In 2013 Paul Graham wrote an essay telling founders to do things that don't scale. I'd put it more bluntly: do things that suck.
Most of the things that work early are things nobody wants to do. They're slow, or boring, or a little humiliating. That's not a coincidence. It's why they work. Everyone can see the obvious move, so the obvious move is crowded. The move that sucks is empty, because most people take one look at it and find a reason not to.
Airbnb is the famous example. In its early days, Graham wrote, the founders were "going door to door in New York, recruiting new users and helping existing ones improve their listings." Brian Chesky and Joe Gebbia took the "professional" photos of their first hosts' apartments themselves. Nobody starts a company dreaming of that work. But Graham's judgment was that about 30 days of going out and engaging with users in person "made the difference between success and failure."
Stripe did the same thing in miniature. When more diffident founders get a yes to "will you try our beta?", Graham wrote, they say they'll send a link. The Collison brothers didn't wait. When anyone agreed to try Stripe, they'd say "Right then, give me your laptop" and set them up on the spot. Y Combinator ended up naming the move after them: the Collison installation.
So why doesn't everyone do this? Graham's answer is blunt too. Founders resist recruiting users one by one partly out of "a combination of shyness and laziness." They'd rather sit at home writing code than go out and talk to strangers, most of whom will probably say no. The other reason is that the numbers look too small to matter. Ten users signed up by hand look pathetic next to the million you're imagining.
But that's exactly the filter. Things that suck have a price, paid in boredom, discomfort or pride, and most people won't pay it. So if you will, you're competing with almost nobody. And the ten users you signed by hand teach you things a million never would, because you were in the room when they used it.
It works the same way outside startups. The salesperson who makes the calls everyone else avoids. The writer who rewrites the paragraph one more time after it's already good enough. None of it is clever. It's just unpleasant enough that most people skip it.
When you're choosing between two ways to do something, notice which one you're avoiding. It's often the one that works, and the reason you're avoiding it is the same reason everyone else is.
So don't scale yet. And when it starts to suck, keep going. That's usually where everyone else stopped.