PDF Summary:The Cold Start Problem, by Andrew Chen

If there’s a secret to success in the tech industry, it’s a thorough understanding of the network effect: The more users you gain, the more valuable your product becomes and the easier it is to profit and grow. But if bigger equals better, is it even possible for a company starting from scratch to compete with established tech giants? In The Cold Start Problem, venture capitalist and former Uber executive Andrew Chen explains how to do exactly that: build a billion-dollar tech company from the ground up.

In this guide, we’ll explain how to build a startup that uses the network effect to compete with the established giants in your industry. In our commentary, we’ll explore tips from other books about scaling small businesses, including Blitzscaling and Crossing the Chasm. Additionally, we’ll provide further advice on how to market your product, incentivize your users to promote it, and ensure it remains profitable.

(continued)...

Chen asserts that once you hit a certain number of subnetworks, you’ll begin growing fast enough to become a legitimate competitor to the current industry leader—this is the Growth Explosion (as mentioned earlier, Chen refers to this stage as the “Tipping Point”).

Chen notes that while you’re trying to reach the Growth Explosion, the strategies you employ to build subnetworks don’t have to be scalable or cost-effective. After growth takes off, you can cease unprofitable strategies and recoup this early heavy investment by leveraging your large, profitable network. Although these unprofitable strategies can be risky, they can empower you to dominate an entire market extremely quickly.

Networks That Never Become Profitable

Chen states that in this stage of subnetwork replication, you’re allowed to be unprofitable for the sake of growth, and after you’ve built a network, you need to start turning major profits. However, in practice, many of the largest tech firms behave differently. Despite having hit their Growth Explosions years ago and boasting millions upon millions of users, many of today’s largest tech startups are still unprofitable, kept afloat primarily through outside investments.

For over 10 years, experts have been warning about the risk of a tech-industry economic bubble similar to the dot-com bubble of the late 90s. They’ve predicted that, at any moment, investors will pull out of tech ventures that aren’t earning enough profit, causing tech businesses to collapse all at once and sparking a major economic recession. However, unexpectedly, the investments have kept coming in. Investors continue to purchase billions of dollars worth of equity in tech startups, and the market continues to value these networks highly. It’s uncertain how long this trend will continue, but for now, it seems that you can afford to prioritize growth over profits at your tech startup indefinitely.

Chen recommends refining a single repeatable process for launching new subnetworks. In the next few sections, we’ll detail several possible options for this process: potential strategies you can use to spread your high-quality product to new subnetworks and then phase out after you’ve achieved growth. You can use one or more of these strategies at once—find a combination that works for your product.

Strategy #1: Pay Users to Use Your Product

One simple strategy Chen offers to quickly and effectively create subnetworks is to pay people to use your product. For example, freelance labor marketplace TaskRabbit offers its laborers a bonus of $25 after they complete their first task.

Another common way startups often “pay” users to use their product is to make it free for all users, with the option for users to pay for premium features. Each free user costs your company money to serve, but they spur enough growth in the number of premium users that it’s profitable overall.

(Shortform note: If you choose to embrace the massive upfront cost of paying users or making your product free to use, you’ll likely need to raise money from outside investors to fund it. To decide whether or not to seek investors, consider some pros and cons: Investors often have a wealth of business experience and can offer expert counsel to help you make wise decisions. However, once investors buy a stake in your company, they gain the right to vote on company decisions. Losing some control over these decisions can force you to compromise your original vision for the company.)

Strategy #2: Artificially Inflate Your Network

Another strategy Chen offers to spark early growth is for you and your employees to temporarily participate in the network to make it seem bigger than it is. This inflates the value of your product to attract users until your network grows big enough to sustain itself. For example, imagine you’re launching an app that lets users find people close to them to play board games with. Before you have enough users willing to host gaming meetups themselves, you might ask your employees to host several events to make your network seem active and attractive to new users.

(Shortform note: Although participating in a network to make it seem bigger to users can attract a sizable user base, take care not to misrepresent the size of your network to investors. Doing so undermines the trust necessary for a long-term relationship with investors, and if your exaggerations are extreme enough, they could constitute fraud.)

Strategy #3: Only Let Users Join by Invitation

A third strategy Chen recommends for launching new sub-networks is prohibiting users from joining unless they have an invitation from an existing user. The up-front cost of this strategy is limiting user growth, which at first seems like the opposite of what your business wants. However, this strategy encourages a more sustainable kind of growth—because users give invitations to people they know, they’re more likely to attract users who share similar backgrounds and interests. This makes them more likely to interact with one another in subnetworks, which is necessary to achieve subnetwork stability.

