The Cold Start Problem: How to Start and Scale Network Effects

by Andrew Chen

A successful network effect requires both a product and its network,

Naval Ravikant, a noted investor and entrepreneur, has observed: Humans are the networked species. Networks allow us to cooperate when we would otherwise go it alone. And networks allocate the fruits of our cooperation. Money is a network. Religion is a network. A corporation is a network. Roads are a network. Electricity is a network.

How many users does your network need before the product experience becomes good? The way to answer this is for companies to do an analysis on the size of their networks (on the X-axis) plotted against a set of important engagement metrics (on the Y-axis).

The solution to the Cold Start Problem starts by understanding how to add a small group of the right people, at the same time, using the product in the right way. Getting this initial network off the ground is the key, and the key is the “atomic network”—the smallest, stable network from which all other networks can be built.

The networked product should be launched in its simplest possible form—not fully featured—so that it has a dead simple value proposition. The target should be on building a tiny, atomic network—the smallest that could possibly make sense—and focus on building density, ignoring the objection of “market size.” And finally, the attitude in executing the launch should be “do whatever it takes”—even if it’s unscalable or unprofitable—to get momentum, without worrying about how to scale.

If you need hundreds of users on the same platform at once, company-wide coordination is needed. In this situation, a top-down enterprise sale that gets a company to mandate usage for everyone might work better.

The hardest problem to solve in creating the first atomic network is, well, attracting the hard side.

The answer is by building a product that solves an important need for the hard side.

Customers request endless features while competitors emerge with a longer list of functionality. Yet I observe that it’s a distinctive quality of networked products that they often do one thing well.

Networked products are fundamentally different from the typical product experience—they facilitate experiences that users have with each other, whereas traditional products emphasize how users interact with the software itself. They grow and succeed by adding more users, which create network effects, whereas traditional products grow by building better features and supporting more use cases.

This is a concept missing from traditional software products, where the richness and complexity of the experience depends on who’s on the network rather than the feature set.

For example, Snapchat lets you send photos to friends. Dropbox is a magical folder that syncs your files. Uber lets you hit a button to get a ride. Slack is a chat product for your coworkers. YouTube lets you watch videos. They’re very simple to use, but also easy to describe to your friends and coworkers as well.

The commonality between all these networked products is that they offered novel ways for people to interact, and over time, the network became the defensibility.

In other words, the ideal product to drive network effects combines both factors: The product idea itself should be as simple as possible—easily understandable by anyone as soon as they encounter it. And at the same time, it should simultaneously bring together a rich, complex, infinite network of users that is impossible to copy by competitors.

the most important reason to have invites in a product with network effects. Invite mechanics work like a copy-and-paste feature—if you start with a curated network, and give them invites, that network will copy itself over and over automatically.

while invites-only strategies are often described as taking advantage of the fear of missing out—FOMO—that’s not the core driver. When a new product carefully curates a network, followed by implementing invites so that it can copy and paste similar networks, then it can grow to take over the market.

imagine arriving at a large dinner party. A good friend welcomes you at the door, and as you step in, you see acquaintances, close friends, and a number of new people who’ve been curated to be absolutely fascinating. If that’s the ideal experience for a dinner guest, it’s also an apt metaphor for the best possible entry into a new product experience. Invite-only products can facilitate this, because every new user that signs up is already connected to at least one person—their inviter.

The connected bring in the more connected.

Part of this is picking the people on the network, but the other is teaching users how to interact with each other and enforcing these “rules” within the product. Quality begets quality. Networked products in these categories can be magnetic to consumers on the quality vector, which is why companies often handpick them—another form of invite-only tactics—to curate the initial network.

Just as creators of new products spend endless hours designing the experience, creators of networked products have an additional task: curating the right people so that the experience of a new member joining the community, marketplace, or other network is just right.

For networked products, the curation of the network—who’s on it, why they’re there, and how they interact with each other—is as important as its product design. Starting with a deliberate point of view on who’s best for your network will define its magnetism, culture, and ultimate trajectory.

Come for the tool, stay for the network” is one of the most famous strategies for launching and scaling networks. Start with a great “tool”—a product experience that is useful even for one user as a utility. Then, over time, pivot the users into a series of use cases that tap into a “network”—the part where you collaborate, share, communicate, or otherwise interact with other users.

Tool, network Create + share with others (Instagram, YouTube, G Suite, LinkedIn) Organize + collaborate with others (Pinterest, Asana, Dropbox) System of record + keep up to date with others (OpenTable, GitHub) Look up + contribute with others (Zillow, Glassdoor, Yelp)

It aids in the Tipping Point of a market because it is much easier to spread a tool than a network—after all, the latter suffers from the Cold Start Problem. Spread a tool far and wide, and then grow it properly, and it might start to build networks upon networks around the tool.

Coupons were invented in 1888 by John Pemberton and Asa Candler, cofounders of the Coca-Cola Company. The early coupons for Coca-Cola showed the classic cursive logo in the center, with the headline “This card entitles you to one glass of Coca-Cola” and along the sides, it encouraged you to go to any dispenser to redeem the coupon.

The common wisdom was, “If you have a chicken and egg problem—buy the chicken.”

