TechNayak

At the Big Thinkers conference by Andrew Tomkins of Yahoo, the word interestingness caught my attention. Though the idea of judging the interestingness of an object like a photo, document, video seems far fetched,  I nevertheless was fascinated by the possibilities of this new metric. I found this patent document filed by Yahoo for their interestingness concept in Flickr.

claim 1:  An apparatus for determining an interestingness rank for at least one media object, comprising: logic for accepting at least one metadatum concerning the at least one media object from at least one user; and logic for ranking the at least one media object based at least in part on the quantity of user-entered metadata concerning the at least one media object.

The apparatus of claim 1, wherein the user-entered metadata comprises a member of the group consisting of: tags, comments and annotations.

Well there are new web metrics being rolled out almost everyday and its only time, at least some of them, become mainstream. One of the strong contenders I feel is the interestingness metric. Based on the number of bookmarks, clicks , click throughs, views, tags, annotations, visits, visit length and many many more factors interestingness can be defined, although this may not be Flickr’s definition. Any piece of information like a music piece , a podcast, a blogpost, a video clip,  a picture  etc can be associated with interestingness - It simply gives the most popular item amongst a list of inventory. These items can also possess geographical information, time line and other information that can be used to extract essential patterns from these items.

The usage, could be a phenomenon on its own. Define a properly ranked algorithm and you could have your top selling items on your e commerce site. You could target geographically distributed audience with content of their choice and also give interesting timeline applications that let you see trends over time. Say for example you run a shopping portal, you could give a time line graph highlighting your most sold items over a period of time. This could also prove a big step in the analytics area where ROI and conversion from visitors to buyers play a significant role in everyday business and concepts like these could really boost your sales. Its still unsolved from a very broad sense, meaning there isn’t a standard set of practices to encompass such a wide variety of data and there isnt such capable systems that could mine patterns out of these collations. But its definitely interesting