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

by Andrew Chen

XChange Leasing unfortunately lost $525 million and failed to professionalize the driver side of the market. The problem was, it attracted drivers highly motivated by money—usually a positive—but who didn’t have high credit scores for good reason. They often failed to make payments, using their Uber-provided car to drive for competitors and avoid the automatic deductions. They would steal the cars and sell them for, say, half price. They would drive for Lyft instead of Uber, as a way to avoid the automatic payment deductions—they would try to have their cake and eat it, too. Uber needed to organize a massive repossession effort to get the cars back, but it was too late—many had been sold illegally, some finding themselves as far away as Iraq and Afghanistan, GPS devices still attached and running. This is a colorful example of how scaling the supply side, when a lot of capital is involved, can be tricky.

Sometimes these large professionalized outfits can get really big. The scale of the winners depends on the difficulty of aggregation—it’s easier to become a larger video creator or app developer than a mega-Airbnb host. A big YouTube star can simply get there by consistently making videos, as some teenage stars do, whereas being a large-scale host requires millions of dollars of real estate as a requirement. A networked product like OpenTable has restaurants on their hard side, which are unlikely to grow to become giants within the platform, because food is inherently fragmented. It’s unlikely to see a huge chain of restaurants representing a huge percentage of the business. On the other hand, the social platforms often get very large and concentrated, with just a few winners at the top.

Unlike the innovators and early adopters, they might not stick around just because they think the product is cool or fun. They’re doing it to solve a problem—often to earn a living—and if a networked product can’t deliver, they’ll leave.

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. I distinctly believe that the former, done well, is the way to go. However, it’s incredibly hard to finesse—no wonder nearly every marketplace company has to deal with labor issues, and every app development platform has ended up competing with or booting off its app developers. Yet the benefits strongly outweigh the costs—it’s one of the key levers to break through the inevitable growth ceilings, and managing these dynamics well will extend the upside of the network.

At the start, the first atomic network for Usenet was at Duke University, where creators Jim Ellis and Tom Truscott were based. Then nearby University of North Carolina was added, followed by institutions like Bell Labs, Reed College, and University of Oklahoma. Because many of the early organizations on the network were universities, every September a new cohort of students would join Usenet as they started school. Over the ensuing months, they would learn the social norms, jargon, and culture, and then integrate into the community, or otherwise quit as people flamed them for not practicing “netiquette.” Many of the early Usenet participants knew each other from academic activities, creating a real-life bond that helped reinforce good behavior.

In September 1993, everything changed. AOL, the largest internet provider at the time, began a massive campaign to mail millions of CD-ROMs and floppy disks to consumers. Instead of a predictable, yearly spurt of users in September as students joined the Usenet network, millions of people began to join from all walks of life. This became a torrent that wouldn’t end.

A few months later, early Usenet pioneer Dave Fischer noted: September 1993 will go down in net history as th...

Your totally appropriate work-related message may not be appropriate if it’s being delivered on the weekend of your recipient.

Cofounder and CEO Steve Huffman wrote, in testimony submitted to the US House of Representatives, a description of Reddit’s philosophy: The way Reddit handles content moderation today is unique in the industry. We use a governance model akin to our own democracy—where everyone follows a set of rules, has the ability to vote and self-organize, and ultimately shares some responsibility for how the platform works. Users can accept or reject any piece of content. While most platforms have some version of the upvote function, an action to convey approval or agreement, we at Reddit see the additional downvote as equally important. The downvote is where community culture is made, through rejecting transgressive behavior or low-quality content.74

In the earliest days, there was very little content to organize. Getting to the first 1,000 videos was the hardest part of YouTube’s life, and we were just focused on that.

Combine this with a big launch and PR blitz across Europe, and the company built over 50,000 listings and was on its way to $130 million in gross revenue in its first year of operations. An article in 2012 described Wimdu’s progress:

All supply isn’t created equal. Wimdu’s top 10% of inventory was at the bottom 10% of Airbnb’s. They went for numbers, but recruited large property owners that managed hundreds of units in the form of low-end hostels. They went the easy route and would get 1000 listings via 10 property owners, but the experience for customers was disappointing. In the early days at Airbnb, we would always talk about creating a positive “Expectations Gap.” In the early days, when we were new, guests go in with low expectations, but then would be blown away by the experience. You need this high NPS to get people to tell their friends, and it makes hosts more likely to join too. Our competitors who took shortcuts couldn’t deliver here.

