Paul Graham says one of the biggest mistakes startup founders make is trying to scale too soon. Instead, in the early days, you should focus on things that don’t scale, like the activities that may seem small or manual, but help you get those first users and keep them happy.

You can't wait for users to come to you. You have to go out and get them.

One example he gives is recruiting users manually. Don’t wait for people to find you. Go out, meet them, email them directly, even sign them up yourself if needed. In the early days of Stripe, the founders literally took people's laptops and helped set up accounts on the spot.

Another big one is delighting users personally. When you have just a handful of customers, you can give them a level of attention that’s impossible later on. Graham shares that doing “things that don’t scale” like talking to customers constantly, making quick changes just for them, and going the extra mile will make them love your product and tell others.

He also talks about the idea of being a “concierge”. In the early stages, you might do manual work behind the scenes to deliver the product experience, even if later you’ll automate it. That’s fine. It lets you learn what users want and refine your service before building complex systems.

The main point: early growth often comes from these high-effort, small-scale actions. Over time, you can automate and streamline. But in the beginning, these personal, unscalable efforts help you build trust, understand your users deeply, and create something worth scaling.

Source: Paul Graham, "Do Things that Don’t Scale"