The Cold Start Problem (2021) explains what network effects are, how they work in practice, while illustrating them with real-world examples, from companies like Zoom, Airbnb and Uber. From getting tech companies off the ground in the 21st century, to the population dynamics of meerkats, The Cold Start Problem is an in-depth look at the way networks develop and interact with each other.
Who is it for?
- Startup leaders looking for rapid growth
- Established CEOs looking to reawaken sluggish companies
- Anyone interested in network dynamics
What’s in it for me? Learn how a company can win out through the power of network effects.
Those crazy cats in Silicon Valley love their buzzwords and slogans. But right now there’s one phrase eclipsing all the others: network effect. It’s used so often, by hopeful entrepreneurs and established CEOs alike, that it’s become something of a joke.
When start-ups pitch to investors, they kinda use “network effect” as an answer to every question. They’re asked How will you deal with competitors? Their answer? Network effect. They might get the question: How will you scale in foreign markets? They’ll answer: Network. Effect.
But what on earth is network effect and why is everyone talking about it?
To explain network effect, let’s take a look at Uber. Uber grew, district to district, city to city, all across the US, until it became the massive global company it is today. And it grew through the network effect.
It worked like this: as more users joined the Uber app, the more likely it was that they’d find others to share a ride with. Naturally, this also meant that drivers found it easier to locate someone who needed a ride. As more and more people interacted with the technology, the more exponentially the network grew. And the more lucrative Uber became. This is network effect.
But to understand things even better, let’s go back in time and look at an earlier technology: the telephone.
At the beginning of the twentieth century, there were fewer than 5 million telephones in the US. These 5 million phones had to serve nearly 90 million Americans. But one phone company, the American Telephone and Telegraph Company – today’s AT&T – was growing fast.
At the heart of that growth was company president, Theodore Vail. He understood network effect better than anyone at the time. In 1900, he observed that a telephone without a connection is one of the most useless things in the world. If no one else has a phone, how on earth can you call them? You need other people to have a telephone for it to be of any worth.
Vail put it this way: its value depends on the connection with the other telephones and increases with the number of connections.
And that is the network effect in a nutshell. Whether we’re talking about apps, from Instagram to Snapchat, or rideshare companies like Uber and Lyft, the product is only as valuable as the network using it.
Take the network away and the product simply ceases to exist.
Launching new tech is very difficult in the twenty-first century.
Okay, now that we know what network effect is, we might be tempted to think, well, that it would be easy to take advantage of. Everywhere you look, from the metro to the sidewalk, you’ll see people squinting into their devices, hooked on apps and games.
You’d think, then, that right now would be the optimal moment for launching new technology. The environment seems perfect to turn a plucky little start-up into the next Tinder or Zoom. Network effect means that new products can easily attract users through word-of-mouth and organic growth. Even a tiny start-up can leave established companies in the dust. Right?
Wrong. Utilizing network effect is far from easy.
In fact, it’s very difficult – we live in the age of squeezed attention spans, where only the most useful, or engaging, apps and technologies will succeed.
Let’s go back a moment to 2008. That was the year when the iPhone apps platform hit phones and devices everywhere. Back then, when there was barely anything on the platform, all a new app had to do to succeed was be more interesting than a commute or waiting for a bus. That was easy.
A decade later and it’s a different story. Today, the App Store has several million apps, all competing for attention. To succeed, any new app has to steal attention from all the other extremely addictive apps – many of which have been optimized over time to engage users. For years, the top charts of the Google Play Store and Apple App Store have looked the same.
Success is even difficult for massive established giants looking to move into a new market where a smaller company dominates. Even if a larger company offers the same product as a smaller competitor, it won’t be able to break through if the smaller company has captured the market and is growing through the network effect.
Take the battle between Snapchat and Instagram. When Instagram tried to copy Snapchat’s features, like its Stories and photo messages, it just couldn’t outdo its rival. The reason? Snapchat had a resilient and growing network that Instagram just couldn’t capture.
