My latest commercial project: SenseArray

I’ve just been putting the finishing touches to the website for my latest project, its called SenseArray.  Its a stand-alone software application that companies can license, which allows them to easily integrate personalized recommendations into their service or website.  I’ve been working with several companies over the past few weeks and months that are deploying the SenseArray technology, and I’m now ready to cast a wider net.

The basic idea behind SenseArray is simple: the success of almost any website or service depends on your ability to figure out what your users want, and then give it to them as expeditiously as possible.  This can be the difference between success and failure, or between some success, and huge success.  The problem is that its hard to figure out what your users want, and often they all don’t want the same thing.

I’m not the first to try to tackle this problem, indeed, its not the first, nor the second time I’ve tackled this problem, as I developed two collaborative filters for Revver, and one for Thoof.

But building collaborative filters isn’t easy, it takes time, and a lot of trial and error.  I’ve done the hard work, and successfully navigated the minefield.  The end result is that SenseArray scores more than 5% better than Netflix’ own algorithm with the Netflix Prize dataset.

Furthermore, all collaborative filters suffer from a common problem, they all suck when it comes to new users, because they don’t know anything about those users yet.  SenseArray doesn’t suffer from this problem, because it can utilize all kinds of data about users, much of which is available right from the very first moment they visit a website (such as their IP address, their operating system, and the referring website).

Anyway, take a look at the new SenseArray website, let me know what you think, and tell your friends about it if you think it might be useful to them!

Leave a Reply

Your email address will not be published.