2 minute read / Sep 7, 2012 /
Using metrics to build communities
There is a Branch called Do platforms need to give users a number to optimize?
Most platforms do provide a number to optimize: Twitter followers, Facebook friends and so on. These metrics build liquidity in the platform as @satyap, @ev and @hunterwalk point out.
But there isn’t just one type of social metric. There are three. It’s important to distinguish the types of metrics a community can provide for users to optimize, because the community will rally around the metrics the community provides. The long term dynamics may not be the expected or desired behavior.
In my view, there are three types of social metrics:
Social validation metrics (Twitter followers, Facebook friends/followers, LinkedIn connections, Klout score) - are proxies for trust or respect within a community. They are useful as trust signals and for initial social graph growth.
Engagement metrics (YouTube views, app downloads, Quora views) - demonstrate the popularity of an item and activity of the community broadly. Seeing a post with 212 likes immediately provides a sense of a large user base. These metrics are the least valuable measure because they can be manipulated easily.
Karma metrics (retweets, likes, up votes, Hacker News Karma, content flags, spam indicators) - Karma metrics measure positive and negative contributions to the community by combining user centric metrics like influence with content quality metrics. Pinterest and Tumblr have gradations of behavior (like vs repin/reblog). Karma metrics enforce a social filter on content because when users retweet or like, they are expending their social capital to promote or endorse some content.
When to use each metric
At the outset of a community, the primary goal is to build a trusted network of contributors who extol and enforce the values of a community. Therefore, user metrics seem to be the most important.
As communities evolve, some sites pursue content metrics, but I think the most vibrant communities ultimately pursue and champion behavior metrics because they can be actively managed to encourage the desired user behavior. The best behavior metrics have some vagueness, some room to evolve in order to respond to changes in community behavior. For example, Facebook is trying to mitigate the fake like problem
Communities are living entities, like cities, that need to be cultivated, nurtured and maintained. Metric selection is an essential component of this evolution.