Meerkats can tell us a lot about the network effect.
By now, we should have a handle on network effect and how it works. But let’s go a little deeper and approach the concept from a more unusual angle.
It’s time to talk meerkats.
If you’ve watched a lot of wildlife documentaries about the African savanna, you’ll have seen a row of meerkats standing to attention, like little people.
Like us, meerkats are hypersocial creatures. And curiously enough, because they’re so social and interact so frequently, they can tell us a lot about network effect.
Meerkats stick together so they can keep a lookout for predators. If they don’t take their turn on “meerkat watch” then it’s likely that a monitor lizard, leopard, or python will sneak up on them and snaffle them up. As you might imagine, whether or not they stick together has a big impact on their population numbers.
If a meerkat mob doesn’t have enough numbers for an effective watch, then it’ll be much easier for individuals to get eaten. Pretty soon, as predators make off with meerkat after meerkat, the population will collapse until there are none left.
Alternatively, if there’s a big, healthy group of meerkats, they’ll be able to keep growing and branch out into multiple mobs. At this point, they’ll reach what’s called the tipping point and begin to increase exponentially. However, if they become too numerous, the population will run out of food. At this point, the numbers will begin to plateau and decrease slightly.
This whole process was described in detail in the 1930s by a professor at the University of Chicago called Warder Clyde Allee. The name of the tipping point – beyond which a population begins to grow fast – is called the Allee threshold.
Meerkat dynamics tell us a lot about the network effect in tech companies. For instance, remember Myspace? Like a healthy meerkat population, the social network reached the Allee threshold – the tipping point – in the mid-2000s and grew rapidly.
But then, not long after its enormous, epoch-defining growth, a powerful competitor, Facebook, entered the scene. Soon Facebook stole much of Myspace’s network, and, like a depleted meerkat population, Myspace collapsed fast. When faithful users logged into Myspace, they found just silence and tumbleweeds.
And with nobody to read their posts or “friend” them, they vanished, too.
When a network begins to expand rapidly, it reaches a point known as “escape velocity.”
All right, let’s recap: once a meerkat population reaches a certain size, it passes a threshold and begins to expand dramatically.
The very same thing can happen to a network. When a network passes beyond that threshold, and grows super fast, it enters something known as escape velocity.
When a company hits escape velocity it’ll start to make thousands of new hires, launch ambitious new projects, and attempt to grow internationally. Its goal is to grow and grow the network effects that have led to its early success.
To really understand how escape velocity works, you have to break it down. If you look a bit closer, you’ll see that there are three distinct forces at work.
The first of these is called the acquisition effect. This is the point when the positive early experiences of users lead them to invite others into the network. It can be seen in the early story of PayPal, the payments giant. When PayPal was getting off the ground, it offered new users cash incentives to invite others to set up an account. Naturally, many took up the offer and the network exploded.
The second of these forces is called the engagement effect. As the network grows, it’s possible to increase user engagement with the product by introducing them to new use cases and deepening their experience. Let’s turn to Uber again. Uber started “leveling-up” riders who were taking airport trips to dining out. This encouraged a section of its network to think of Uber as something that could do more than just get them from A to B.
The third and last force in play in escape velocity is the economic effect. This is when the economic performance of a product begins to match the rapid expansion of the network. For instance, a multiplayer game like Fortnite, which sells customized items and weapons, will start to monetize more effectively as a gamer’s friends sign up to play together. Or take Slack as another example. As more teams within a company sign up for the service, it’s more likely that the company as a whole will become a paying customer.
All of these factors are part of a process that will light a rocket-booster under a network’s growth.
The “cold start problem” is the first obstacle that businesses need to overcome.
I hope by now you have a good idea of what network effect is and how it happens. You’re also probably a bit fed up with me saying it again and again.
