Making AM Collections discoverable in the library environment

Every time you search a database or library catalogue, metadata is working hard behind the scenes. At AM, we know that well-structured metadata – in formats that work with existing library systems – is essential to helping users discover our collections. Jennifer Wedge, AM's Metadata and Discovery manager explores what we do to support that.

Whether you’re searching a library catalogue or your favourite streaming service, you’re relying on metadata, the information used to describe a resource. Metadata can be as simple as a file creation date, or can take the shape of detailed and structured records governed by international standards. Metadata is vital for discoverability and ensures that wherever they start their search, researchers can either find exactly what they’re looking for, or serendipitously discover something relevant. 

Metadata is a big part of the publication process at AM; it is drawn from source archives, with additional data supplemented by the AM Editorial team in order to facilitate browsing and searching across each collection. 

How AM supports discovery in the library environment

Making AM content discoverable beyond AM’s own platform means ensuring metadata integrates smoothly into existing library systems. This is where I come in. As Metadata and Discovery Manager, I make sure that AM’s metadata is available in formats libraries can use, either directly or via discovery services.

A large part of that is the provision of records that comply with MARC 21, the standard format for bibliographic data used in academic libraries. Monographs and serials digitised within AM’s collections are catalogued using familiar MARC practices, but our collections go far beyond books and journals. AM’s primary source collections can include anything from promptbooks to personal correspondence, artwork to alarm clocks. 

Translating all of that into MARC requires a blend of automation and human cataloguer oversight, taking metadata created by our Editorial team and converting it into structured MARC records using a tailored process for each collection. 

Balancing efficiency and accuracy in MARC creation 

AM publishes exciting new collections regularly and we strive to release accompanying MARC records alongside them so that researchers can start discovering content right away. With collections containing hundreds or even thousands of items, creating these records is no small task.

Attention to detail is crucial: the placement of specific punctuation in a title statement may be seemingly minor, but for a metadata librarian that misplaced colon can signal wider issues within a record. That’s why we’ve worked hard over the past couple of years to improve the quality and consistency of our automated MARC workflows.

Previously, records were generated through a rigid mapping process that kept all formatting as it appeared on the platform. Today, workflows are customised for each collection, with improved formatting and added enhancements like authority-controlled names and subject headings to optimize discoverability. 

In our Nineteenth Century Stage collection, for example, authority control is used to disambiguate theatres with similar names and link records related to specific locations, helping trace productions and organisations more effectively.

Cataloguing primary sources: Key challenges 

Creating MARC for our collections is fascinating, but it’s not without its challenges:

• Scale: AM’s largest collections contain tens of thousands of digitised items, making some degree of automation essential.
   
• Diverse material types: The publication 1980s Culture and Society contains more than a dozen distinct document types, from zines, posters and objects to correspondence and films. Where possible we are formatting the MARC to specific material types, but this approach significantly increases the workload. Many of the documents we publish also do not easily fit into MARC material types.
   
• Subject metadata: Assigning accurate, consistent subject headings across such a wide variety of historical material is a significant task. Our Editorial team creates bespoke controlled vocabularies for each publication, utilising feedback from contributing archives, expert editorial boards and the communities involved. Mapping this to library thesauri can sometimes feel like a step backwards from the nuance of in-house vocabularies, but the inclusion of Library of Congress Subject Headings (LCSH) is also the most frequent request we hear from librarians. Much work has gone into constructing appropriate LCSH for many of our new collections. Where this is not possible due to work involved, we are trialling the use of FAST (Faceted Application of Subject Terminology), which maps more easily onto our Editorial metadata.
   
• Sensitive content: The collection Amnesty International Archives contains material relating to human rights violations from the second half of the twentieth century. Careful thought was required in how to represent this in the MARC records, from appropriate subject metadata to the inclusion of content warnings.
   
• Institutional needs: We aim to produce metadata that’s flexible enough to suit a wide range of library environments and preferences. For some collections this means doubling up our approaches. When reassessing the MARC records for Mass Observation Online, for example, some customers wanted the granularity of  individual records for each document whilst others preferred a more top-level approach that directed users to grouped collections of document types. Both item-level and collection-level MARC records were created and are available for download.

Despite these challenges, each collection offers the cataloguer its own quirks and joys to work on. From transliterating Russian film titles for Socialism on Film, to navigating established meeting names in The Olympic Movement, every collection brings a different form of puzzle-solving to the intellectual work of cataloguing.

Keeping the conversation going 

No single metadata solution will work for every institution, but our goal is to make AM’s collections as accessible and discoverable as possible while saving time for librarians. 

We’d love to hear more from the library community about how we can continue to optimize AM collections for the discovery environment. By working together, we can make it easier for researchers to uncover new narratives in primary sources.

You can ontact us via. our website with your reccomendations .
 

About the author 

Jennifer Wedge is the Metadata and Discovery Manager at AM.