Librarians: Make Metadata Automatic
No metadata strategy survives first contact with the user.
That was the message from a diverse panel of data librarians at the recent Metadata Madness East conference in New York. Despite hailing from different industries, including retail, advertising and the non-profit sector, all agreed that users within an organization need to be coaxed and coached into creating reliable metadata.
One way to do that is to trick users into creating metadata, said Abigail Meisterman, metadata architect at Tiffany & Co. “When you don’t have [experienced] metadata creators – librarians and whatnot – you have to use your users, and that will be marketing and all sorts of different content creators, and metadata may be beyond them,” she said. “So you sort of have to trick them and say, you’re just doing this itty-bitty piece of work [filling out fields in an XMP file], and then we take those XMP fields and map them to our custom metadata schema and asking the [digital asset management] system to translate it. So when they think they’re just adding another SKU they’re really creating all this metadata.”
The more steps that can be automated, the better, the panelists stressed.
“We have a very low-tech way of capturing metadata from our content providers,” said Tony Gill, global director of library science and information management at AdPeople Worldwide. “We have users fill out a simple Excel form, with little drop downs with fields they can choose from. So they’re just filling out a form and they’re uploading metadata, and then the team of librarians sort of read between the lines and add the information to their ever-growing library of content knowledge.”
Jane Greenberg, director of the Metadata Research Center at Drexel University said organizations should take better advantage of the metadata that is automatically generated when many types of files are created.
“When you create a video a lot of metadata is automatically generated. Even something as simple as a Word doc generates metadata” she said. “We don’t always do a good job of leveraging the metadata that is associated with those assets.”
The bottom line: The less an organization’s metadata strategy relies on the diligence of its users the better for all concerned.