Network effects is such a thrown around concept especially in pitch meetings. It’s almost become a cliche, an answer to every question, a silver bullet to any problem. But even though the term is used excessively, few understand the depth of what it truly means. Andrew Chen’s book, “The Cold Start Problem” is a great explainer to the concept.
How do you define “Network effects”?
In simple terms, a network effect is defined as, what happens when products get valuable as more people use them. The network is defined by the people who use the product to interact with other users for communication, commerce etc. The effect part describes how value increases as more people start using the product. For Uber, the more users join the app, the easier it gets for drivers to find someone for a ride. This fills in their time and increases their cash flow. This encourages new drivers to join the app, eventually creating a stable network.
Metcalfe’s law comes up as a core pillar in the study of network effects, popularised mostly during the dot-com era also referenced as the “Dot-com” bubble. The law states that “The systemic value of compatibly communicating devices grows as a square of their number” i.e. each time a user joins an app with a network behind it, the value of the app increases to n². So based on the law, for a network of 100 nodes that doubles to 200, the value of the network quadruples. Now, anyone who has built and scaled a networked product will show strong dissent to the statement. This law unfortunately has not aged well and is painfully irrelevant. It conveniently ignores the different phases in building a network, the quality of user engagement and the multi-sidedness of many networks. In-short, it looks good on paper but fails the real life messiness.
Is there a law that better suits Network effects ?
The book calls this the Meerkats law. There are so many animals that benefit from living in a pack, similar to having more nodes in a network. Be it meerkats. Goldfishes, penguins, bees or sardines, they benefit from living in a group. The ecological term for that group size number is called “Allee threshold” — where the group of animals above that threshold would be safer and grow faster as a population. However if for whatever reason the population of these animals declines beyond this threshold, the benefits of the pack quickly go away, making them susceptible to collapse. On the other hand, if the population grows too quickly, then overpopulation negates the advantages of the pack which causes their population to plateau. This is because there are limited resources that can support a finite population often called “carrying capacity.”
Let’s look at this in the context of technology products. If a messaging app doesn’t have enough people in it, some users will delete it. As the user base shrinks it becomes more likely that each user will leave, ultimately causing inactivity and the collapse of the network. But, if the app scales above a certain number, people do find other users to chat with which will encourage them to bring on more users and the population keeps growing. This threshold is called the tipping point. But like anything, this can’t last forever. Overcrowding can cause too many messages for users which is distracting and draining or too much content on the feed or too many choices that finding the right thing becomes exhausting.
Taking another example from Uber: When there are very few drivers in the city, it takes a long time to get a ride. When the ETA is thirty minutes to get a ride, the value to the user is essentially zero. The conversion rates are low causing the drivers to also leave the app eventually leading the network to a collapse. But when there are enough drivers on the platform i.e. above the tipping point, things start to work . The riders start to get cars in 15 minutes, then 5 minutes. Eventually though the value of network plateaus — there’s diminishing returns to having more density of drivers. For a user it doesn’t make a world of a difference if the car arrives in 4 minutes versus 2 minutes. As the network grows to become overcrowded, it becomes rather inconvenient. Imagine the wait time for a ride is essentially zero and the driver is just outside, your last minute plans of wearing shoes, putting things in your purse, locking the door and going downstairs is taken away from you.
How do you identify if a product has a successful network effect ?
A successful network effect requires both a network and a product. Any network is useful when you have someone to connect with. So the first question is, does the product have a network i.e. does it connect people with each other ? Then the second question is, does the ability to acquire new users or to become stickier or to monetize become even stronger as its network grows larger ? If the answer to both the questions is a “yes” for a product, then the product is one that classifies as the one with successful network effects.
There are 5 stages to fully harness the power of network effects:
Stage 1 — The cold start problem
Why do most networks fail ? It is because solving a cold start problem requires getting all the right users and right content on the same network at the same time. You need to find an approach that focuses on building an atomic network which by definition is the smallest possible network that is stable and can grow on its own. For Zoom the network size can be just 2 people whereas for Uber it may be 100s of drivers in an area. An important question to ask is who are the first most important users to get on to a network and why and how will you seed the initial network so that it grows in the way you want ?
Stage 2 — Tipping point
It takes an enormous amount of effort to build the first atomic network but having just one won’t create a multi-billion dollar business. You have to build many many networks to expand into the market. The good news is that each launch makes the next set of adjacent networks easier and easier to build until the momentum becomes unstoppable.
Stage 3 — Escape velocity
This stage is all about working fiercely to strengthen the network effects and sustain the growth. You should be using three underlying forces to do so. First, the Acquisition effect which lets products tap into the network to acquire valuable users at low cost. For example: Paypal’s viral referral programs or Linkedin’s recommendations for connecting. Second, the Engagement effect causes users to engage more with the product either through new use cases or features. For example: Instagram added stories and reels which causes users to spend more time on the app. Finally, the Economic effect which improves monetization and conversion rates as the network grows. Combining these three effects creates a flywheel, powering the network into billions of users.
Stage 4 — Hitting the ceiling
A rapidly growing network wants to grow as well as tear itself apart. This is when the network hits the ceiling and the growth stalls. This can be due to a variety of forces, starting with a high CAC, shitty click throughs, overcrowding, context collapse and many other negative forces. But by adding the right features to aid discovery, combat spam and increase relevance you can increase the carrying capacity for users. So the team then addresses these problems and another growth spurt emerges followed by another ceiling. Each time the team has to do this, it becomes tougher and tougher.
Stage 5 — The Moat
The final stage focuses on defending yourself against competitors. This is tricky when everyone is able to take advantage of the same dynamics. For example: Airbnb had to fight off its European competitor by competing on the quality of the network and not via traditional vectors like additional features or reduced pricing. These strategies have to differ depending on the size of the network. David can’t have the same strategy as Goliath and vice versa.
The subsequent parts will go into depth for each of the stages and it aims to provide a roadmap for any new product team. Stay tuned for Part 2.