So, now we’re going to talk about something else. We’re going to talk about the cold start problem – which is the title of the book after all.
To get an idea of what a cold start problem is, imagine trying to start a car on a frosty morning, You’re finding it hard to get the motor going. However much you try the ignition, nothing seems to happen. Your breath is misting in the freezing cold. Eventually you give up, and call for a mechanic to help.
Just as it is for old cars, so can it be for businesses. It can be difficult for a business idea to get up and running. And this is the cold start problem.
So, what’s the root of the problem? Well, it boils down to either networks of people not connecting with an idea or them jumping on the idea for a short period of time, but finding it unsatisfactory. It can happen to the smallest start-up or the biggest organization.
To use an example, imagine a movie-streaming app by an established brand. Imagine it launching to great fanfare. Users flock to the service at first. But it doesn’t offer a wide selection of content to start with. There just isn’t a lot of content in the library. So users don’t stick around. They end up drifting away, never to return. And that is the end of that. That’s the cold start problem.
The solution to the cold start problem lies in that first network a product secures. Those initial users are the most important in the story of a company – if you don’t win them, you don’t win anyone.
To succeed at this early stage with these early users, a company should try and build what’s known as an atomic network. An atomic network is a small, secure network that can grow on its own. It’s the opposite approach to just launching a poorly-tested product into the ether and hoping people will flock to it.
Take the story of business communications service, Slack. Slack didn’t start out as Slack, but began life as a start-up called Tiny Speck. Its first product was a multiplayer game called Glitch. Tiny Speck had great hopes for Glitch, but it failed – badly. It was given disastrous reviews and was soon abandoned.
But, undeterred, the people at Tiny Speck turned to the little communication tool they’d been using while developing the game. It had been called various things, like “Frankentool,” “Honeycomb,” and “Chatly.io.” Eventually it became known as Slack.
To begin with, Slack was tested with friends of Tiny Speck. These included other start-ups, like Rdio, Wantful, and Cozy. Eventually, 45 companies signed up to use the product in all. This was the first, tiny, atomic network that is so important when getting around the cold start problem.
Those start-ups loved Slack – a service that served their needs perfectly. So they passed the news onto their friends. The rest, as we know, is history.
Sustaining a large network can be extremely difficult.
Let’s imagine that your tech product – let’s say, a new app that lets people find pet-sitters in their local area – has bridged the cold start problem and taken off. Network effects have kicked in, and you’re now sitting at the head of a massive global company that allows pet owners to go on vacation, safe in the knowledge that their kitten or labradoodle will be in safe hands.
There’s nothing more to do but sit back and reap the rewards, right? Wrong. It’s not that easy; it’s never that easy. As it grows, the network itself can present serious problems.
The first kind of problem relates to immediate growth. Not long after the company flies into escape velocity, its growth will likely hit a ceiling. This can be for a variety of reasons. It might be caused by a market becoming saturated for a particular product or service. Or it might be down to users tuning out from stale marketing channels.
Another problem, as the network grows, is the arrival of bad actors. It’s especially acute in social networks. Usenet – an online forum from the earliest days of the internet – is a perfect case study. Originally Usenet was a place where people could talk about anything from wine-making to philosophy, and it was free from spammers and trolls for years.
But then, sadly, in September 1993, the spammers and trolls arrived as the internet took off. Usenet became unusable. Its original context – as a place where users enjoyed in-depth discussion – collapsed.
These problems are intrinsic to large networks, and they have to be managed. With the first problem – remember, this problem is slowing growth – a company can find solutions by pushing for another growth cycle. By building new networks, they will soon enter escape velocity a second, third, even fourth time.
With the second problem – that’s the collapse of context in a network – you can look to modern communication services, like WhatsApp and iMessage, for an answer. Rather than allowing networks to become so large that the original context is lost, it’s possible to have smaller, self-contained bubbles, where select numbers of people can interact in the way they choose.
