Starting the business you have been describing for years

The idea is good enough that you have explained it at three dinners. You know who it is for, roughly what it costs, and why the existing options are annoying. Someone has told you to just do it, and you agreed, and that was two years ago.

Nothing about that is a knowledge problem. You are not waiting on a fact. You could sketch the whole thing on a napkin tonight, which is precisely why more thinking has stopped being useful.

What is missing is smaller and more awkward than a plan: one person, not related to you, who has paid you something.

The work that feels like starting a business is mostly not it. The name, the logo, the domain, the company registration, the accounting software, the landing page, the reading. All of it is real work, all of it is legible as progress, and none of it involves a stranger deciding whether to give you money.

That is what makes it so effective as avoidance. It cannot be criticised, because every item is genuinely on the list eventually, and it produces artifacts you can point at. Months of it feel like momentum right up until you notice the thing has not been sold to anyone.

The useful test on any task is blunt: does this bring a first customer closer this week? Registering a company does not. Asking eleven people whether they would pay does.

Almost everything before a first sale is a guess, including the good parts. The price is a guess, the target customer is a guess, and the thing people say they want at a dinner is famously not what they buy on a Tuesday.

A first customer settles more of that in an afternoon than six months of preparation. It tells you the problem is real, that your version of the solution is close enough, and roughly what it is worth. It also tells you when you are wrong, which is the more valuable half and the half people spend years avoiding.

Small counts. A single client, one paid pilot, one person who buys the rough version. The point is not the money; it is that somebody made a decision that cost them something.

The reason the business has not started is usually that the next step is not a step. "Launch" is a category, "build the product" is a season, and neither fits in an evening after work. Steps that cannot be finished cannot be crossed off, so nothing about the week tells you whether you moved.

What fits a weekday is narrower and slightly uncomfortable: writing down exactly what you would sell and to whom, in one sentence; sending that sentence to five people who have the problem; asking one of them what they currently pay for the bad version.

None of that requires quitting anything, and all of it is more informative than another month of preparation. The version of this that works is boring and small, and it happens on ordinary Tuesdays rather than in a decisive weekend.

The standard story says you cannot start while employed, and it is mostly told by people selling the leap. In practice the job funds the experiment and removes the pressure that makes a first version desperate. Plenty of things worth building were built in the hours around one.

What the job does take is the evening energy, which is a scheduling problem rather than a commitment problem. Two protected hours a week that survive a bad month will outrun five hours a week that only exist in a good one.

The honest version of the question is not whether to quit. It is what the smallest weekly commitment is that you could keep for six months while everything else stays the same.