Intranet Insights

Asking authors to tag their work with keywords is frequently the way content is assigned metadata and classified in intranet content management systems. While this can be effective — depending upon the nature of your content and how like-minded your authoring team members are — it often is not.

Rather than focusing on such ‘explicit’ metadata, I recommend using the technology to your advantage by gleaning ‘implicit’ information about content wherever possible. By way of example on the public web, think of the recommendation engine at Amazon.com: The books that Amazon suggests are determined primarily by information implicit in your earlier actions.

Here are a few examples of implicit metadata:

• Author
• Author’s department
• Date created
• Date modified
• Links to and links from (these are the foundation of Google’s multi-billion dollar enterprise)
• Title (whether this is considered metadata or is part of the content depends on several factors: Author training, strength of writing guidelines etc.)
• Where created (by this I mean “where was the author’s starting point in creating this content?” It requires an authoring environment where authors navigate the site prior to launching a form to add a piece of content)
• Automated keyword extraction

Any of these can be leveraged to drive both reporting and end user UI.

Another type of metadata that can help free you from a dependency on authors tagging content is ‘user-centred’ metadata. Here are two examples:

• Clickstream data (i.e. where are viewers of a particular piece of content coming from? Where are they going after reading it?)
• Popularity

Finally, here are two examples of ‘user-generated’ metadata:

• Star rating (give users the option of rating each piece of content)
• Thumbs up/down (a simpler way of promoting/demoting content)