I think one of the biggest hurdles to ranking and rating systems is how the data comes in. Imagine a two-dimensional histogram where Y is the number of ratings and all of the items are arranged on the X axis.
What I saw with Oink was this X axis (the number of items to be rated) growing faster than the Y axis (the number of ratings): in effect, noise growing faster than signal.
They weren't aggressive enough in filtering out duplicate entries and/or working with businesses to upload their catalogs onto Oink so users didn't have to add (and re-add) them. Not saying it was an easy task, but essentially this is the main challenge to such a product.
What I saw with Oink was this X axis (the number of items to be rated) growing faster than the Y axis (the number of ratings): in effect, noise growing faster than signal.
They weren't aggressive enough in filtering out duplicate entries and/or working with businesses to upload their catalogs onto Oink so users didn't have to add (and re-add) them. Not saying it was an easy task, but essentially this is the main challenge to such a product.