It can be complicated but this might simplify things:Data Inputs:
Store Price search feeds
Studio release calendar feeds
Manual entry (based on reviews, etc.)
We pick up new movies from any of those inputs. The primary criteria for whether a something is new is the movie's title and a release date if available. (Studios that are too fucking stupid to use unique movie titles make this task more difficult than it needs to be.)
New movie titles get assigned an ADT internal part number. (It's the sku= part of the URLs on the price search pages.) This process is manual for every new movie added.
We then take the various price search data feeds we get from the stores and associate their internal part numbers with our part number.
Some of the process of building the ADT->store association is done automatically, much of it is manual. We also manually associate directors, cast lists, categories and most studios. It's a shitload of work.
Just to give you an idea of the amount of data involved, we've already added 1368 new movies in 2007. For those 1368 movies there's 11278 performer associations too. The stores have provided us 123866 new items of which we have 89239 active and available in the price search. (We have 1,284,938 total items currently linked up to stores in the price search which means 6.9% have been added in 2007. We've added lots of new VOD options this year.)

[Link]