A company can only pay attention to so much at once. That is not a figure of speech. Attention is a budget. It is finite, and you can roughly count it.
Robin Dunbar put a number on the human version: about 150 relationships before the brain runs out of room to track them. A company is a pool of those budgets, stitched together. The question that decides its future is what it spends that attention on.
Attention faces one of two ways. Outward, at customers and the market and the signal of what buyers are doing. Or inward, at itself: keeping its own people aligned, its meetings synced, its processes fed.
Every unit of attention spent inward is one not spent on the customer. That is the tax. It never appears on an invoice, and it is the most expensive thing the company pays.
Why mass levies the tax
Here is the part that should change how the next hire feels. A stack of laws, from fields that never speak to each other, arrive at the same conclusion. Adding mass spends the budget inward.
Parkinson watched bureaucracies breed administrators who mostly make work for one another. Brooks showed that each person added to a team multiplies the lines of communication faster than the output. Past a point, the team gets slower.
West and Kleiber found that as an organisation grows, its overhead rises faster than its production, the way a larger animal must live at a slower pace. Dunbar set the ceiling on how many people any of them can track before formal process has to stand in for memory.
Four fields. Public administration, software engineering, biology, anthropology. They converge on one finding, with no collusion: capability does not scale with headcount, because headcount is paid for in attention, and attention is fixed.
When that many independent disciplines land on the same answer, it is not an opinion you can manage your way around. It recurs with the regularity of a law.
What capability is
If capability is not headcount, what is it? The cleanest answer comes from cybernetics. Ashby's law of requisite variety says the ability to control a complex environment is a matter of variety: having a range of responses that matches the range of things the world can do to you.
Too little variety and the environment overwhelms you. You cannot meet what you cannot match.
Variety has a partner the org chart forgets: attention. Variety is the range of responses you can make. Attention is the selectivity to spend that range on the disturbance in front of you, not on the noise.
Enough responses to matter, and the focus to aim them. That is what a go-to-market needs. Neither half is bought by adding bodies.
Note that the most consequential idea in modern AI carries the title Attention Is All You Need. In its own register it made the same point. The breakthrough was not more machinery. It was a mechanism that decides what to attend to and drops the rest. Different field, same lesson. Once you can act, attending to the right signal is the whole game.
The tax is escapable
The encouraging part: the inward tax, unlike the laws beneath it, is not a fact of nature. It is the consequence of one way of growing, by adding central mass. And nature is full of the alternative.
A murmuration of thousands of starlings holds together with each bird minding about seven neighbours, no conductor anywhere. An ant colony coordinates the work of a city with no meetings, leaving signals in the environment for the next worker to read. Variety and coordination, without the central mass that taxes attention.
A company can borrow the same trick. Automate the bounded work, so output grows without adding a node someone has to manage. Rent senior variety fractionally, so you carry it only while it performs. Instrument the relationships, so the system tracks the thousands and hands you the few. Each move buys capability without buying the tax.
Spend it outward
Which leaves the only question that matters. You have a fixed budget of attention. Call it your 150. You can spend it coordinating yourselves, or you can spend it on the customers and the work that make revenue.
The company that wins is not the one with the most people. It is the one arranged so that the smallest possible share of its attention faces inward, and the largest possible share faces the market.
You do not need to be bigger. You need to stop paying a tax you were never told you were paying, and point the mind you already have at the people who pay you.
The operating question is not "how many people do we need?" It is this: of all the attention this company spends, how much of it faces the customer, and what would it take to spend the next unit outward instead of in?
Founder-led. Not founder-limited.