To illustrate, let’s return to our example of an app that helps people find board game groups. By making this app invite-only, it’ll spread among people who know each other, who are more likely to live in the same city and like the same games. This will create lasting, stable subnetworks. In contrast, if you advertise your app to everyone, the majority of users may join only to find that no one around them has the app or no one near them wants to play their favorite game. Because the app doesn’t function for them, they’ll delete the app immediately after joining, leaving you with no users at all.

(Shortform note: Brands with non-networked products have been experimenting with invite-only access to their wares, too. For instance, when the luxury activewear brand WONE first began, it forced customers to reach out and request approval directly from the WONE team before they could purchase any clothes. By communicating directly with every customer and getting to know many of them personally, the team turned their store into a community, creating a new kind of value. Consequently, shoppers spread the word about the brand, allowing their customer base to grow without any traditional paid marketing.)

Step #3: Streamline and Accelerate Growth

Once you’ve successfully launched a number of subnetworks, you’ll hit the Growth Explosion. This is when across all your subnetworks, you have enough users for the network effect to kick in with gusto, accelerating growth to the point where you can challenge the market leader in your industry. Chen states that at this stage, you can shift your focus away from manually starting new subnetworks and instead work to refine your product in ways that amplify the network effect. This allows you to grow your user base faster than you could ever imagine.

(Shortform note: Although Chen assumes that the readers of his book want to grow their company into an industry-dominating giant, you may want to keep your business intentionally small and simply stop at this step. In Anything You Want, Derek Sivers describes how he turned down weekly offers from firms who wanted to invest in his startup and help him expand. His rationale for doing so was that rapid expansion can sometimes distract you from giving the best possible product to your existing customers, which is the key to not only financial success but also fulfillment.)

To explain how to rapidly grow your company, we need to break down exactly how the network effect drives growth. Chen explains that having a bigger network makes it easier for you to 1) attract new users, 2) retain existing users, and 3) profit more from the users you have. We’ll discuss each of these aspects of the network effect in turn—it should be your goal to optimize each one.

Goal #1: Optimize Network Growth

First, Chen asserts that the larger your user base, the easier it is to attract more new users. If you have more users, more people will share your product with their personal contacts, creating naturally accelerating growth. This word-of-mouth advertising is more effective and more economical than traditional paid marketing.

To optimize your network growth, refine your product in ways that make users more likely to show it to others. Chen suggests implementing features in the product that naturally expand the network: For example, TikTok allows users to export and share videos outside of the app. The company places its logo in the exported video so that viewers know if they want more videos like that, they can join TikTok.

(Shortform note: To optimize network growth, it’s not enough to simply give users the ability to expand your network by using and sharing your product—you must also give them an incentive to do so. In Contagious, Jonah Berger describes different kinds of value you can offer to users for sharing your product. By making your product interesting or cool, you can give users social currency, making them seem interesting and cool for knowing about it. Similarly, by making your product scarce or exclusive, people will feel special to be included in its user base, giving them an incentive to brag about it to their friends.)

Goal #2: Optimize Network Preservation

Chen explains that the second positive aspect of the network effect is that the more users you have, the easier it is to make a product compelling enough for users to continue using it. This is because having more users means you have more user behavior to monitor—analyzing this data gives you the insights you need to make your product more alluring.

To optimize your network preservation, Chen recommends identifying which customers spend the most time on your product and then experimenting to discover which aspects of your product influence them to stick around. Use this information to refine your product and make it even more compelling.

To illustrate this idea, let’s return to our “board game app” scenario. By analyzing user data, you may discover that users who play two or more different games are 300% more likely to use the app for six months or more than users who play just one game. With this in mind, you could add a “discover new games” feature to the main menu of your app and track whether this encourages more users to play multiple games.

A Low-Budget Alternative: Conversations With Customers

If you want to refine your product to be more engaging but you don’t have the time or resources to analyze data and run experiments like Chen suggests, you can gain a wealth of relevant information from simply talking to customers face-to-face. In The Mom Test, Rob Fitzpatrick explains that customer conversations can be a remarkably reliable guide to making successful product development decisions.

Fitzpatrick recommends keeping your discussions focused on objective facts about your customers’ past behavior rather than on their opinions, preferences, or predictions, which they often misreport out of politeness or lack of self-awareness. For example, if your product is a social network, you could start by asking a customer, “When was the last time you opened our app?” If it’s been more than a week or two, signaling a lack of engagement, you could ask “What’s the last social networking app you’ve used?” and “Did you post something when you used it or just browse?”

