Community By The Numbers - Part 1 - Active Users

This is the first of hopefully a few posts about different stats that social networks, investors, and entrepreneurs should pay attention to. While a lot of the value in a social network is subjective, I personally believe there is a lot to be said for close tracking and analyzation of an entire range of statistics and numbers.

For the first post, I wanted to discuss the most basic, but probably most important statistic, Active Users. Many social networks publicize/boast how many total registered users they have. Unfortunately, this is the most misleading statistic of them all. There are many sites out there that are good at driving registration, but horrible at keeping users involved on a regular basis (i.e. Classmates). In addition, over time, registration numbers naturally rise, but active users can drop dramatically. For example, old timer Friendster, who advertises around 36 million members in fact only has a very small percentage that still use the site (my personal estimate would be around 1 million).

Bottom line…you can have 150 million registered members, but if none of them are active, then the community is worthless.

So what defines an “active user”? Personally I think there are multiple levels of an active user involving the type of activity on a site (blog posting, uploading pictures, adding events, etc.). But for the purpose of this first post, the most general, and best place to start is to look at login activity.

There are a few basic statistics to look at in this category:

Percentage of users who have logged in within the past day/week/two weeks/month

This gives a very general idea of how often a user is returning to the site. The statistic gets far less interesting/valuable beyond one month. If a user only logs in every couple of months, then I don’t consider them an active user. Most communities require active participation to make them truly worthwhile. A person who logs in only 6 times a year, isn’t of much value.

Average logins per user per day/week/month

This is an important statistic to be compared with the one above. For example, a community that had 50% of its users login within the past month, with an average login of 10 times per user within the past month is much different and better then a community who has had 75% of its users login within the past month, but only with an average of 2 logins per user.

IMPORTANT!!!

When calculating the above statistics, all users who have registered in the past month should NOT be included in the outcome. Including them, skews the statistic upward. If they registered within the past month, they will of course had logged in. In addition, many communities see a fall out rate over a certain period of time that is much larger in the first month. For example, many forums see users register, make one or two postings within the first month, but then never come back. These users should not be considered “active users.”

Flaws/Considerations

1. Login statistics can be skewed downards due to a browsers treatment of session and cookie variables. Many times a user can be cookied and returns to a site automatically being logged in. Even though they may have returned multiple unique times, it will only look like one login.

2. The reason this will be a series of posts, is because the analysis of an active user can’t stop here. Statistics based on logins will include “lurkers”, or users who login but don’t contribute content. While these users are still valuable because they generate page views and unique visits, a community must have a critical mass of contributers in order to be valuable.

[tags]Social Networks, Social Networking, Online Communities, Statistics, Friendster, Active Users, Brian Balfour[/tags]