Reverse Network Effects - What Large Networks Die From
For the success of any social network, the network effect is vital, which consists in the fact that the value of the network grows with an increase in the number of participants. According to Metcalf's empirical law , network utility is approximately equal to half the square of the number of participants. However, the network effect is not a guarantee of the cloudless existence of a large network. Metcalfe's law did not stop Facebook from getting around Friendster and MySpace.
Sangit Paul Chaudry, an entrepreneur, analyst and business consultant from Singapore, believes that the cause of death of many large networks is the so-called reverse network effects , due to which the value of the network does not grow, but decreases with an increase in the number of users. How do they work?
The usefulness of any network comes from three sources - connections, content, and influence (connection, content, clout):
Naturally, real networks to one degree or another rely on all three sources, and the most successful of them fully exploit each.
However, with the growth of network size, all three aspects can degrade - these are the inverse network effects.
The flow of new users can actually worsen the quality of communication between network participants - this can happen when a closed and little-known community suddenly becomes more open and popular. The flow of contact requests from unfamiliar people, spam and vandalism make communication on the network difficult. The signal is drowning in noise. To avoid this, it is necessary to create “friction”, which prevents an uncontrolled increase in the number of participants. Without such “friction,” the network may show explosive growth, but a recession will follow very quickly.
A typical example is the anonymous ChatRoulette video chat, which first attracted a large audience, but the anonymity and lack of need to register led to the fact that it was filled with psychos and exhibitionists, after which normal people ran away from there.
One of the reasons for the success of Facebook was precisely in the fact that at an early stage of its development, he was able to avoid this effect. Initially, only Harvard students and other elite universities could register there - this guaranteed a certain circle of communication and increased the value of the network due to the smaller, but not more, number of participants. However, over time, this effect began to create problems - on Facebook every day more and more advertising, spam, friendship requests are unknown from anyone and uninvited invitations to applications and games.
Content-based networks are subject to the reverse network effect of degraded content. With the growing number of users creating content, it is becoming increasingly difficult to find something worthwhile. The ease of content creation, which is very important for attracting a critical mass of users, is becoming an obstacle to growth.
There are two ways to deal with this effect - by tightening the requirements for content creators and creating personalization, recommendations, and search mechanisms.
The first method includes mechanisms for voting, karma, pre-moderation and restrictions on rights for beginners. Spam and inappropriate material can be deleted by administrators in response to complaints from participants or automatically hidden from view, receiving many negative reviews. Such a scheme is well known to any user of Habrahabr and, in one form or another, is present in any social network focused on content.
At the same time, the abundance of even quality content creates problems over time. It becomes difficult to find interesting materials, and personalization mechanisms are needed. Any successful project allows you to customize the event feed to your liking, often automatic recommendation mechanisms work.
Unfortunately, the reverse network effect cannot be completely defeated - a network that is too large becomes so demanding on server resources that the network owners will have to break their own filtering mechanisms and recommendations in order to earn enough money to maintain the infrastructure. Paid accounts appear from which less relevant content can be pushed for money, and the signal-to-noise ratio inevitably drops.
Making it difficult for beginners to enter and creating obstacles for creating content in order to avoid the first two reverse network effects, it is easy to fall under the influence of the third. The network can turn into a closed and conservative community with a rigid hierarchy in which it is almost impossible for a newcomer to break through to the top. Without an influx of fresh blood, the community sooner or later decays and dies.
If the influence of the participants is proportional to the quantity and quality of the content they created, with the growth of the network, beginners are less likely to compare with the “veterans”, since their advantage only grows with time. Therefore, in order to maintain sustainable growth, the network must remain sufficiently democratic, with the possibility of rapid “career take-off”.
Reverse network effects can easily bury any large network. An increase in the number of participants alone is not guaranteed to be successful. To compensate for these effects, it is necessary to achieve a balance between conflicting requirements:
The reverse network effects cannot be defeated completely, and therefore large, well-established networks will periodically die under their own weight, making room for new ones. There is always a chance to create a “second Facebook”.
Sangit Paul Chaudry, an entrepreneur, analyst and business consultant from Singapore, believes that the cause of death of many large networks is the so-called reverse network effects , due to which the value of the network does not grow, but decreases with an increase in the number of users. How do they work?
