Getting Taxonomies and Metadata Right


      Bookmark and Share

BEST PRACTICES SERIES

We are all overwhelmed with the variety, velocity, and volume of information. It is becoming increasingly difficult to figure out where the information we want is. In one of my cloud systems? Outlook? A folder? Findability can seldom be helped by search alone. Metadata and taxonomies are key to boosting findability. However, if "metadata" makes eyes roll, "taxonomy" can be a snoozer ... until you understand how important both are to boosting findability and making your content stand out and be found. This is true whether you're standing up a CMS or an intranet inside the firewall or if you're competing for eyeballs on the web. Developing a structure for navigating, finding, and understanding content-and the related descriptions in that taxonomy-is difficult and can be a surprisingly contentious process.

Everyone has his own idea of what constitutes good tagging or how folders should be structured. Is it better to have deep and narrow structures or wide and shallow-many tags or few? If you have not already had to deal with this, chances are you will sooner or later. I developed some tactics to build a consensus (especially between business and IT). Even though I succeeded, learning by doing is not my favorite approach. I often wondered how others managed to succeed, especially with larger efforts. Recently, I learned about a specification developed by the Learning Resource Metadata Initiative (LRMI) and its efforts to enhance findability and shape education technology content.

LRMI was established by the Association of Educational Publishers (AEP) and Creative Commons and recently came under the governance of the Dublin Core Metadata Initiative. Additionally, LRMI aims to assure that content could be found in web searches by teachers (using Google, Bing, and Yahoo) and then, in a future phase, to also be found by both primary and secondary school students to discover resources based on their goals. Does this plan sound overly ambitious? Wait, there's more. To help students and teachers, the system would need to leverage another initiative called the Learning Registry to understand more about student goals. The Learning Registry grew from the Shared Learning Infrastructure, driven in part by the Bill & Melinda Gates Foundation and Carnegie Corp., linking the registry (among other things) to digital education content using Common Core State Standards. This definitely dwarfs all my former taxonomy projects.

To find out more, I contacted two LRMI heavy hitters: Dave Gladney, LRMI project manager from AEP, and Phil Barker, who is affiliated with the U.K. initiative Centre for Educational Technology, Interoperability and Standards and with Heriot-Watt University. What was the incentive for educational publishers to participate? Findability by their customers for the educational resources, and for their part, educators demanded more productive and faster searches.

Key to LRMI's success was transparency and repeated polling of experts, any who wanted to contribute to the specification. Gladney said that LRMI "posted frequent updates online and in discussion groups. We also opened up these discussions via a free and open Google Group ... with official public comment three times during the development process." The result was that LRMI should be "lightweight, easy to implement, and easy to map to existing education metadata schema." This process continually reinforced for the LRMI leadership that enhancing search and discovery was key. And how about contention among participants? Barker cites the potential for getting stuck in the weeds and said that LRMI's strategy was that the specification must "be as widely applicable as possible ... not to include anything that ties it to one educational system." "K-12" is meaningful in some countries but not others. This meant allowing content creators to pick and choose, deciding whether to map between terms in different educational systems or use only terms relevant in one particular system.

Is this succeeding? Recent polls show more than half of publishers that are aware of LRMI already use it or plan to do so within a year. This year's educator survey showed that 84% would be happier with search engines using LRMI tagging. Transparency, simplicity, and mutual incentives are key. LRMI's lessons are useful for us all.