Review of Content Policy

Anyone who has attempted to manage content on a large intranet website will know that if you give users an easy way to contribute they will create a lot of content.  Also, if you don't police a strict content policy most of that content will be of a very low quality, but there will be the a small proportion of gems.  With no information architecture, there will be masses of duplication and navigation will be reduced to matching keywords or phrases.  So with over a billion potential users, it looks like Knoll will become a bohemoth version of the classic failed intranet project that didn't get its content policy in order - a giant mess that would benefit from having most of its content culled.  Youtube is in a similar position, but Youtube is there to entertain, not to inform.

 

Obviously Google must have thought about this.  When an intranet website becomes a content dumping ground, the webmaster has a real problem and needs to make tough decisions about locating and deleting poor content whilst being sensitive to internal politics.  Google doesn't have that problem.  Firstly, there is no internal politics so they can get rid of what they like. Secondly they don't actually have to delete poor content, they can just make it impossible for the public to find. Given their existing experience in ranking content, identifiying poor content automatically should be easy, particularly as content is rated by users, and detecting duplicate content is already built into knol - see http://groups.google.com/group/knol-users/browse_thread/thread/56670af4eb52c236.

 

So, this is how I see Knol could potentially work in the future, if Google wants to acheive the aim of a Knol being the first thing people turn to when they want to know about a certain topic. The user does a search on Google for, let's say, the search term 'Albert Einstein'.  Knol by that time contains dozens of articles about Einstein, most of them terrible. Google's ranking technology combined with the ratings system knows that one article is better than the others, so it's presented as the first result in Google.  The rest of the matches on the results page could come from other sources, rather than revealing dupicate articles in Knol, which appear hundreds of results pages later.

 

Obviously the process would need to be more subtle than that to ensure that equally good articles with oposing perspectives got equal voice, and that top ranked pages don't become stale over time. I expect google's enormous resource of boffins will be able to crack the problem, if they haven't already.