One of our design goals for Operation Omar Popyoular is to ensure that our core data about individual content items (such as an album) is as correct as possible. Another design goal is to to attempt to enrich that core data with as much useful metadata as possible in order to enable additional functionality. As it turns out, these steps aren't exactly trivial. But they can produce unexpected opportunities, as we've just discovered. One tricky aspect of music metadata is how to correctly establish a release date for an album. Looking around at various types of music services you'll quickly see how a famous record from, say, 1973 will be listed as a release from, say, 1998. That would be the year of that particular release of that album.
This makes perfect sense to record companies. It makes very little sense to most consumers, and generally screws up simple yet important things like viewing a list of albums chronologically.
As it turns out, we've managed to do a pretty good job of resolving the release date/year commonly associated with a given album. And as it turns out, solving this particular challenge is important to the service which most likely will support us with a pilot implementation of Operation Omar Popyoular. So we isolated this particular task and pre-delivered a large number corrected release dates, with more to come.
As a new venture, we expected our business to change in unexpected ways along the way. This particular side effect was a pleasant one. We now know that our platform can work really well for pure metadata tasks, not just as a full-blown content discovery mechanism. We hope that means we can put it to work for a larger set of services with similar challenges.