Metadata, like the electrical grid and the highway system, fades into the background of everyday life, taken for granted as just part of what makes modern life run smoothly. When metadata is doing its job well it fades into the background, almost to the point of being invisible.
– Jeffrey Pomeranz, ‘Metadata’
We’ve all embraced Amazon, Spotify and eBay and use metadata in these systems to search, sort, filter and group information. We don’t need to think about how these systems works – they just do. If we have any complaints, its usually because there is not enough metadata.
Nevertheless, there’s a consistency to these systems. Take Spotify, Google Play and iTunes. They all operate according to the same set of metadata standards (ID3).
Many communities of practice have evolved metadata standards over time. For example, in the library world, subject-based metadata has been systematically and cooperatively created over decades in schemes (thesauri and ontologies) which are used worldwide.
Metadata schemes, initially developed for use by humans, have informed the designs of the earliest library software. And it is the extension of these metadata schemes and the development of data exchange standards which underpins most of the successful search platforms we use today.
When standards become ubiquitous, metadata really does become invisible. We find what we want quickly and easily and trust the integrity of information we find. Unfortunately, because it’s invisible, we don’t always appreciate the value that metadata is delivering to us.
Businesses who develop fit-for-purpose standardised metadata to describe their documents can find the “right information in the right place at the right time”. The value to the business is enormous. The inverse cost of not finding the right information is proportionate.
The cost of poor metadata is increasingly visible and many businesses are getting serious about data/metadata management. But metadata management for unstructured information can be a minefield if not done well, because there are multiple stakeholders who must be taken into consideration.
A few pointers for making your metadata work:
- Metadata is the foundation of all information systems and most businesses have dozens of information systems. Check your information asset register to find out what metadata other systems are using.
- Enterprise metadata management reduces costs for everyone. Metadata and taxonomies that are developed and stored in SharePoint, eCM systems and search engines are difficult to share and reuse.
- Reduce the costs of developing metadata by using established enterprise or industry sources. For each class of metadata, consider which are the authoritative data sources for your business or industry. For example: Assets and Suppliers in your ERP system, Customers from your CRM system.
- Use metadata repository software that enables you to capture alternate labels (non-preferred terms, synonyms, acronyms etc.). There are many search engines that will use this information.
- If you’re working with unstructured content, you will be working with records. Understand that there are standards for metadata describing the content, context and structure of records, and rules for retention disposal and data protection.
- You may have a local project to complete but there will be other stakeholders who are impacted by your project. Find out who they are and what their requirements are. Try not to be Robinson Crusoe.
- At some point your unstructured information will be moved, transferred, migrated into another database. It’s a given. If you build your database to enterprise standards, the cost of future migrations will be significantly reduced.