Web-Wide Data: "When visiting and interacting with
websites we share important data about ourselves:
Netflix knows some movies I watch,
Amazon knows some books I read, and
Last.fm knows some music I listen to.
While this data enables these services to provide additional value -
Amazon suggests books; Last.fm notifies me of concerts - there’s two major issues: (i) the data is not accessible by the user; (ii) the data is site-centric.
For example, despite Netflix knowing some movies I watch, they don’t know about my
movie activity across other sites. And there’s no easy way for me to let them know.
Therein lies the problem to site-centric datasets that aren’t user controlled: each site represents a fraction of our total web activity within a given vertical. Increasingly our interactions within a vertical are web-wide. For movies, we read reviews on
Rotten Tomatoes or
IMDB, watch trailers on
Apple, buy tickets on
Fandango, rent from Netflix, and buy discs from Amazon."