How a smart idea gets initial traction is one of the most important moments in any business. It’s a confusing, inventive, fascinating period, and one Andrew Chen spends a lot of time thinking about.

Chen is a partner at the exceptionally busy and successful investment firm Andreessen Horowitz, a.k.a. a16z. He just wrote a book on the topic titled The Cold Start Problem, which lays out how startups have cracked that code.

He’s been out on the podcast and webinar circuit talking through the highlights of the book, and during conversation with Stripe he focused on why the first users are so, so important. First users of what, and so important to whom? I know what you mean, but spell it out.

Andrew Chen

“If you open up Dropbox at your office and it is just empty, nobody uses it. Or nobody uses Airbnb, you’re a lot less likely to use it,” said Chen. “You coworker next door who might be excited, even if you’re both excited, but you’re not using it (What does it mean/refer to here?) on the same projects, it’s just not going to work.”

He has some specific knowledge of the gig economy, as he was responsible for growing the ridership at Uber and helped the company reach 1 billion rides, then 2 billion just six months later. The second billion ride milestone wasn’t easy, but it was easier than the first few hundred.

“This is the cold-start problem, you have a phase in all these products where you don’t have any users so the product isn’t going to work, but you won’t have any users if the product doesn’t work,” said Chen. “Only when you solve that are you able to tap into your network effects effectively.”

The spark that launched companies like Uber, Lyft, Airbnb and every delivery app after that has one thing in common, said Chen. He calls it the “atomic network” or the minimum viable network of users to create a stable, useable product upon which to scale.

“At Zoom, it’s two to three people; Slack, it’s five to 10 people—it’s very important to get five to 10 people using it at the same time, you need that dense network,” said Chen. “For Tinder, they found they needed 500 people on a college campus.”

Tinder famously threw lavish parties for those college kids, spending all sorts of money on fancy cocktails at swanky venues for college kids. To get in, they had to sign into the dating app. The next morning, perhaps after some hangover pancakes, they started swiping. Today, it’s so ubiquitous that everyone knows what “swipe right” means.

“If you can build one atomic network, you can build a second and a third, that’s what you see,” said Chen. “Tinder threw college parties to get people using the app. Then they could throw a party at all these different colleges and eventually they merge into this larger network.”

He saw early on the atomic network in Clubhouse, the live chat platform that exploded onto the social media scene in 2020. Chen was invited when the app had just 100 users.

“We invested in the Series A, we invested when there were 500 active daily users. So, it was very small,” said Chen. “But if you’re inside an atomic network that is functioning, you should be able to tell that it’s working. We invested just based on the idea that this one atomic network was working. Even with 500 daily users we were able to value the company at nearly $100 million.”

Whether you think listening to people chit-chat is worth $100 million is up for debate, but there are plenty of examples in the food-delivery space. Snackpass is building out its atomic networks, for example, and Chen sits on the company board. The social food-ordering and sharing app started on college campuses, and it’s expanding out into the real world.

The atomic network for the company was initially just a handful of dorm mates. From there it grew virally with very, high organic engagement. That led to $70 million in investments for the company.

Chen likes to see evidence of those networks and people sticking with the product for the long term.

“I like to see annual growth oof at least three to five times a year, that maps to 20 to 30 percent monthly growth. I ideally want that to be organic traction so it shows that the product can leverage the network,” said Chen. “I don’t like when startups give money to Facebook and Google. And engagement, you like to see really, really high engagement. After 30 days, is it over 20 percent? If they’re able to retain 20 percent of users on day 30 that’s very exciting.”

To reach massive scale, Chen said founders should take any means necessary to create those atomic networks that lead to traction and then scale.

“How fast do I have to grow to get to 100 million active users? When you do the math, what you often find is the consumer companies need to hit 100 percent numbers every year. That’s why the tech industry is obsessed with very fast growth rates,” said Chen. “It’s not just about getting to network effects, once you get to network effects, you need to figure out how to scale it.”

That may be the topic of his next book, but founders and visionaries need to start atomically small to solve that cold-start problem.