiring is the honest answer — the model above will say so.

And here is the encouraging part. The technology now reshaping work helps most where you would least expect — with newer and less-experienced people, not the veterans. In a field study of customer-support teams, an AI assistant lifted output by 14% on average — but novices and lower-skilled workers improved 34%, while the most experienced barely changed ^[3]. It handed the biggest gains to the people who had the least to begin with.

That is no longer a frontier experiment. Stanford's AI Index reports a growing body of research that AI "boosts productivity and, in most cases, helps narrow the gap between low- and high-skilled workers", with 78% of organisations using AI in 2024, up from 55% a year earlier ^[2]. The tools are widely available; the differentiator is whether you redesign the work to use them.

Can you take efficiency too far?

Yes — and it is the mistake that looks smartest on a spreadsheet. Keep stripping out people to hit a headcount target and you eventually cut the experienced staff who catch what the system gets wrong. Throughput still climbs for a quarter; then rework, errors and key-person risk quietly eat the gains. This is the over-leveraged firm — efficient on paper, fragile in practice. The skill is stopping just before that edge, where enough human judgment remains to keep the output trustworthy.

The whole idea on one curve

Put those three questions on a single picture and you get the leverage curve. The left is hiring into a process you never fixed — the output of each person stays low. The middle is the healthy zone: redesigned work, the same team doing more. The right is cutting too far. Drag the marker to move between them.

Where does your function sit?

Most teams have never located themselves on that curve. Five quick questions will place you — and point to the next move.

How to move into the middle

Wherever you land, the shape of the move is the same: stop buying capacity by the seat and start designing it. These four plays take a function into the healthy zone — and keep it there.

None of these is a tool you purchase; they are changes to how work is shaped, who does what, and which number you manage by. Capacity used to be a question of how many people you could afford. Increasingly it is a question of how much useful work each person can do — and for most teams the ceiling is far higher than the next hiring plan assumes. The job is to design the work, and to stop just before efficiency starts costing you quality.