By Abigail Wickes | Sat, Mar 10, 2018
The KBART format (Knowledge Bases and Related Tools) was developed by a NISO Working Group as a way to standardize the title level information content providers send to discovery services to facilitate customer access. This standard, developed by library, content provider, and discovery service industry stakeholders, creates efficiencies across the supply chain, and creating KBART files for title based projects has been an exceptionally helpful investment for Oxford University Press.
Before OUP provided KBART files, our discovery service partners had to ingest title level data for their knowledgebase software from either public facing title lists created for marketing purposes or from MARC records. Since the purpose of the knowledgebase is to facilitate customers’ access to holdings, marketing title lists and MARC records which were not designed for this purpose presented some problems:
- Marketing title list information was not placed consistently among different products. One product title list might have ISBNs in the first column and online ISBNs in the fifth column, and another product title list might have only print ISBNs and place them in the third column. Parsing this data required discovery partners to create elaborate scripts that were also very breakable if any column of data was relocated.
- Information from MARC records was not designed to facilitate access. Although OUP MARC records provide reliable title level bibliographic data, there is no clear indication as to what sales package a given title may be part of. This puts library customers in the tedious position of having to cherry pick what titles they may have access to from products that may contain thousands of titles.
By providing KBART files for our title based products we solve both of these problems. The format requires title level bibliographic data to be located in the same column across all serials and monograph products, so discovery services can use one script to find all required data in a standardized format. KBART files are based on standard sales packages, so libraries can select their holdings more efficiently; the library can select the sales packages they have access to, and they may also select individual titles outside of these packages as needed.
KBART also provides the flexibility to include additional metadata and access points in our files. Based on feedback from library and discovery service colleagues, OUP provides subject taxonomy and sales package data as well as MARC record numbers in all of our files along with the required data. We will also soon be incorporating OCLC numbers into our KBART files.
One of the main limitations of KBART is that files are based on titles in standard sales packages, and there are many instances where a library’s holdings do not perfectly match a standard sales package. I’m grateful to be at the forefront of that conversation through participation in the NISO KBART automation working group, and the group’s recommendations including an overview of the current landscape, use case examples where automation will solve current problems and inefficiencies, and a prototype for automated delivery will be available in early 2018.
Because this data transfer is so complex and involves so many different stakeholders—libraries answering patrons’ questions about accessing titles, discovery services needing additional information from content providers, and content providers answering both customer and vendor questions—clear lines of communication are another crucial element for successful downstream relationships. Before my role was created at OUP this work was carried out across different teams, which led to confusion for all parties involved on who to contact with questions. Publishers and discovery services working together to optimize discovery workflows benefits everyone, and investing resource to facilitate this communication has been incredibly worthwhile.
Providing KBART files has made many processes much simpler for our customers and downstream partners, but there are still many opportunities to enhance our metadata offerings. We’re so glad to join this conversation through Metadata2020, and we hope to discuss projects to improve searchability of products, expanding taxonomies, and delivery of metadata to downstream partners in this space in the future.
About the author
Abigail Wickes received her MLS from UNC Chapel Hill in 2012 where she focused on cataloging, metadata, and institutional knowledge management. Since 2015 she has been the Discoverability Associate at Oxford University Press, managing relationships with discovery partners and working to ensure OUP metadata meets industry standards. She would love to talk to you about publisher-provided metadata (especially KBART and MARC), how publishers can make life easier for librarians, and working in publishing with an MLS. She lives in Durham, NC with her husband, daughter, and miniature dachshund.