howsocial.ru intends to be able to measure your social impact on the online world.

To achieve that, it combines the impact you have in a number of platforms in a single percentile index named CoSI (Collective Social Impact). At the moment you can submit your profile link for Twitter and Friendfeed and we take into account your public data to assess the combined effect of the proverbial ripples you make when you throw your pebble of contribution to these communities.

We make a point of adopting the same concept for all the platforms we include in our service in order to ensure - to the extent that it's possible - that the impact of a user we measure on one platform is comparable to and can be combined with the impact on another. As no two platforms are the same we adapt the information available to us from each platform to our model.

What we do for each platform

Our model takes into account and combines a number of factors. Clearly, the absolute number of your connections is an important parameter to the effects your contributions have to the online community. This is a simple metric shared by most platforms in one form or another whether they are social networks, micro-blogging or simply blogging.

Secondly, we include what portion out of the total of a certain user's connections tend to contribute or reiterate what that user has said compared to behaving like black holes: taking everything in but saying nothing as a result. Such a piece of information can be estimated by looking at the ratios of people they follow to the people that follow them for each connection.

Thirdly, once that user has submitted something to his sphere of influence it is important to measure what comes back from that action. And although it's computationally expensive and of dubious accuracy to do a text analysis across a variety of platforms to measure replies, one can come up with alternate metrics. The method we adopted is to count the time between the first and the last of a fixed (and usually predetermined by technical constraints) number of responses an action has elicited. Obviously, a brief response window translates into rapid and sizable effect while a long window corresponds to a rather flat reaction. Again, such an approach is useful for the user's contacts as well - in order to measure how well they do (on average) in creating discussions and attracting attention.

And finally the above can be repeated for the equivalents each platform has for favoriting and/or liking particular actions. This is similar to what Google does to evaluate a webpage's ranking by considering the incoming links it has as a vote of confidence, reliability or value. We take this approach and apply it to people.

Needless to say, each of the above methods we use to calculate a certain user's impact on the particular platform is distinct and different quantitatively. As such, we normalise and weight them accordingly to reflect that method's particular importance with respect to the others.

Putting all platforms together

Once a model is in place for the general functionality we can start adapting it to each particular platform. That way we can get the Social Impact index per platform.

Twitter, on which we based our initial concept, is straightforward to implement on. The number of followers and their followers is readily available information from the Twitter API. Similarly, the concept of responses to an action described earlier have a direct equivalent in @ replies whereas likes also appear in Twitter as 'starring' tweets.

Friendfeed, which built on the proven success of the Twitter model, is also directly implementable simply by replacing followers with subscribers, @ replies with comments and stars with likes.

Furthermore, in order to combine into a single Social Impact index all the platforms a certain user contributes to, that contribution has to be weighted with respect to that user's activity there. In other words, this means that we also take into account if that user is mostly active for example in only one of the services.

And of course, we need to take into account the difference in the overall popularity of our platforms (independent of particular users). For example, this means that Twitter which (at the time of writing) has an order of magnitude more users than Friendfeed is considered a more influential tool - and that is something our model also weighs in.

Disclaimer

We realise that howsocial.ru provides just one of the many possible ways to measure the impact of a person. The current description of our model aims to help our users to better understand and utilise howsocial.ru at the circumstances they consider it serves better. At the same time we are available for suggestions as to how it can be improved or clarifications about its details.