When we talk about digital assets, we have to talk about metadata. Unlike text files or XML documents, digital assets like images, video and audio files do not have any text or structure that can be indexed and queried. To understand what you have and to find what you need, you must collect extra information that describes your assets. In other words, you need metadata.
Every business process that touches your digital assets probably requires specific metadata, which practically guarantees that your metadata is more varied and dynamic than the digital assets themselves. In a talk I recently presented at the Henry Stewart Digital Asset Management conference in Los Angeles, I listed the following common pain points that enterprises face when attempting to manage all this metadata:
- Ownership: Most organizations don’t worry about its metadata or who is responsible for it. Without a clear owner, metadata is created and updated in an adhoc manner to satisfy each individual business process. A publishing company, for example, may have cover images that it uses on its marketing website, inside eBooks, and that it distributes to online bookstores. Each of these business processes requires metadata, such as the filename, size, resolution and description of the product. Since there is no clear owner for this data, the art department will record the information in a Word document, the marketing department will keep the same information in a spreadsheet, and the production department keeps the data in a database. Not only does this result in duplicated effort, but the information itself will have no consistency.
- Definition and Modeling: Allowing individual users and departments to decide what metadata must be kept results in duplicated metadata fields or fields with unclear purpose and meaning. Images may end up with metadata fields named “title”, “description” or “caption”, depending on the context in which the image is used. Taking the time to determine the information that needs to be kept, creating consistent names, and defining the allowed values is critical to effectively find and reuse your assets.
- Storage and Management: All DAM systems provide a method for assigning metadata to assets. But is the DAM the best place to create and maintain your metadata? It makes sense to store file attribute information like filename, type and size in the DAM, but should descriptions be stored and maintained in the DAM or in an external system like a product or title management database? What about rights and permissions information or taxonomic and semantic metadata? There are many specialized systems that give users domain specific tools that can do a much better job creating and maintaining specialized information.
- System Integration: If you decide to store your metadata in multiple systems, then you must find a way to integrate these systems so that information can be shared between processes. Even if you store all your metadata in the DAM, you will find that you need to feed your metadata to other applications, such as search engines, fulfillment systems, production workflows, and distribution channels.
- Staffing and Training: Once you’ve created a good metadata model and developed the systems to create and maintain metadata, you need people to enter and maintain the required information. Here you confront two problems: first, does your staff have the training to know how to enter the right information? Second, does your staff have the time to collect and enter the metadata? I worked with a media company that had designed a very rich metadata model and constructed a custom DAM to hold its images. The system worked well, but after three years, only 80,000 out of a total universe of nearly 1 million images had been entered into the DAM. The problem? The metadata model was so extensive that no staff had the time to enter all the information required. You need to take a realistic look at how much your staff can support and find ways to automate the process by leveraging technology and existing sources of information.
- Determining Value: Creating metadata models, building and integrating systems, and training and hiring staff takes time and money. If you can’t justify the value of your metadata, the resources will never be allocated and you will never be able to collect the metadata that you need to manage your assets. The first step of any metadata management project is to quantify the value of the metadata to the business. How many more orders could be generated if you can find the rights assets at the right time? How much will you decrease operating costs if you can find and reuse assets rather than recreating them? How many licensing penalties can you avoid if accurate permissions metadata is attached to the right assets?
There is no one-size-fits-all method that is guaranteed to overcome all these problems, but securing high-level business sponsorship is certainly a prerequisite for securing the right level of planning, staffing, and system investment.
It is also important that people have realistic expectations about what Digital Asset Management Systems can accomplish. No DAM is going to track all the metadata that you need without configuration and integration with other external systems. If you don’t have technical staff that can do this task, you must find an integrator who can work closely with your enterprise to do this work for you.
Above all, remember that metadata is not static. Once you determine a metadata model and have configured your system to manage that information, you will soon need to make changes. Be prepared to constantly review, revise, and improve the metadata you collect and how you store and distribute it. Don’t look at metadata collection and management as a one-time project. It is a vital part of your business and needs to be treated as an ongoing function.