Metadata tagging is the process of creating a term that describes a keyword or phrase and assigning those tags to a media asset.
A media asset, as described in our blog about digital asset management, is any content that comes in one of five forms —audio, images, video, documents, or HTML. These tags don’t appear to the user but are in the source code and can be edited by the person managing these assets.
Metadata tags are an important aspect of digital asset management. This data essentially helps both internal teams find content by searching for key terms attached to these files.
It can also help external people, both partners, and potential licensing customers, search through an archive to find content that’s available to them, enabling controlled access.
In this blog, we’ll be diving into these key areas to provide an in-depth understanding of metadata tagging:
How is metadata tagging done?
Metadata tagging is typically done by teams who manage an archive of content, using some sort of asset management system or solution. Anyone that has a lot of assets stored needs metadata tagging to quickly find and surface content. This is especially true for video content.
Companies who have large video files, particularly if they are a studio or a production company, need the ability to know what is contained within lengthy run-time files. A two and a half-hour movie has many scenes that need to be documented with metadata so that they can extract moments for promotional purposes or determine where third-party approvals are needed to use the content.
There are many challenges that organizations face with their structural metadata processes. These challenges include:
- Keeping a standard for tagging content as multiple people touch these assets
- Inconsistencies in what’s tagged and what’s not
- It’s extremely laborious to tag content manually
- If content is not tagged accurately, that asset is at risk of being lost and never being found or used again
Let’s explore how these challenges, when solved, shine a light on the business value of metadata.
What is the business value of metadata?
First and foremost, the key business value of indexing media with metadata tags is classification. Classification with metadata enables teams to track key elements of an asset such as important dates related to a digital asset’s record schedule. This ensures that everyone can see what and when a file has been modified or updated. When done accurately this saves time and makes teams work more efficiently by not duplicating work.
From an information security standpoint, metadata can be a useful tool. This is often called administrative metadata, for it is used to help flag security settings, validate access to specific groups or individuals, and edit rights. This puts the owner of the content firmly in control of use and the distribution of their media assets.
Companies can use metadata to capture and improve the customer experience with their content. With descriptive tags, teams can use metadata to capture users’ ratings of content. For example, these tags can indicate that content is “valuable” or “outdated” and thereby not actively available for use.
And lastly, as you can imagine when dealing with hundreds, even thousands, of assets, information findability is paramount. Metadata streamlines search efforts by accelerating and enhancing the retrieval mechanism used within an archive. This enables users, both internal and external, more accurate search results that match a query on a certain field.
How do I create an indexing strategy?
Creating an indexing strategy, the traditional way, which is not the most efficient these days, would start by:
1. Defining who’s responsible for tagging content
2. Establishing a tagging standard for past and future content
3. Creating milestones for past content tagging
To set up tags, you need to consider the following so that teams or on the same page with implementation.
Keywords – These are descriptive terms that are assigned using a free-form text field or through an enforced list.
Taxonomy – This structured metadata approach consists of hierarchical keywords.
Field-level Text – These are short descriptors and categories relevant to a particular collection of assets.
Picklist – A picklist consists of descriptive terms a user can choose from. A picklist can consist of suggested terms or enforced terms.
Rights Fields – This approach is used to provide information related to usage rights or asset ownership.
As we suggested, the manual method, or “traditional” way of tagging media, is not the most efficient method anymore. This is because artificial intelligence has made this entire process faster and more accurate.
How to leverage AI in metadata tagging
With AI, and its various cognitive engines from facial recognition to object detection, companies can lean on technology rather than their teams to tag assets. By having AI do most of the heavy lifting, it can cut down the time it takes to tag all of the content in your archive from months to days.
Veritone Digital Media Hub, a media asset management platform, leverages Veritone aiWARE, the first OS for artificial intelligence. Using all the cognitive engines, both propriety and third party available with aiWARE, it can accurately tag content through face, audio, and text recognition (to name just a few).
Some of the largest media companies in the world use Digital media Hub to automate, curate, and activate how they manage, distribute, and monetize their content. A core aspect of Digital Media Hub is its AI tagging capability, and how users can edit this metadata and visualize it within the UI, adding structure to all the valuable unstructured data contained within media files.