The operating structure of a one-person FDE company

A one-person FDE company rarely fails on skill. It fails because the same twenty working days have to produce both the next deal and the current one, and delivery is the part that makes noise. Selling is the part that gets postponed until delivery is finished, and delivery is never finished.

Two clocks, one calendar

There are two kinds of hours in this business and they are not interchangeable. Selling hours produce conversations, samples and proposals; doing hours produce something a customer can check. Neither is optional. A week of only selling ends with a full pipeline and nothing delivered; a month of only doing ends with a happy customer and no next one.

The ratio is the business. Everything below is a way of moving it: raising the price, shortening delivery, reusing what was built, or selling something that does not consume the calendar at all.

Four incomes on one bench

A one-person practice usually ends up with four revenue lines, and they are not four separate businesses. They are one bench serving four purposes.

Custom delivery. The two-week first stage, priced around 10,000 yuan in one practitioner's account, with cumulative custom work reaching 50,000 before anything else took over. It pays now and, more importantly, it produces the material every other line depends on.

Consulting, or being the boss's outside brain. One practitioner moved from cohort coaching at roughly 6,000 per round to engagements around 30,000, with a ticket of 20,000 to 50,000 and more than 500,000 accumulated over five-plus months. The margin per day is the best of the four. The prerequisite is a diagnosis you can only earn by having delivered something.

Cohorts and training. Cheap per head, and not a business on its own. Its real function is to test whether the material is teachable and to qualify people who might become clients.

Product. A tool, a template, an internal agent, anything that produces output without consuming your hours. It is the only line that can grow while you sleep, and the only one that can be sold without a calendar.

The feeding order matters. Delivery produces the diagnosis that makes consulting credible; consulting produces the patterns that suggest what to productise; the product shortens delivery and frees selling time; the cohort tests the material and produces leads. Reverse the order — product first, delivery later — and you have a guess with a subscription.

What a month actually holds

A model, not a measurement. Twenty working days. One 10,000-yuan first stage takes ten days of build plus two of communication and sorting: twelve. Eight remain for selling, support and rework. A consulting engagement at 30,000 needs perhaps four days of preparation and two on site. So the honest ceiling is one delivery plus one engagement per month, not four deliveries.

The practitioners who describe this say the same thing in different words. One reports monthly profit above 100,000 in a one-person shop and says capacity, not demand, is the limit. Another puts it more bluntly: the ceiling is twenty-four hours a day. Both are self-reports, and both point at the same constraint — the constraint is not the market.

The signal that forces productisation

Productisation is usually attempted too early, for a reason that feels like ambition: the first delivery goes well, the second looks similar, and a product seems obvious. The test that matters has four parts, and all four have to hold.

Is the same problem appearing independently in three or more customers? Is the cost of generalising lower than the benefit, given that generalising often costs more than writing the first version? Can someone who does not write code use it? And after it is productised, is there still work left for the person on site, or does the field role degrade into configuration?

A simpler signal sits underneath: the third time you write the same script, or the second time you explain the same rule to a different customer, the thing is a product candidate. The first time it is a delivery. The counter-signal is just as strong: if every customer's version needs a different data source, a different approver, or a different definition of the number, generalising will cost more than it saves, and the honest move is to keep it custom and raise the price.

What the structure cannot do

Four limits are structural rather than temporary. A one-person company cannot carry procurement-heavy deals, because the cycle consumes the selling time it was supposed to fund. It cannot promise round-the-clock support, because there is no second shift. It cannot run two sites at once, however good the tooling is. And it cannot be the only person who understands the system it leaves behind — that is a handover failure waiting for a date.

There is a fifth, less obvious one. One person is not the same as alone. The most consistent lesson in the accounts of people who have done this is that the isolation, not the workload, is what degrades the work: no one to check a diagnosis, no one to say the scope is wrong. If you have not yet run a full enterprise project with a team, doing one round with people who have is worth more than a year of solo practice, which is the argument in how people actually become FDEs. The seat model — one person plus a paired agent — is the closest a solo practice gets to a second opinion that is not yourself.

The test

For four weeks, log two numbers every working day. Days that produced something a customer could check. Days that produced a conversation with someone who could buy. If either falls below a third of your working days, the problem is structural, and the fix is a product decision or a price decision — shorten what you deliver, or charge enough that the delivery stops eating the month. Working evenings is not one of the options; it only moves the ratio in the wrong direction.

The pricing side of that decision is in scoping and pricing. The structural side is the sentence above: one of the two clocks has to stop being paid for by the other.