A common feature on many websites letting users contribute content is a list of the most interesting content on the website right now. It is often presented as highest rated content, most read content, or just most popular content (whatever that is…).
Is the rating or the view count really the best way of finding out what is more popular – or even more important: what is more interesting?
The fallacy of the most read criteria as an indicator for interestingness is its re-enforcing effect. Once something has reached most read status it will be highlighted and thus be read even more – making it harder for actual more interesting content to emerge on the list.
Other indicators of interestingness could be where the clicks are coming from, what is more commented and who comments, what and who marks something as favorite or rates it good, what tags does it have and is that something that is tagged often just right now, what is searched for, its incoming links or google page rank?
The fallacy of the rating, comments, or other user-contributed measures as indicators for interestingness is that it can be gamed. Users will try to “break the algorithm” for the purpose of self-promotion.
Even Twitter’s “Trending topics”, which relies on often mentioned words on tweets seem to have problems.
The most popular thing is not necessarily the best thing. The most views or most votes doesn’t always win.
So how do we measure interestingness? With one or more of the above mentioned methods – or even by other new and creative ways not mentioned above. Whatever you do, consider if it will be safe to rely solely on one method, as each potentially have severe pitfalls.