Yes, it would be ideal to be able to grow a network with positive unit economics from day one, but sometimes it’s just not possible. Or it might be too slow. A new networked product is better off taking on the risk of subsidizing the network up front, and then improving the economics over time.

What looks like unprofitability in the short term might lead to dominance in the long term, if the market reaches a Tipping Point in your favor.

One of the more intriguing dilemmas within the Cold Start Problem is what happens when your initial network pulls you into a gray area.

When this happens, do you try to fix the loopholes or add more controls, potentially impacting the usability of the product? Is it even a good idea to scan through the contents of your users’

As a rough benchmark for evaluating startups at Andreessen Horowitz, I often look for a minimum baseline of 60 percent retention after day 1, 30 percent after day 7, and 15 percent at day 30, where the curve eventually levels out. It’s usually only the networked products that can exceed these numbers. That’s because networked products are unique in that they often become stickier over time, which cancels out the inevitable customer churn.

To amplify the Engagement Effect as it relates to reactivation, the key question to ask is, what is the experience of a churned user? If you’re inactive, what kinds of notifications are you getting from other users, and are they compelling enough to bring you back? Almost always, churned users don’t receive any communication at all. You can boost reactivation success rate significantly just by sending a weekly digest of the activity in a user’s network, or “Your friend X just joined” notifications. The other question to ask is, if a user wants to reactivate, how hard is it? At Uber, we had a staggering statistic where several million users were failing their password recovery per week—how do you make this much easier, and treat reactivation with the same seriousness as the sign-up process? While reactivation is typically not a concern for new products—they should focus on new users, since their count of lapsed users won’t be large—for products that have hit Escape Velocity, there will be a pool of many millions of users to draw upon. Reengaging them can become as big a growth lever as acquiring new users.

Perhaps it shouldn’t be a surprise that it was the PayPal alumni who mastered these dynamics, because payments are a naturally viral interaction—after all, there’s no better value proposition than someone wanting to give you money!

In fact, you had to sign up if you received money, and once that happened, you might send money to others, who might create even more signups.

Max Levchin described to me how this worked: Growth started to pick up after we gave $10 to every PayPal user who invited a friend, and also dropped $10 into the account of the person as soon as they signed up. The motivation to sign up was strong, but so was the motivation to continue inviting. People were already inviting each other, but this incentive supercharged viral growth. While it might have seemed like this would burn a lot of money, it increased engagement as well—we saw users send money back and forth within the network, and every time it happened, we recouped a little bit of the incentive. It was more cash efficient than you’d think.

What ad agencies have called “viral marketing” is usually taking consumer goods or services—with no network effects—and building an advertising campaign around them, anchored by a bit of shareable content. What I’m referring to—network-driven viral growth—is much more powerful. Networked products are unique because they can embed their viral growth into the product experience itself.

This is the Product/Network Duo at work again, where the product has features to attract people to the network, while the network brings more value to the product.

The real magic starts to happen as the viral factor starts to approach 1. After all, at a viral factor of 0.95, 1,000 users show up and then bring 950 of their friends, who will then bring 900, and so on—ultimately the amplification will be 20x. This is the mathematical expression of when a product “goes viral” and starts growing incredibly fast.

it’s a combination of big virality projects, lots of little optimizations, and strong retention that ultimately drives big viral factor numbers.

The cornerstone to amplifying the Acquisition network effect is to understand how one group of users taps into their respective networks to bring in the next group of users.

It may be easy to copy features, but it’s nearly impossible to copy a network. The winners end up with pricing power, and generate enormous economic benefits.

Both network saturation and market saturation can slow down growth. Market saturation caps the total number of people on the network—you eventually run out of companies that can sign up to your collaboration tools, or gamers for your new massively multiplayer game. But network saturation also caps the effectiveness of your engagement over time as the interconnections slowly diminish in incremental value. These two dynamics together drive the saturation effects that slow a network’s growth. The way to fight these inevitable forces is to constantly evolve your product, the target market, and the feature set—there’s no way around it.

The best practice is for products—whether they have network effects or not—to constantly layer on new channels.

The Law of Shitty Clickthroughs is best countered through improving network effects, not by spending more on marketing.

Yet the tyranny of the majority often overwhelms the scarcer part of the network.

A networked product generally wants to nudge its ecosystem toward professionalization, because it helps scale the hard side.

there are tremendous positives to professionalization within a network. Encouraging successful players to get even bigger can inject huge amounts of growth. Because the most successful members of the hard side are the most likely to professionalize, they are also the most likely to have expertise on how to be successful. They can hire and train employees, and help the company expand into other services and categories. They can raise investment and spend on big projects where others won’t, giving them an ability to scale quality and consistency on your network. Over time, these businesses often become a network’s best partners, with a deep symbiotic relationship.

A networked product is successful when people on a network know how to interact with each other.

The dilemma for networked products is stark. Embrace the professionalization of the hard side, and reap the benefits of increasing scale. Yet this leads to power concentration, and potential misalignments—though hopefully no protests outside your windows. Or reject the trend, and see the hard side struggle with scale.

Context collapse is what happens when too many networks are simultaneously brought together, and they collapse into one. It’s most problematic for social networks because it inhibits the behavior of content creators—the hard side—as they no longer are able to post photos that satisfy everyone in every context.