Object Lessons and Cautionary Tales: Articles About Big Entertainment Metadata

by Chad Beer

Working at the enterprise level with systems aimed at managing the information and assets of a single company, we feel the weight of user expectations honed by navigation and search on experiences on the web. Well-designed taxonomies and metadata models pave the road for intuitive search and navigation for site after site.  The e-commerce search experience has generally become excellent. Entertainment services like iTunes and Netflix are just as successful at not only helping us find what we want but also predicting what that will be. The bar has been set very high.

These expectations are challenging but instructive.  Enterprise level administrators of Digital Asset Management systems, Marketing Resource Management systems, or other information management solutions don’t have the seemingly limitless human and financial resources available to web and entertainment giants. Add limitations to time and bandwidth and competing business priorities, and it’s easy to understand why enterprises won’t or can’t develop solutions as comprehensive and sophisticated as iTunes, Amazon, or Google.  Fortunately there are plenty of tools available to meet the information, asset, and content management needs faced by most companies.  Technology solutions are evolving in complexity, and industry best practices are becoming more and more sophisticated. We don’t all need to become giants to achieve great things.  That said, we have to model ourselves after somebody.  Users (all of us) expect greatness, so we continue to strive for it while tending to the realities of other demands.  We continue to learn from the accomplishments of today’s web based entertainment icons.

A recent article in The Atlantic drives this point home and inspired an informal survey of articles about Big Entertainment metadata.  Below is a brief list of highlights from this survey.  Each offers a peek under the hood at the metadata strategies driving these services.  These tales offer us lessons and inspiration, and point out that the challenges of Information Management are universal no matter how big or small you are.

"The Tragedy of iTunes and Classical Music", The Atlantic, July 28, 2015

This article was inspired by a discussion of what some iTunes users consider a weakness of the iTunes metadata model: a lack of fields that allow for detailed classification that meets the needs of classical music enthusiasts.  It presents a lesson about understanding the needs of all users when designing a metadata model, accounting for the peculiarities of all asset types, and perhaps most importantly recognizing the customer’s demand for that classification.

"How Netflix Reverse Engineered Hollywood", The Atlantic, January 2, 2014

Another article by The Atlantic, offering an in-depth look at Netflix’s predictive movie genres.  This is a compelling view into the depth and variety in their schema and data structure. The story it tells emphasizes the degree to which human (not automated) insight and intelligence still underpin the analysis of content for developing categorization strategies and metadata schema.  This is also a surprisingly entertaining read about the development of a metadata model.

"The New York Times Built a Robot to Help Make Article Tagging Easier", Nieman Lab, July 30, 2015

This is a fascinating article about an automatic tagging tool developed by The New York Times.  While many of us don’t have the resources for such robust custom-built solutions, this tale addresses puzzles of workflow integration and efficiencies, including the role of content creators in the tagging workflow.

"NetflixGraph Metadata Library: An Optimization Case Study", Techblog.Netflix.com, January 18, 2013

A deep-dive look at a project undertaken by Netflix to reduce the memory load of their genre classifications.  This look under-the-hood is technically focused but very accessible in its discussion of the information architecture behind these tags and their relationships.

Reading these articles, anyone working with metadata can take comfort in the knowledge that even the big guys face the same challenges we all do. Better than providing solace, however, they drive home the real value of thoughtfully designed metadata models and workflows, and can provide relatable anecdotal support for the efforts of Information Management professionals everywhere.

Chad Beer is a Manager at Optimity Advisors.