For an established company, building a moat is the best defence against upstart competitors.
Can you remember a time before Airbnb? Today, it’s the world’s biggest rental service. Everything about its success boils down to networks – from the way it scaled so quickly to the way it retains its number-one spot at the top of its market.
But it wasn’t always so secure. In Europe, it faced a strong competitor called Wimdu. Wimdu emerged in Berlin, in Germany, with lots of funding, hundreds of dedicated employees, and more traction than Airbnb in the places it operated.
So how did Airbnb come out on top?
Airbnb won out, because it built a bigger moat. And this is what any successful, mature company must do today.
But what the heck does it mean to build a moat?
As a company matures, it will, naturally, face competition. The trouble is, it’s often quite easy and inexpensive for smaller competitors to leverage the same network effects that made it successful in the first place. Things like viral, organic growth and increasing monetization as new users join.
So to stay ahead, a company must not only have the best brand, product, and partnerships, it must build a large, impenetrable network – or moat – around itself. It must compete at the level of the network above everything else.
This is exactly what Airbnb did in its battle with Wimdu. Rather than competing in the short-term by reducing price, Airbnb focused on the quality of the networks it was building in Europe.
While Wimdu was unselective with the kinds of properties that it listed, Airbnb wasn’t. Airbnb made sure that its users had the most wonderful rental experiences – way beyond their initial expectations. While those who booked on Wimdu might find themselves in overcrowded youth hostels, Airbnb made sure its users got what they were looking for.
And by doing this, alongside a targeted marketing campaign, Airbnb secured faithful networks across the continent. This was the moat that allowed it to triumph.
Any established company will face continuous challenges from smaller competitors. To survive today, it must constantly watch over its network, like those meerkats on guard against snakes and jackals. If it thinks the battle is won and relaxes, then a rival will sneak in and steal away its network bit by bit. Until one day, there is nothing left.
Here’s a recap of the main message:
The network effect is what happens when products or companies become more valuable as more people use them. A case in point is the telephone at the beginning of the twentieth century: without people on the other end of the phone, the technology was worthless. But as more people had phones installed, the network grew to become enormously powerful. To kick off the network effect, though, can be very challenging. This is called the cold start problem. To solve it, businesses must build small, sustainable networks, before scaling up. As they expand, companies will then have to deal with problems within the network effect itself – like hitting a growth ceiling and attracting bad actors. Finally, to win out in the long term, established companies must build a moat – a large, stable network.
About the author
Andrew Chen is a general partner at venture capital firm Andreessen Horowitz. Before that, he led the growth teams at Uber during their early years. He’s also a board member of fast-growing startups like Substack, Clubhouse, Z League, All Day Kitchens, Sleeper, Maven, and Reforge. He runs a popular blog, and his work has been featured in Wired, the Wall Street Journal, and the New York Times.
Andrew Chen is a general partner at Andreessen Horowitz, investing in early-stage consumer startups. He is a board member of fast-growing startups like Substack, Clubhouse, Z League, All Day Kitchens, Sleeper, Maven, and Reforge, and previously led the rider growth teams at Uber during their high-growth, pre-IPO years. He has a popular professional blog, and has been featured in Wired, the Wall Street Journal, and the New York Times. He holds a BS in applied mathematics from the University of Washington, where he graduated at the age of nineteen. He splits his time between San Francisco and Los Angeles.
