I've had a number of conversations recently about the importance of metadata to a DAM system and frankly , I think it' all about the metadata. It's not the sexy part of the process and may not get the attention it deserves...but it's critical and will either sink or swim a project.
I can understand that the nuts and bolts of a data model may not hold the same level of excitement to senior management as the "WOW FACTOR: of a new technology, or the content or GUI's but the value of the DAM will rest on the success of the metadata model, the type of metadata captured, by who and when, etc.
So why is it so important?
Let's think about a basic media workflow. Content is captured or created, digitized (or born digital), added to a DAM, used in a production cycle and then archived. Without a structured data model there is no way to efficiently use this content beyond these immediate needs. Information is entered on an as-need basis in whatever way makes sense for that individual at that time with no stake in the future of the asset. The accuracy, consistency and completeness of the record is unknown. After the project is completed there is no way to assure that the records archived are search-able for re-use within the enterprise AND if search-able there is no guarantee to the value or completeness of the search returns. I won't get it to the rights implications here but obviously that would be an issue as well.
A good data model needs to be an important and early part of a DAM design process. It needs to be streamlined, structured and governed so that there is controlled consistency to the data and that the attributes are keyed to the needs of the users. Ultimately, the data model should include information across different functions and parts of the asset life-cycle including:
- Capture and Creation specific data
- Asset inventory data (ie. info re: physical and digital asset, formats, locations, filenames, etc)
- Production cycle data
- Cataloging metadata to enhance search usage and assure rich, complete and consistent data about the about-ness for each asset.
- Rights data including ownership, usage, restrictions, etc.
- Distribution data
The depth of information for any of these can vary as needed but the idea is that with this structure the archived assets within any DAM can be examined from a multitude of perspectives and purposed and bring value to the entire organization..beyond that first use. Additionally, in a perfect world the DAM users should be able to pick and choose the data attributes from the data set available to them (via admin rules) that they choose to display and work with for their specific needs.
Building the structure, rules, style guidelines and attributes take some time and effort but ultimately pay off. Look at how MLB.com is able to tap into the data rich aspects of baseball to enhance their products...it's about being smart with what you have and how you organize it.
Let's play ball and not discount the importance of the DAM data!
Well said, Phil. In point of fact, the success of every bit of information management, structured, semi-structured, and unstructured, depends on well-defined, persisted, and stewarded metadata. But one area where I will respectfully disagree is this...I think metadata is DAMN sexy! :-)
Posted by: Bob Averack | September 07, 2009 at 02:44 PM
Great point Phil! Especially "the attributes are keyed to the needs of the users".
Posted by: Scott Santulli | September 05, 2009 at 09:20 AM