While Wimdu was able to quickly announce impressive numbers, the hard side of the network wasn’t fully formed or curated to reflect quality.

The smaller player and bigger player use different strategies. And the most intense competition tends to happen as networks compete over the most valuable users of one network to another—this is “Competing over the Hard Side.” The drivers, creators, and organizers that do the hard work in the network are incredibly valuable, and by shifting them over, a new network can rise while an incumbent crumbles.

Instead, it’s often the dynamics of the underlying network that make all the difference. Although the apps for DoorDash and Uber Eats look similar, the former’s focus on high-value, low-competition areas like suburbs and college towns made all the difference—today, DoorDash’s market share is 2x that of Uber Eats. Facebook built highly dense and engaged networks starting with college campuses versus Google+’s scattered launch that built weak, disconnected networks. Rarely in network-effects-driven categories does a product win based on features—instead, it’s a combination of harnessing network effects and building a product experience that reinforces those advantages.

It’s also not about whose network is bigger, a counterpoint to jargon like “first mover advantage.” In reality, you see examples of startups disrupting the big guys all the time. There’s been a slew of players who have “unbundled” parts of Craigslist, cherry-picking the best subcategories and making them apps unto themselves. Airbnb, Zillow, Thumbtack, Indeed, and many others fall into this category.

To figure out a response, it’s important to acknowledge a common myth about defensibility and moats: that somehow, network effects will magically help you fend off competition. This is a myth repeated again and again in startup pitch presentations to investors and entrepreneurs. It’s a lie that entrepreneurs tell to themselves.

Every dominant network might seem invincible, but the networks-of-networks framing argues that some parts of the network are weaker than others.

When an incumbent has its network cherry-picked, it’s extra painful along two dimensions: First, any network that is lost is unlikely to be regained, as anti-network effects kick back in. And second, the decline in market share hits doubly hard, which has implications for being able to raise money.

The Big Bang Launch is convenient for larger, more established companies as a method to launch new products because they often have distribution channels, huge engineering teams, and sales and marketing support. But counterintuitively, for networked products, this is often a trap. It’s exactly the wrong way to build a network, because a wide launch creates many, many weak networks that aren’t stable on their own. When companies don’t understand these nuances, it leads to disaster.

While this might superficially look like a large user base, it actually consisted of many weak networks that weren’t engaged, because most new users showed up and tried out the product as they read about it in the press, rather than hearing from their friends.

Early on, with fewer than 10,000 daily active users, Snapchat was already hitting 10 photos/day/user, several orders of magnitude more than equivalent services—showing it had mastered the hard side of the network. Twitch, Instagram, and TikTok innovated in a similar vector, giving creators new tools and media types to express themselves.

When networks are built bottom-up, they are more likely to be densely interconnected, and thus healthier and more engaged.

But these bonuses weren’t just any bonuses—they were targeted at quickly flipping over the most valuable drivers in the networks of Uber’s rivals, targeting so-called dual apping drivers that were active on multiple networks. They were given large, special bonuses that compelled them to stick to Uber, and every hour they drove was an hour that the other networks couldn’t utilize.

When a networked product takes competition seriously, it has to collect metrics to figure out the comparative position of all the players in the market. This in turn allows product teams to experiment and execute, while keeping an eye on results—it allows them to set goals, not only against their product’s success, but also their competitors’ declines.

Thus the quote from early Microsoft execs, “For every copy of VB we sell, there are ten copies of Windows that go along with it.”

Bundling has both helped and hurt companies implementing it over the years. Many of Microsoft’s security issues, instability, and less elegant interfaces can be traced directly back to the decision to focus on the needs of developers, particularly enterprise customers who made big investments in custom software that required reverse compatibility.