Why Many Would-Be Network Effect Companies Fail in Their Infancy

The “cold start problem” evolves into ever-shifting imbalances in supply and demand.

Written by Tony Yiu
Published on Dec. 09, 2020
Why Many Would-Be Network Effect Companies Fail in Their Infancy
Brand Studio Logo

A network effect is one of the most powerful forces in business. Once it gets going, it’s like a snowball that grows ever larger by sucking in more and more suppliers and customers. It’s also the force that keeps new would-be competitors at bay.

Many of today’s most successful companies exhibit massive network effects: Visa, Apple, Google, Facebook, Amazon, Microsoft, Square, Airbnb, etc.

A network effect is a feedback loop. In the business world, it occurs when there are at least two sides in a marketplace — app suppliers and app users; content makers and content viewers; restaurants and hungry people.

The power of the network effect stems from the fact that an increase on either side of the network makes the overall network more attractive. More restaurants means more selection for hungry people — this attracts more hungry people to the service. More hungry people means more potential customers for restaurants — this attracts more restaurants to the service. More of one side attracts more of the other, and, theoretically, this perpetuates a virtuous cycle of growth.

As the network keeps increasing in size, the business that created the network (and facilitates the transactions) monetizes it by taking a cut of every transaction.

So if network effects are so great, why doesn’t every business get one? It’s because it’s extremely hard. For every successful network, there are probably hundreds that attempted to do a similar thing but failed.

Read More From Tony YiuShould You Turn Off Your Predictive Model or Keep the Faith?


The Challenge of Starting a Network: The Cold Start Problem

It turns out that getting that snowball to roll even a little bit when a company is first starting out is very difficult. Imagine a food delivery network with almost no restaurants — how does such a network attract users? Or that same network with no hungry users — how does it attract restaurants?

This is known as the cold start problem. The same characteristics that make the network effect so powerful once scale is achieved are the ones that make it extremely vulnerable and hard to scale in the beginning.

The answer that most new companies arrive at is money. They attempt to use financial incentives and marketing to jumpstart one side of the network (or both): free trials, referral fees, discounts, etc. So it basically becomes a race against time and investors’ willingness to fund the cash burn. If the company reaches escape velocity in time and growth becomes more organic (the positive network effects kick in), everything is great. If not, then either bankruptcy or a cheap liquidation awaits.

There’s no easy solution to the cold start problem — otherwise it wouldn’t be the cause of death for so many once-promising companies. One smart way companies attempt to sweeten the deal is by providing peripheral benefits for joining the network — for example, a network attempting to serve restaurants on one side might provide analytics and order management to those restaurants as bonus benefits for signing up. But still, no matter what, at some point, the network has to grow big enough to make it worth everyone’s while.

Follow Tony’s WorkAlpha Beta Blog


A Dynamic Problem

The challenge of growing a network is also a dynamic one. As the network gradually scales, supply and demand imbalances can create major headaches.

Take Uber for example. In its early days, it was all about getting drivers on its platform. The value proposition back then was that, compared to taxis, Uber was cheaper and faster (in terms of wait time) for a ride. This meant that the supply of drivers in a city had to be high enough (and they had to be driving primarily for Uber) — so Uber started off by subsidizing its drivers with significant sign-on and incentive bonuses. But over time, the growth bottleneck moved from the supply side (drivers) to the demand side (riders). Past a certain point, further decreases in wait times met with diminishing returns on the rider side (no difference between waiting three minutes vs. five minutes for a ride).

In order to reach the lofty valuations (and market size) that Uber and its venture backers hoped to achieve, Uber not only needed to cannibalize the taxi market, but it also needed to (a) convince people that usually drove to start taking Ubers instead and (b) get people that already took Ubers to use the service even more. It needed growth, and growth was achieved by attracting more and more riders. That’s why all of a sudden you started hearing much more about how drivers couldn’t make a decent living driving for Uber. The subsidies and financial incentives that were once used to attract drivers and keep them happy were now directed toward growing the rider base.

Growing a network is a constant challenge: balancing the needs and satisfaction of all sides, ferocious competition as everyone rushes to scale (and aims to be the first to hit escape velocity), the massive cash burn, etc. Each problem in isolation can be enough to sink a business. But a company that seeks to build a network effect must constantly solve ever-evolving variations of all three over and over again. Like I said above, there are no obvious solutions. But here are a few suggestions:

  1. Focus on one side of the network in the beginning. Attempting to grow both sides simultaneously requires more effort, more people, and more money. Rather, if you can get one side of the network to par, then that will organically begin to attract the other side.
  2. Provide value beyond just what the network provides. If there are other ancillary benefits to be on the network (analytical tools, support of the creative process, etc.), that can help sweeten the deal and attract initial users.
  3. Focus on a highly engaged niche market to start instead of trying to be everything to everyone. Create a platform that caters to their specific needs and build your initial user base from there.
  4. Save your money and marketing for whichever side of the platform is harder to grow over the medium term.

Building a network effect from scratch is hard. But if you’re the last one standing, the rewards are immense.

Hiring Now
AdTech • Digital Media • Machine Learning • Marketing Tech • Software • Travel • Hospitality