100 People Love

Paul Graham dropped it almost as an aside in his 2013 essay "Do Things That Don't Scale." But the sentence became the most repeated piece of startup advice in Silicon Valley: it's better to make something a hundred people love than something a million people kind of like.

The arithmetic feels wrong. A million users sounds like a business. A hundred sounds like a hobby. Graham's insight cuts against that intuition because he's not talking about market size — he's talking about the physics of growth.

Love is a different substance than lukewarm approval. People who love a product tell other people. They forgive bugs, tolerate missing features, and evangelise without being asked. People who "kind of like" something use it when convenient and abandon it the moment a shinier alternative appears. One group compounds. The other churns.

Airbnb is the canonical proof. In early 2009 the company was nearly dead — revenue flat, the founders maxing out credit cards. Paul Graham gave them a specific directive: go to New York, meet your hosts in person, and make them love the product. Not "like" it. Not "find it useful." Love it. Brian Chesky and Joe Gebbia flew to New York and went door to door. They photographed apartments with professional cameras. They redesigned listings by hand. They asked hosts what would make the experience perfect and then built those things, one host at a time. The New York hosts became fanatics. They told other hosts. The growth curve bent upward. Everything that followed — the global expansion, the $100 billion IPO — traces back to the decision to make a few hundred people in one city love the product instead of trying to get millions worldwide to tolerate it.

Stripe ran the same playbook from a different angle. Patrick and John Collison didn't launch with a marketing campaign. They showed up at startup events, found developers frustrated with payment integration, and said "give me your laptop." They installed the Stripe API on the spot — seven lines of code, working payments in minutes. What became known as "the Collison installation" wasn't a growth hack. It was an act of obsessive service aimed at making each individual developer love the experience. Those first few hundred developers didn't just prefer Stripe. They were furious that payment integration had ever been hard. That fury — love's close cousin — drove organic adoption that no advertising budget could replicate.

The non-obvious insight is that the number 100 is a proxy, not a target. Graham isn't prescribing a literal headcount. He's saying you should optimise for intensity of need. If you can find even a small group of people for whom your product solves a burning problem — a problem so acute they'll endure anything to get the solution — you have something real. The intensity is the signal. A million people who shrug at your product is noise.

This inverts the standard startup playbook. Most founders obsess over acquisition: how to get more people through the door. The 100 People Love model says the question isn't "how many users do we have?" but "how much do our existing users care?" Get the intensity right and the numbers follow. Get the numbers right without the intensity and you've built a leaky bucket.

What Graham calls "love" is really a cluster of observable behaviours. Users who love a product have high daily engagement. They refer friends without being incentivised. They submit feature requests instead of switching to competitors. They write public praise without being asked. They complain loudly when the product breaks — not because they're angry, but because they depend on it. These behaviours are measurable. The model isn't asking founders to chase an emotion. It's asking them to optimise for a specific, quantifiable pattern of intensity.

The model also carries an implicit rebuke to the dominant startup orthodoxy of the 2010s. The "growth at all costs" framework — fuelled by cheap venture capital and winner-take-all market dynamics — told founders to acquire users as fast as possible and worry about retention later. Companies like Groupon, MoviePass, and Homejoy followed this logic to spectacular initial growth and equally spectacular collapse. Each had millions of users. None had users who loved them. The wreckage of premature scaling is the best evidence for Graham's thesis: the count of your first users predicts nothing; the intensity of their engagement predicts everything.