The goal of these questions is to learn about your customers’ desires and the habitual way they currently fulfill those desires. Once you’ve done this, you can tailor your product to more consistently fulfill those desires, making it more engaging. In this example, you might discover that your users are more interested in browsing than they are posting, so you decide to develop a more engaging browsing experience.

Goal #3: Optimize Network Monetization

Third, Chen explains that the more users you have, the easier it is to make money from your product. Even ignoring the fact that more users means more paying customers, a large network is much more valuable to users than a small network, and users are willing to pay more for it as your network grows in size.

To optimize your network monetization, give your users the opportunity to pay extra for additional network-based features so that as your network grows, these premium features become more valuable. For instance, your board game app could allow group organizers to pay for more visibility on their postings. This premium feature becomes more valuable as the number of group seekers on the network increases, so the app leverages the increased network effect to get more users to pay.

Don’t Stress Out About Raising Your Prices

As your network advances and you refine your product, it becomes more valuable, so it’s perfectly reasonable you would charge more for it. However, as Chris Guillebeau describes in The $100 Startup, the owners of small startups often hesitate to raise their prices out of fear that their current customers will grow frustrated and quit buying from them. Guillebeau reassures his readers that customers have come to expect occasional price increases. The vast majority of the time, price increases can have an enormous impact on your profits while barely disturbing your customer base.

Guillebeau also recommends charging more than you think you should if you choose a tiered pricing model. He suggests not only offering customers a premium-tier option but also offering a super-premium tier that’s far more expensive than you would expect them to buy. This way, you not only make more money whenever anyone decides to invest in that tier, but you also leverage the anchor effect to make your standard premium-tier price feel like more of a deal in comparison.

Step #4: Push Through Negative Effects of Growth

If you’ve optimized your product to take advantage of all aspects of the network effect, your user base will skyrocket, and you’ll find yourself in a highly profitable—if not totally dominant—market position. However, Chen warns that at a certain point, extreme growth instigates a unique set of new problems. If you don’t take steps to solve these problems, they’ll degrade the quality of your product and halt your rapid growth.

(Shortform note: Because there will always be problems for you to solve, even after you’ve achieved “success” by growing your company into a market leader, trying to solve all the major problems your company faces will likely cause you to feel burned out. For this reason, in Clockwork, Mike Michalowicz recommends designing your business so your team can successfully run it without your direct input. This way, you can take vacations as necessary to preserve your physical and mental health and renew your motivation to build a great business.)

Let’s take a look at three problems uniquely inherent to large networks and explain how to resolve them to maintain growth.

Problem #1: Market Saturation

At a certain point, your network will reach a natural maximum size. Chen explains that this occurs when you’ve effectively captured the entire market: All the users that would feasibly use your product are already using it. This is called market saturation.

To solve this problem, Chen argues that you must innovate beyond your original idea, creating new features or products that do one of two things: Either target new kinds of users and integrate them into your existing network, or offer new paid features to the users you already have. Both strategies allow you to tap into new revenue streams and continue growing your company.

To illustrate, let’s return to our board game app example. Once everyone in your market (social board gamers) is using your app, you’ll need to expand into a new market. You could target a new kind of user by expanding your app into a tool for hobbyists of all kinds, or you could offer a new paid feature to your existing users—for instance, turning your app into a platform that lets users play board games with each other online.

Market Saturation Shouldn’t Be Your First Assumption

Take care not to make big, risky business decisions like targeting new users or introducing new features until you’re sure that your market is saturated. In How Brands Grow, Byron Sharp argues that often, entrepreneurs falsely assume they’ve become the leaders of a saturated market. They fail to recognize that they’re competing in a much bigger market than they think they are; consequently, they set sales goals that are too low for the market they’re selling in. This complacency prevents them from diagnosing flaws with their business that keep them from attracting more customers.

For example, if user growth for your language-learning app has stagnated, don’t jump to the conclusion that the language-learning market is all tapped out. Instead, consider that there may be obstacles dissuading users in some demographics from using your product. In this case, perhaps your app’s cartoonish design or beginner-focused advertising is keeping more serious language-learners from using your app.

To avoid making this mistake, Sharp recommends making your target audience as broad as possible—aim to recruit everyone who might find value in your product, across all demographics. Then, set aggressive growth targets until you’re the clear market leader. For example, if you’re running a social network, don’t assume the market is saturated until you have a user base as big as Facebook’s.

Problem #2: Community Dilution

According to Chen, another problem that occurs when your network reaches a certain size is community dilution: when certain kinds of users joining your network make it less valuable. This occurs when you have an influx of malicious users such as scammers, abusers, and trolls, but it can also happen simply when your product goes more mainstream. Users who are accustomed to interacting with a specific close community may leave the platform if a broader audience is likely to be there.

