How much metadata is too much metadata?
By Adam Wilkins
Director of Strategic Consulting
Sitting in boardrooms across the country, this question is often asked of me. Now it’s important to remember the purpose of metadata: to provide context to a piece of content, like a document, so that it is more consistently findable when needed. It is great to see that people are starting to understand what metadata is (see post “what is metadata”), but I do find that many people are still wanting to tag their documents in all manner of ways, “just in case”.
If you allow for this via a poor design, you may actually make documents harder to find! For example, an engineering company I worked with wanted everything from the drawing number to the sheet size associated with every document, a total of 28 metadata properties to keep track of! The result was that users either refused to fill in all the information or did so inconsistently. If you are relying on metadata to filter search results this means documents are going to disappear.
Then of course we have the other end of the spectrum, those that don’t really care about metadata because they store every document in a folder, where nobody else can find them.
So like most things in life there needs to be a balance. Too little metadata or too much metadata and you can’t search for things in a meaningful way.
Before I answer the question of how much is too much, lets first eliminate what I call “System Metadata”. This is metadata like “date added”, “who added”, “last modified date”, “version number” and other similar tags that nobody needs to enter. If the value can be automatically populated by the system in any way, then let it be auto-populated. For system metadata – you can have as many as you wish – but over use needs to be considered in your taxonomy and enterprise search strategy. Always remember that each piece of metadata is intended to help search results. If what you are considering is not part of a search then don’t add it as metadata.
The other group to consider is “User Metadata”. These are tags that users will need to enter when they add a document and sometimes when a document is updated. This is where you need to be extra careful. Ideally you should have 4 values to enter, but no more than 8. And where possible make the entry with pick lists or similar pre-defined values, avoiding free form text entry.
The values entered should classify the document sufficiently so that the use of folders becomes unnecessary but that it will be “findable” thru a search. This starts to encroach on taxonomy…see blog post for details.