Michael Daconta | Metadata: The stuff of web 2.0

Reality Check'commentary

In a recent article on CNet's news site, the head of the world's largest game studio expressed his unhappiness about the average scores for the company's video game titles listed on Metacritic.org.

Why would a video game mogul fret over scores from one Web site? Because of the power of metadata.

Metacritic.org combines reviews and scores from across the Web and boils them down to a single number for movies, DVDs, music and video games.

The act of boiling down those divergent rating schemes into a single, easily comparable number is the secret sauce that gives Metacritic.org extraordinary clout in the video game industry.

This site's approach to ranking is an example of a type of metadata description that measures the degree of a characteristic along a continuum.

Designing degree characteristics is an advanced skill by which a modeler defines a scale and its key inflection points.

And it's one that has a variety of applications for adding value to government data.

For example, measuring the freshness of a data item is more complex than just calculating a simple elapsed time since its creation.

Though a measure of elapsed time would be accurate, it does not provide that magical, boiled-down view that saves me time and makes me more productive by making it evident whether the freshness score is acceptable.

Carrying this example further, the endpoints must make immediate sense for the organization.

This point is often misunderstood and sometimes even used as an argument against metadata. Critics argue that the real world is too messy to be modeled. That's a ridiculous argument and here's why: Metadata is not supposed to be reality, just as a map is not the terrain. You don't need to know every nook and cranny to turn left or right. Thus, effective metadata design focuses your data on what the consumer wants.

For example, 'current' could mean within an hour, and 'too old' could mean older than one year. Thus we define the scale and inflection points using our intimate knowledge of how the organization does business to boil down complex reality into manageable action thresholds.

This skill is exactly what Web 2.0 uses to turn its crowds into productive metadata collectors. Why do they do this? Rankings, ratings and reviews are the mother's milk of Web 2.0.

Sites such as Amazon, Netflix, iTunes, Edmunds and Digg put us to work every day, asking for feedback and creating metadata that makes those sites more valuable.

People don't do this out of charity. They do it because they understand that a small input from many people equals a big payback for the entire community.

In simpler terms, these sites use metadata because it works.

Another example of such pragmatism is metadata in the link titles of news aggregator sites such as Reddit and Digg.

For example, [NSFW] is a metadata tag ' standing for not safe for work ' that you will see on the title of a link you should not click on while at work. Other common examples of link-title metadata are [Image] and [Graph]. I recently proposed a standard set of rules for this type of metadata.

What can government organizations learn from the rise of Web 2.0 metadata? The most obvious lesson is that the value of data is not inherent in the data itself.

Web 2.0 sites use metadata to make data useful, to make it relevant for a particular user at a particular time for a particular purpose.

Let's close with a specific example.

A list of contacts by itself is not very useful. But if I tell you it is an employee directory, it becomes a bit more useful. If I then organize it by department, by geographic region and by cross-cutting function, it becomes still more useful. If I carry it further by tagging contacts with seniority, professional affiliation and educational affiliation ' now you are getting metadata to enhance the value of that data for your audience, just like a Web 2.0 site.

Daconta (daconta@ oberonassociates.com) is the former metadata program manager at the Homeland Security Department and author of 'Information As Product: How to Deliver the Right Information to the Right Person at the Right Time.'

Reader Comments

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Please type the letters/numbers you see above