Good metadata means good government
Last month, one of my Facebook friends shared a YouTube video of a local news exposé on how the Whole Foods supermarket chain was selling “organic” produce from China. Especially egregious was the selling of a product labeled “California Blend” vegetables that had a “Product of China” label on the back (in small print).
So when does “California” not mean “California” the place? The manufacturer’s answer was that “California Blend” was a style of mixing vegetables and did not refer to the place. Well then, the label should have distinguished this by adding the qualifier “style” as in “California-Style Blend” to avoid ambiguity to the consumer.
It is important to note that this exposé and soon-after lawsuit would not have been possible without the implementation of the Agriculture Department's Country of Origin Labeling rule that became effective for all covered commodities March 16, 2009. COOL is an example of both good metadata and good government. Today, another COOL controversy has erupted because orange juice from Brazil was found to contain a fungicide that is illegal in the United States.
4 search engine snags and their satisfying solutions
Improved image tagging aids warfighters' situational awareness
There are plenty of examples, similar to COOL, that demonstrate how smart organizations are designing and using the right metadata in their IT systems.
My favorite part of the movie “The Social Network" is when Mark Zuckerberg’s classmate asks him if a girl named Stephanie in his art history class has a boyfriend. Zuckerberg practically leaps out of his desk, races back to his dorm room and adds the key metadata field “Relationship Status” to the code for Facebook. In Facebook, the core datasets are status messages and photos, while all of the static profile information is metadata that describes whether and why you care about those data records.
Now let's move from describing status messages to describing movies. Netflix has revolutionized the distribution, viewing and recommendation of movies via good metadata. Netflix recommends movies based on key usage-based characteristics such as “cerebral war documentaries," “sentimental biographical father-son movies," “dark revenge action thrillers” and “feel-good romantic coming-of-age movies." These narrow, dynamically generated categories describe a user’s tastes and drastically improve the recommendations.
Finally, let’s move from describing movies to describing all products that can be shipped by the King of Metadata: Amazon.com. The diversity, fidelity and quantity of metadata about all of the company's products is impressive. A product page on Amazon has metadata sections such as “Customers Who Bought This Item Also Bought," “Frequently Bought Together," “What Other Items Do Customers Buy After Viewing This Item," “Tags Customers Associate with This Product” and “Customer Reviews."
Many customers, including me, can attest to the effectiveness of this type of metadata in either helping them select the right product or influencing them to buy additional products they had not planned on. In a nutshell, the right metadata works, and understanding how to design that metadata is the key challenge.
In my last book, “Information As Product," I included a chapter on designing metadata. That chapter explained the different types of metadata design techniques to include identification, static measurement, dynamic measurement, degree, categorization, relationships and commentary.
The key to metadata design is developing the best descriptive fields that increase the usage value of the data to the end user. By usage value, I mean those metadata fields (like relationship status in Facebook) that significantly drive and affect user behavior. These types of fields can be discovered by asking questions such as “How do our users get value out of our data?” and then walking backward from those answers to the fields that best distinguish the good from the bad (or the proverbial wheat from the chaff).
From COOL to Netflix, Amazon and Facebook, it is evident that smart organizations are learning how to leverage the right metadata, which in turn allows their users to discover and leverage the right data at the right time.