For example, this is arguably what happened to Facebook—when it fully saturated the market, many younger users didn’t want everything they posted to be seen by loose acquaintances and older relatives, so they fled to Instagram, Snapchat, and TikTok.

Depending on your product, one potential solution to community dilution is to allow users to create their own subnetworks. Then, if one subnetwork becomes diluted with people that ruin a user’s experience, they can create another one. For example, users can create private Snapchat stories and handpick which of their friends can see them.

You’ll also likely need tight moderation tools to regulate malicious users abusing your network. Since it’s difficult to manually moderate a large network, Chen suggests creating automated moderation tools as well as giving users the ability to flag malicious users for manual moderation.

(Shortform note: When regulating the malicious use of your product, be cautious and intentional about what content to remove. Many social networks have been criticized for heavy-handed moderation that some users deem to be censorship. To avoid this criticism, publish a transparent set of rules governing what actions are allowed, and make sure to apply them consistently across your network.)

Shifting User Motivations Can Dilute Your Community

Even if you give users the tools to create their own subnetworks, community dilution is arguably inevitable if users’ motivations to use the platform fundamentally change. For instance, some people argue that when a social network gets big enough, users begin to care more about building an audience than connecting with others on a personal level. This causes a drastic shift in the network’s culture that’s arguably impossible to undo.

Allowing users to create their own subnetworks may further exacerbate this shift, as it allows them to create forums purely devoted to their broadcasting. For example, although users can use Snapchat stories to keep in touch with their friends, influencers use them to create content channels with millions of followers. This new lucrative goal incentivizes users to care more about gaining a following than connecting with their friends, changing the culture of the network and resulting in dilution.

Problem #3: Too Much Content

Finally, Chen explains that when a network grows to a certain size, it often begins to offer more content than users are able to effectively sort through. This can degrade the user experience and cause users to leave. For example, if your board game app receives too many postings about different groups that play a wide range of different games, it becomes more difficult for users to find the few groups relevant to them. This also hurts group organizers, who find it harder to stand out amid the ocean of postings.

To fix this, connect users to the content they came to the network for. Chen highlights one solution in particular: Since you have access to a huge network of user data, you can create precise discovery algorithms to help users find the experience they’re looking for.

(Shortform note: Algorithms designed to solve the problem of “too much content” by giving users more of what they want to see unfortunately have some harmful side effects. Researchers both inside and outside of social media corporations have concluded that some networks’ discovery algorithms disproportionately spread political disinformation and hateful sentiments. Similarly, one study found that Facebook’s algorithm favors news outlets that align with the user’s existing beliefs, contributing to political polarization.)

Step #5: Ward Off Challengers

Once you’ve pushed through all the negative effects of growth, you’ll be a dominant market leader, earning massive profits. Chen explains that at this point, the network effect works in your favor: Potential competitors lack the benefits of your network that we discussed in Step #3, so it’s much more difficult for them to field a comparable product.

(Shortform note: Although your hard work has brought you a massive advantage over competitors in the form of network effect benefits, such an advantage can turn into a disadvantage if it negatively impacts your mindset. In Ego Is the Enemy, Ryan Holiday argues that success often curses people with the egocentric feelings of entitlement, paranoia, and an obsession with control. Any of these three mindsets can trigger irrational decisions and sabotage your success, so make a conscious effort to monitor your feelings and keep your ego in check.)

Chen notes that now, all that’s left to do is maintain your position as a market leader by closely monitoring your competitors and making sure they can’t convert substantial portions of your users. Adjust your user acquisition strategies and your product to convince users to choose you over competitors, then verify that your adjustments are effective. This requires a lot of data analysis—by studying user data you’ve collected as well as data purchased from external sources, you can get ideas for how to convert competitors’ users, and you can track the impact your strategies have on user behavior.

Avoid Making “Beating Competitors” Your Primary Goal

Chen frames this stage of the startup process as a cutthroat competition against your closest competitors. However, in The Infinite Game, Simon Sinek argues that you should see your competitors as friendly rivals and avoid making it your goal to thwart their success. Companies that obsess over beating their competitors have less motivated and less inspired employees than companies that focus on achieving a noble mission.

Sinek would likely approve of Chen’s suggestion to gather data on your competitors with the intention of improving your product. However, as opposed to Chen’s suggestion to use this data to steal competitors’ users, Sinek would argue that you benefit most by using this data primarily to learn from your competitors and use their ideas to offer a better service to users.

According to Sinek, learning from your competitors is much more difficult when your only mission is to “beat” them. Setting competition as your primary goal motivates you to focus on ways that your company is better than them, meaning you ignore their strengths and don’t learn from them.