The One Article I Recommend Most to Founders

Chasing scale too early kills growth. Here’s what to do instead.

If you’ve ever had a conversation with me in the early stages of your startup, there’s a good chance I’ve pointed you to Paul Graham’s classic essay “Do Things that Don’t Scale.”

I’ve been following Paul Graham’s writing since my time working with early Y Combinator startups like Xobni and Dropbox. His essays helped shape how I think about building and scaling startups—and this particular one is the piece I’ve shared more than any other over the years. I’ve probably sent the link to over 100 founders, usually when I sense they’re getting distracted by growth tactics that are way ahead of where they are in the journey.

The article came out in 2013, but the advice is still completely relevant today. Graham, who cofounded Y Combinator, the most successful startup incubator and accelerator in history, delivers a message that’s simple, powerful, and still not followed nearly enough:

In the early days of a startup, scaling is the wrong goal. Your job is to learn. And the best way to learn is to do things that don’t scale.

What Paul Graham Gets So Right

Graham’s key point is that most great startups begin with unscalable effort. That means personally onboarding users, giving them white-glove support, and sometimes even helping them use your product in real time. Not because it’s efficient, but because it’s how you understand what matters.

He tells stories from companies like Airbnb and Stripe, which both started with incredibly high-touch approaches. Stripe’s founders literally went door to door helping people install the code to accept payments. Airbnb’s founders took photos of hosts’ homes themselves.

That kind of work doesn’t scale, but it builds insight, and that insight is what eventually leads to product-market fit.

Where I Build on Graham’s Advice

This is where I usually step in with founders. Once you’ve done the unscalable things to get people using your product, how do you know if it’s working? How do you avoid chasing false signals?

In my experience, whether it’s at Dropbox, LogMeIn, or advising early-stage teams, you need a minimum number of active users to understand whether the product is truly resonating.

• For B2C, I’ve found that number is usually at least 1,000 active users

• For B2B, 100 active users is often enough to start seeing patterns

To be clear, I don’t mean 1,000 signups. I mean 1,000 people who are actually using the product consistently.

Below that, you’re often guessing. You might get lucky with one or two enthusiastic users, or misinterpret interest as value. But when you reach these thresholds, you start to see the signal through the noise.

Find the Must-Have Users

Even more important than hitting those numbers is identifying the users who would be very disappointed if your product disappeared tomorrow. I call these your must-have users.

Everyone else? They’ll churn eventually.

These must-have users are gold. They’re the people who found real value. And if you take the time to talk with them—not just survey, not just analyze, but actually talk—they’ll give you the insights you need to build an engine that creates more users like them.

In fact, these early users hold the keys to figuring out how to build a long-term scalable growth engine. They can help you understand not just what works, but why it works. And once you know that, you can reverse-engineer acquisition and onboarding flows that consistently lead to more must-have users.

This is also how you deepen your understanding of product-market fit. PMF isn’t just a milestone you pass—it’s the single most important driver of scalable growth. And the teams that grow fastest aren’t just the ones that hit PMF, they’re the ones who truly understand it and leverage it. They know exactly who the product is for, what value it delivers, and how to create more of that value for more of the right users.

Another major benefit of doing this work manually in the early days is that it helps you work out the kinks in getting started with your product. You’ll see exactly where new users get stuck, where they lose interest, or where they misunderstand the value. These friction points are often invisible in analytics but obvious in a direct interaction. Fixing them early can dramatically increase activation and retention down the line.

When I’m trying to extract these kinds of insights, I often build off the same questions used in my Product-Market Fit Survey, especially when someone answers that they’d be “very disappointed” if the product went away.

Here are a few of the follow-up questions I rely on:

• What type of person do you think would most benefit from this product?

• What is the main benefit you receive from using it?

• How can we improve the product for you?

• Have you recommended the product to anyone? (If yes, how did you describe it?)

These questions help you uncover the core value your product delivers, who it’s best suited for, how users describe that value to others, and where your product still has room to grow.

The Temptation to Scale Early (and Why It Backfires)

I get it. Scaling is exciting. You want to show traction. You want your dashboard going up and to the right. But here’s the trap: if you start trying to scale before you understand who your must-have users are and why they stick, you’re just buying vanity metrics.

And worse, you lose sight of the people behind the numbers. I’ve seen teams spend months optimizing a funnel, only to realize later that they were focused on getting users to the wrong experience in the product.

There’s a reason Airbnb and Stripe didn’t launch with scalable marketing engines. They weren’t ready. They didn’t need volume, they needed clarity. The same is true for most startups, even if it’s uncomfortable.

Final Thought

So yes, read Paul Graham’s Do Things that Don’t Scale if you haven’t already. It’s one of the most important pieces ever written for early-stage founders. And even though it was published over a decade ago, it feels like it could have been written yesterday.

But don’t stop there.

Once you’ve done the unscalable work to get real usage, make sure you have enough of it to see the patterns. Find the users who truly need your product. Learn everything you can from them. They are the blueprint for scalable growth. And only then should you start building systems to grow.

Scalable growth is absolutely the long-term goal. But ironically, the fastest way to get there is by doing the very things that don’t scale. Aggressively executing this early, hands-on approach gives you the insight and clarity you need to build a growth engine that actually works—and lasts.

That’s how you build something real. And that’s how you build something that lasts.