Sharp Distribution Beats Perfect Product.
A short, opinionated essay on the first hundred users, the first ten thousand views, and the small mathematical dignity of being heard at all.
There was, in my second year of university, a boy in my hall called Marcus, who had built — over the course of a long, sleep-deprived autumn — what he and the seven people in his Discord agreed was unmistakably the best note-taking application ever made. It had, I will not lie to you, several beautiful features. It had a feature where, if you held down the option key, the entire interface dimmed by a small and pleasing amount. It had a feature where the cursor, having been idle for a particular interval, blinked in time with your breathing — assuming, that is, that you were breathing in 4-7-8, which Marcus believed everybody secretly was, or should have been.
Marcus's application now lives, as far as I can tell, on a single Heroku instance which serves a 503 error on alternate Tuesdays. The seven people in the Discord have moved on. Two of them, last I heard, are using Notion. Marcus is in graduate school.
I think about Marcus a great deal, because Marcus is what happens when a person believes, sincerely and at considerable personal cost, that the product is the thing.
The product is not the thing.
The arithmetic nobody admits to.
Distribution, as a word, has the slightly unwashed quality of all words borrowed from logistics. People in software like to say it the way people in the eighteenth century used to say commerce — quickly, with the eyes elsewhere, as if the saying of it might attach. They prefer to talk about product. Product is dignified. Product can be put on a slide. Product permits, on certain occasions, the use of the word craft, which is the highest aesthetic compliment a man in a Patagonia vest is permitted to give.
And yet. The arithmetic of any consumer software business, when you do it on the back of a receipt at half past one in the morning, comes down to two numbers, multiplied. The first is how many people will hear of the thing. The second is what fraction of them will, having heard, do anything about it. You are permitted, in software, to be excellent at one of these. You are not permitted to be terrible at both, which is what, in practice, the perfect-product fundamentalists usually are.
Marcus had a conversion rate, as best as I could tell, of about forty per cent. This is, by the standards of the industry, extraordinary. Forty per cent of the people who landed on his website became users. The trouble was that the people who landed on his website numbered, in any given week, around eleven, of whom seven were Marcus.
The doors of Toronto.
I sold things door-to-door for nine months when I was nineteen — knives, in case it matters, although it scarcely does — and I learned in that time three things which have never quite left me. The first is that almost nobody is rude. (Almost everybody is, on closer inspection, simply tired.) The second is that a sufficiently specific opening sentence will buy you four further sentences, almost always, almost regardless of what the first sentence said. The third — and this is the one I want to talk about — is that distribution is not, in fact, a separate discipline from product. It is the same discipline, conducted in a different room.
What I mean is this. The thing the customer is buying — the thing she is, in her tired, polite way, being asked to consider — is not the knife. It is also not the explanation of the knife. It is the small, slightly miraculous experience of having been correctly understood by the person on the doorstep. If you have understood her correctly — if your opening sentence has lodged in the small, particular shape of her Tuesday — she will give you the four further sentences, and quite often the cheque.
Software is not different. Software is door-to-door, conducted at scale, conducted in writing. The opening sentence is the headline. The opening sentence is the tweet. The opening sentence is, in increasingly many cases, the eighteen-second clip somebody has watched while waiting for their pasta to cool. If the opening sentence is wrong, no amount of dimming-on-option-key will save you. The customer has already turned away from the door.
A small confession about the first hundred.
The first hundred users of any product I have ever shipped have come, almost without exception, from a single small act of indignity on my part. I have, at various times, hand-written eighty-four cold e-mails in a single sitting; I have stood, slightly damp, outside a hackathon in Berkeley handing out stickers I had printed at a Walgreens; I have, on one memorable evening, persuaded my own grandmother — who is eighty-two, and who has not, to my knowledge, ever installed an application of any kind — to download a beta and give me feedback over WhatsApp. (Her feedback was: the green is too green. She was, of course, correct.)
The thing about the first hundred is that they cannot be acquired through cleverness. They can only be acquired through embarrassment. Founders who refuse to be embarrassed do not have first hundreds. They have, instead, a very nice landing page and a Stripe dashboard whose graph is a flat, dignified line.
Sharp, not loud.
I want to be precise, because I am about to be misunderstood. Sharp distribution does not mean shouting. It does not mean the sort of LinkedIn essay that begins I just hit a million ARR and ends with a thread of seventeen tweets each containing one platitude in a different font. Shouting is the distribution strategy of people who have not yet noticed that everyone else is also shouting.
Sharp distribution means knowing — with an almost rude specificity — where the people you want to reach already are, and what they are already, in the back of their heads, half-thinking about. It means going to that place. It means saying the sentence that would, were you in their shoes, cause you to look up. It means, in a way that the Patagonia-vested fundamentalists find vulgar, doing the work of being interesting in public.
It is harder than building, although it does not look it. The reason it looks easier is that, when it is done well, it leaves no fingerprints. The customer thinks she found you. She did not, of course. You found her, with the patient malice of a man who has spent six weeks trying to work out where she has lunch.
Marcus, and what I would tell him.
I had a coffee with Marcus, briefly, last December. He was, if anything, more talented than I remembered. He had built another thing — a small, beautiful something to do with calendars — and he wanted, I think, to be told that this one was the one. I did not, on the whole, think it was. The thing was elegant. The opening sentence was not.
What I should have said to him — what I have, in the imaginary version of that coffee I have rehearsed since, said many times — is this. The product, Marcus, is fine. The product has always been fine. Spend a quarter of the time you spent on the cursor blinking in 4-7-8 on the question of who, exactly, is going to find this, and how, and on what afternoon, and in what mood. Then, if you are still talking to me at the end of it, we will discuss whether the cursor blinks correctly.
I did not say this. I drank my coffee. I told him the calendar thing was lovely, which was true, and I went home and worked, instead, on my own opening sentence — which I am, you will be relieved to know, still working on now.
— B.