Building Startups by Doing the Unscalable: Lessons from Paul Graham

By now, you’ve probably picked up on the fact that I’m a big fan of the chaotic, exhilarating world of startups. It’s where innovation meets grit, and nothing exemplifies this better than Paul Graham's essay "Do Things that Don’t Scale". He drops some golden advice on how early-stage companies can push through the noise by doing the unscalable things—recruiting users manually, delivering extraordinary customer service, and perfecting the product step by step. Let’s dive into the key lessons from his iconic piece!

1. Recruit Users Manually

Graham starts by emphasizing how early user acquisition is often manual. Startups shouldn't expect users to come to them magically. You need to seek out and recruit your first users. He points out how companies like Stripe went door to door to personally onboard their first users. He stresses: "Instead of sitting back and hoping users will come to them, most startups need to get out there and drag people in."

Key takeaway: In the beginning, beg, borrow, or convince people to use your product—even if it means reaching out one by one.

2. Focus on Customer Delight

Once you have users, go out of your way to make them happy. It’s not enough to have them try your product; you have to ensure they love it. In Graham’s words, "You can turn users into fanatics by taking extraordinary measures." This means prioritizing their needs with hands-on attention. Airbnb famously provided professional photography for listings, a personalized service that helped win over hosts.

Key takeaway: Do the hard work to make every early user feel like royalty.

3. Perfecting Your Product

Graham suggests startups shouldn’t worry about scalability at first. Early on, the focus should be on building a product that a few people absolutely love, not one that many people kind of like. The intense, non-scalable efforts to please early users will help you refine the product better than a thousand hypothetical surveys.

Key takeaway: In the early days, spend more time refining your product than worrying about scaling.

4. Be Scrappy and Personal

Another theme is that startups should be scrappy and willing to do anything to get going. A personal touch is invaluable, especially in a world increasingly dominated by impersonal, automated interactions. He cites how early-stage companies have a secret weapon—direct access to their users.

Key takeaway: Personal interactions help you learn from users and build lasting loyalty.

5. Finding Early Evangelists

Graham notes that the first users of a product are not just customers; they can become champions of your brand. The best startups find users who will evangelize their product to others. Dropbox is a great example—its early users loved it so much they spread the word organically, effectively becoming free marketers.

Key takeaway: Treat your early users well, and they will help you grow your user base.

6. Act like a Consultant

Sometimes you’ll need to function more like a consultant than a software engineer. Graham points out that a lot of successful startups found initial traction by doing things that looked more like consulting gigs—building custom solutions for individual users that eventually turned into a scalable product.

Key takeaway: Custom solutions in the early days can lead to scalable innovations later on.

7. Launching Incrementally

Lastly, Graham advocates for launching early and iterating quickly. Don’t worry about perfection on day one; worry about learning from real users. Early feedback is more valuable than waiting for a perfect launch. Airbnb and Dropbox are prime examples of this. As Graham states, "Launching teaches you what you should have been building."

Key takeaway: Launch early, gather feedback, and iterate fast.

Conclusion

In essence, startups often find their footing by doing work that won’t scale—like hands-on user acquisition, custom solutions, and extraordinary customer service. As Graham puts it, these "unscalable" efforts lay the foundation for scalable success.

For the full essay, check it out here.

This was Paul Graham's take, and now it's your turn: what "unscalable" thing are you doing right now to grow your startup?