Search is hard – Google is truly gifted

Search is just hard – that is it.  It is hard because all the content we search has little or no meta-data of any use.  Managing meta-data is an uphill battle.  Meta-data is designed by and for professionals searchers – and often is out of synch with a users’ context.  As you have guessed I’m in the throws of a search project.  Progress is steady, but the content is nightmare.  We have created enough scripts to analyse the trajectory of a spaceshuttle’s flight- but in the end it is the content clean up or rather the state of the content that will determine if we succeed.

A promising new tool we are considering is Baynote.  Part of our requirements include fairly extensive reporting – the purpose of which is to manage the search features such as best bets, spell check, sponsored links and are also to learn what our users are and are not searching for.  Baynote meets these requirements nicely by heuristically building contexts around ’search hits’ based on the most fundamental Web 2.0 principle - the wisdom of crowds.  In Baynote a hit gets a better ranking based on the terms searched by others who were seemingly, (heuristically speaking) satisfied with a certain search result;  a result which matched their search term or phrase.  Technology rarely wows me but this tool is seems to be onto something.

Any search or statisical experts will have assumed the reference to Bayesian Probability - and I believe they are correct, although oddly there is no reference to it in Baynote’s collateral – I would have thought they would pay homage to Thomas Bayes but they don’t seem to…!?  Anyway – creating search reporting is one thing, analysing and making decisions on the reports is an entirely different issue -Baynote seems to handily solve both of these issues – by heuristically creating user context around searching.    I’m excited to see how this interesting search product unfolds in the market….