Business, Money, Finance, Careers, Business Development, Entrepreneurship, Science, Technology, Leadership, Design, Management, Economics, Self-Help
Table of Contents
Part I: Network Effects
1. What’s a network effect, anyway?
2. A brief history
3. Cold Start Theory
Part II: Cold Start
4. Tiny Speck
5. Anti-network effects
6. The atomic network — Credit cards
7. The hard side — Wikipedia
8. Solve a hard problem — Tinder
9. The killer product — Zoom
10. Magic moments — Clubhouse
Part III: Tipping Point
12. Invite only — LinkedIn
13. Come for the tool — Instagram
14. Paying up for launch — Coupons
15. Flintstoning — Reddit
16. Always be hustlin’ — Uber
Part IV: Escape Velocity
18. The trio of forces
19. The engagement effect — Scurvy
20. The acquisition effect — PayPal
21. The economic effect — Credit bureaus
Part V: The Ceiling
23. Rocketship growth
24. Saturation — eBay
25. The law of shitty clickthroughs — Banner ads
26. When the network revolts — Uber
27. Eternal september — Usenet
28. Overcrowding — YouTube
Part VI: The Moat
30. Virtuous cycle, vicious cycle
31. Cherry picking — Craigslist
32. Big bang failures — Google+
33. Competing over the hard side — Lyft and Uber
34. Bundling — Microsoft
35. The future of network effects
A startup executive and investordraws on expertise developed at the premier venture capital firm Andreessen Horowitz and as an executive at Uber to address how tech’s most successful products have solved the dreaded “cold start problem”—by leveraging network effects to launch and scale toward billions of users.
Although software has become easier to build, launching and scaling new products and services remains difficult. Startups face daunting challenges entering the technology ecosystem, including stiff competition, copycats, and ineffective marketing channels. Teams launching new products must consider the advantages of “the network effect,” where a product or service’s value increases as more users engage with it. Apple, Google, Microsoft, and other tech giants utilize network effects, and most tech products incorporate them, whether they’re messaging apps, workplace collaboration tools, or marketplaces. Network effects provide a path for fledgling products to break through, attracting new users through viral growth and word of mouth.
Yet most entrepreneurs lack the vocabulary and context to describe them—much less understand the fundamental principles that drive the effect. What exactly are network effects? How do teams create and build them into their products? How do products compete in a market where every player has them? Andrew Chen draws on his experience and on interviews with the CEOs and founding teams of LinkedIn, Twitch, Zoom, Dropbox, Tinder, Uber, Airbnb, and Pinterest to offer unique insights in answering these questions. Chen also provides practical frameworks and principles that can be applied across products and industries.
The Cold Start Problem reveals what makes winning networks thrive, why some startups fail to successfully scale, and, most crucially, why products that create and compete using the network effect are vitally important today.
Read an Excerpt
Downloadable PDF preview of THE COLD START PROBLEM. Writing the first book, published by Harper Business, on viral growth, network effects, and competitive strategy. Interviewed some amazing CEOs, including Slack, Zoom, Clubhouse, Tinder, Twitch, Reddit, and others to put it together. Ties together many of the ideas on user growth, metrics, network effects, and product strategy.
“A practical handbook for entrepreneurs struggling with how to effectively apply what can be a devilishly tricky concept.” – Business Insider
I’ve read a preview copy of Andrew’s book, and I can tell you it’s very good. It’s full of real-life examples, actionable advice, and concrete takeaways. – Lenny Rachitsky, Startup Investor & Advisor, Author, Former Airbnb Growth PM
Excited for The Cold Start Problem, @andrewchen’s new book full of wisdom for starting and scaling. – Darius Contractor, Chief Product & Engineering Officer, Vendr; Former Airtable, Facebook, & Dropbox Product Executive
Extremely excited to see @andrewchen’s new book come out. Doing research on GTM strategies for network-driven companies (social networks, marketplaces, professional networks) was immensely fun, and I learned a lot through it. And now finally it’s here!! – Li Jin, Founder & General Partner, Atelier Ventures; Former a16z Partner
‘Andrew’s blog is an inexhaustible source of numbers and ideas on concepts that matter’ – Max Levchin, co-founder of PayPal
‘Here’s one message you want. Chen’s weekly newsletter offers thoughtful essays on startups and marketing’ – Wired
‘One of the best entrepreneurship blogs of all time’ – Eric Ries, author of The Lean Startup