The usefulness of any network comes from three sources - connections, content, and influence (connection, content, clout):
- Connections allow participants to communicate, find new people, maintain and develop a network of contacts. This source of value plays a primary role for telephone networks, Skype, dating sites.
- Content is the main source of value for networks such as YouTube, Flickr, StackOverflow.
- The influence allows advanced and professional users to gain and maintain a constant audience of fans and subscribers - this aspect is very pronounced on Twitter or LiveJournal.
Naturally, real networks to one degree or another rely on all three sources, and the most successful of them fully exploit each.
However, with the growth of network size, all three aspects can degrade - these are the inverse network effects.
Communication
The flow of new users can actually worsen the quality of communication between network participants - this can happen when a closed and little-known community suddenly becomes more open and popular. The flow of contact requests from unfamiliar people, spam and vandalism make communication on the network difficult. The signal is drowning in noise. To avoid this, it is necessary to create “friction”, which prevents an uncontrolled increase in the number of participants. Without such “friction,” the network may show explosive growth, but a recession will follow very quickly.
A typical example is the anonymous ChatRoulette video chat, which first attracted a large audience, but the anonymity and lack of need to register led to the fact that it was filled with psychos and exhibitionists, after which normal people ran away from there.
One of the reasons for the success of Facebook was precisely in the fact that at an early stage of its development, he was able to avoid this effect. Initially, only Harvard students and other elite universities could register there - this guaranteed a certain circle of communication and increased the value of the network due to the smaller, but not more, number of participants. However, over time, this effect began to create problems - on Facebook every day more and more advertising, spam, friendship requests are unknown from anyone and uninvited invitations to applications and games.
Content
Content-based networks are subject to the reverse network effect of degraded content. With the growing number of users creating content, it is becoming increasingly difficult to find something worthwhile. The ease of content creation, which is very important for attracting a critical mass of users, is becoming an obstacle to growth.
There are two ways to deal with this effect - by tightening the requirements for content creators and creating personalization, recommendations, and search mechanisms.
The first method includes mechanisms for voting, karma, pre-moderation and restrictions on rights for beginners. Spam and inappropriate material can be deleted by administrators in response to complaints from participants or automatically hidden from view, receiving many negative reviews. Such a scheme is well known to any user of Habrahabr and, in one form or another, is present in any social network focused on content.
At the same time, the abundance of even quality content creates problems over time. It becomes difficult to find interesting materials, and personalization mechanisms are needed. Any successful project allows you to customize the event feed to your liking, often automatic recommendation mechanisms work.
Unfortunately, the reverse network effect cannot be completely defeated - a network that is too large becomes so demanding on server resources that the network owners will have to break their own filtering mechanisms and recommendations in order to earn enough money to maintain the infrastructure. Paid accounts appear from which less relevant content can be pushed for money, and the signal-to-noise ratio inevitably drops.
Influence
Making it difficult for beginners to enter and creating obstacles for creating content in order to avoid the first two reverse network effects, it is easy to fall under the influence of the third. The network can turn into a closed and conservative community with a rigid hierarchy in which it is almost impossible for a newcomer to break through to the top. Without an influx of fresh blood, the community sooner or later decays and dies.
If the influence of the participants is proportional to the quantity and quality of the content they created, with the growth of the network, beginners are less likely to compare with the “veterans”, since their advantage only grows with time. Therefore, in order to maintain sustainable growth, the network must remain sufficiently democratic, with the possibility of rapid “career take-off”.
Summary
Reverse network effects can easily bury any large network. An increase in the number of participants alone is not guaranteed to be successful. To compensate for these effects, it is necessary to achieve a balance between conflicting requirements:
- The network should create some “friction” at the entrance, preventing the uncontrolled influx of participants
- A well-scalable content moderation and rating system should be created.
- Convenient and effective referral and personalization mechanisms must exist.
- You cannot allow hazing and the emergence of a rigid hierarchy.
The reverse network effects cannot be defeated completely, and therefore large, well-established networks will periodically die under their own weight, making room for new ones. There is always a chance to create a “second Facebook”.
