Analysing network effects

What is the impact of the network effect on a startup? How to project your network effects over time? 

Li Jin and D’arcy Coolican wrote about the dynamics of network effects.  

There are three key factors that affect the dynamic of the network effect, namely:

a) the value proposition;

b) the users / inventory;

c) the competitive ecosystem;

The value proposition: what value proposition drives the NE (network effect) in your company? How strong / weak is it? How strong / weak may it become in the future? How will your value proposition evolve as you further develop your product and add new layers of product on top? Examples: ride sharing, social lending, social networks, decentralized platforms*. 

Users / inventory: not all users are created equal. And the type of users being added impact the NE as well. 

Differentiated versus commoditized inventory: a marketplace like Uber, where the expected quality of the service is not attached to the provider (driver) will have, down the road, less NE than a marketplace where the service provider can be differentiated. This happens because after hitting a given threshold (e.g. 5 minutes wait or less) the quantity of inventory becomes less valuable. Which brings us to the next topic.

Type of incremental user: there are three types of users: pollutants, neutral and contributors. E.g. twitter: think of a troll, a passive reader and a prolific content creator. Make sure your acquisition efforts are targeting the right user and your curation systems regularly clean your user pool of any bad actors that might be degrading the experience. 

Competitive ecosystem: your direct competitors and your substitutes affect your NE.

Network overlap: If someone has a network of users that overlaps with your business, they may enter your market. Sometimes substitutes are less obvious threats than direct competitors because the overlap is not as visible.

Switching costs: naturally, a low switching cost also lowers the NE.

Multi tenants: when a single platform isn’t enough to perform one job, the user needs to use multiple platforms (multi tenant), which usually leads to less differentiation and higher pressure on price. 

*(examples taken from the original article)