Network

Network Effects of Platform Businesses

Image courtesy of BrandDignityOpens in new window Networks are characterized by nodes interconnected by links. Networks can be physical, such as telecommunications networks where copper pairs are physically connected to premises, or logical—where the connection exists as a piece of information—such as when you are connected to friends on social media.

Networks have a range of properties and typologies. They can connect different types of nodes (heterogeneous), the links can be unidirectional or bidirectional, the nodes can be more or less meshed, etc.

These distinctions are important when we discuss ifeatures of platform businessesOpens in new window. For example, does the platformOpens in new window allow two-way communication between participants (e.g. friends on Facebook) or one-way communication (e.g. celebrities followed on Twitter), and what are the implications of these network rules?

For our purpose, platform businesses are powered by these communities of producers and users that can be helpfully modeled and explained using network concepts.

This can be counter-intuitive, since with traditional business models the opposite can be true, and the value of an exclusive luxury car does not increase—and may actually decrease—if another one is produced (or bought by a neighbor).

However, having more collectors (buyers and sellers) on eBay clearly increases the value of the overall platform since the sellers will add inventory and the buyers will provide liquidity to the platform and increase the number of transactions.

Unlike traditional businesses, platforms often exhibit network effects, and these have profound competitive implications. Figure X1 illustrates the number of relationships in a network with 2, 5 and 12 connected users.

Figure X1 Network Effects

The number of links in the network increases exponentially with one, 10 and 66 relationships (and therefore possible connections), respectively.

The relationship between the value of a network and its size, or between its utility and the numbers of transactions, has been modeled in a number of different contexts to help quantify network effects.

Network effects can be:

1. Direct. As in telecommunications, where the value of the network increases directly with the numbers of users connected to it. This is also what happens with the numbers of users connected to it. This is also what happens with “peer-to-peer” or social networks, where more users tend to benefit the entire platform.

2. Indirect. As in operating systems or app stores where the value of the network increases not only with the number of users (direct), but with the number of application developers attracted by the growing user base (indirect effect).
   
   Indirect network effects result in positive feedback loops across the platform as more developers creating more apps then attract more users, and more users make the platform more attractive to developers. Games consoles, operating systems and app stores are typical platforms exhibiting strong indirect network effects.

Mature platforms such as AlibabaOpens in new window, eBayOpens in new window or CraigslistOpens in new window benefit from huge indirect network effects.

More than 423 million active buyers make an average of 58 purchase a year on Alibaba; 25 million sellers with 1 billion listings attract 164 million active buyers on eBay; and Craigslist users post well over 80 million classified ads each month.

It is worth noting that network effects can also be induced by platforms and maximized or internalized through platform design and governance decisions. For example, features that allow user to provide feedback on their experiences add value to the platform. It increases direct network effects, as the more users participate by giving feedback on books on AmazonOpens in new window or flats on AirbnbOpens in new window, the better the information to other users.

By the same token, gaming platforms provide software development kits and a range of incentives for developers to select their platform over other competing ones. They do this to maximize indirect network effects since more and better games will attract more buyers of the game console itself.

Network effects are very important since they can represent a significant barrier to entry for competitors and therefore contribute to protecting a business. In certain circumstances, network effects may lead a platform to reach a critical massOpens in new window, and even in some markets tipOpens in new window to a “winner takes all” natural equilibrium, where a single business ends up serving the entire market.

Network effects are not just about the number of platform participants, but about their propensity to interact on the platform as well. Inactive users contribute less to network effects on a platform than active ones that participate frequently.

In real life, platforms often benefit from—and have to manage—a combination of different types of network effects. For example, Facebook started with direct network effects since Harvard students found it easier to connect to one another once their entire class had joined. These effects were later combined with indirect network effects with the addition of developers on the platform offering games and applications (e.g. Farmville, horoscopes, etc.).

After proving that it could scale its free services, Facebook started monetizing its business by inviting advertisers to the platform and allowing them to target very specific segments of platform participants.

It is worth noting that while network effects may lead to viral growth, the two concepts are distinct. A new book can become very successful and be bought/downloaded by lots of people very quickly (viral growth) but have no network effect (since new readers do not add value to previous ones).