Entrepreneurship.

The framework

1. 01

   Manual recruitment, in the dullest sense of the phrase.

   The first users don't come to you, because there is no reason yet for them to have heard of you or to trust what they hear. So you write down thirty named people — not segments, names — and approach each one individually, in whatever way genuinely reaches them: an introduction, a message that proves you have read their work, a conference, a doorstep. Ten personal approaches will teach you more this week than a launch will, and the list itself is the asset, because building it forces you to be specific about exactly who this is for.

2. 02

   Overservice is the half founders skip.

   Recruitment is only the first half of Graham's essay. The second is what you do once somebody has said yes: do by hand what the product will one day do automatically, deliver it yourself, answer inside minutes, get on a plane if that is what it takes. This isn't generosity — it's the fastest instrument you have for seeing the gap between what somebody said they wanted and what they will actually use, and it buys a kind of early loyalty no amount of spend reproduces. The right level is slightly embarrassing. If it feels comfortably efficient, you aren't learning much.

3. 03

   The manual work is a research method, so keep the notes.

   Every hand-delivered order and every over-long support call produces something a dashboard will never give you: the thing the customer did that you did not expect. If nobody writes those down, you have paid the full cost of unscalable work and banked none of the benefit. Keep one running document — what we did by hand this week, what surprised us, what we changed because of it — and read it back at the end of each month. That document becomes your roadmap, and later your automation queue, in roughly the order the entries appear in it.

4. 04

   Set the exit condition now, while it's still cheap to say.

   The phase ends, and it ends far better when the trigger was named in advance: at so many customers, or when one step starts eating a given number of hours a week, or by a date on the calendar. Then decide what each manual step becomes — automated, somebody else's job, or a deliberate luxury you keep because it has quietly become part of what people are buying. A few unscalable things are worth carrying forever at low volume. But if the business still functions only while you're in every conversation, that isn't a phase you're passing through. It's the shape of the company you have built.