Do Things that Don't Scale

One of the most common types of advice we give at Y Combinator is to do things that don't scale. Many founders believe startups either take off or don't, but in reality, startups take off because the founders make them take off. Founders must recruit users manually at the start, as nearly all startups do. Stripe, a successful startup we funded, is a prime example of aggressive early user acquisition, a technique we refer to as "Collison installation."

Founders resist recruiting users individually due to shyness, laziness, or the belief that the numbers seem too small to matter. However, they underestimate the power of compound growth. Measuring progress by weekly growth rate is crucial; a consistent 10% weekly growth leads to significant numbers over time. Airbnb is a classic example of a startup that took heroic measures to get rolling, including going door to door in New York to recruit users.

Startups are fragile initially, and it's a mistake to judge them by the standards of established ones. Founders often dismiss their startups too early without realizing their full potential, as even Bill Gates did when he briefly returned to Harvard after starting Microsoft.

To recruit users manually, if you build something to solve your own problems, you only need to find your peers. Otherwise, observe which users are most enthusiastic and seek out more like them. Extraordinary measures should be taken not just to acquire users, but also to make them happy. Startups should focus on providing an insanely great user experience, even with an early, incomplete product.

Sometimes the right unscalable trick is to focus on a deliberately narrow market, like Facebook did initially with Harvard students. For hardware startups, doing things that don't scale might involve assembling products manually, as Pebble did with their watches.

Consulting-like techniques can help recruit lukewarm users by acting as if you're building something just for them, tweaking until it fits their needs perfectly. This can also apply to software startups, as we did at Viaweb by building online stores for merchants.

The Big Launch usually doesn't work for startups; what matters more is how happy the initial users are. Partnerships with big companies also usually fail to provide the expected growth. Startups should think of their ideas as vectors, with the unscalable things they'll do initially to get the company going being just as important as the product itself.

In conclusion, startups need to do something unscalably laborious to get started, and this effort is not just a necessary evil but can change the company permanently for the better.

The original article: https://paulgraham.com/ds.html