User-Centered Provisioning of Interlibrary Loan: a Framework

In Brief:

Interlibrary loan (ILL) has grown from a niche service limited to few privileged scholars to a ubiquitous expected service. Yet, workflows still assume specialness. Users’ needs should come first and that means redesigning ILL into a unified linear user-centered process. It is not just a request form, rather we need improved mechanisms for users to track, manage, and communicate about their requests. This article explores how ILL developed, problems with the current ILL ecosystem, and changes that can make ILL centered on users’ needs and processes rather than backend library systems.

By Kurt Munson


Interlibrary loan (ILL) provides library users with a critical tool to acquire resources they need for their information consumption and evaluation activities whether research, teaching, learning, or something else. The 129% increase in ILL volume between 1991 and 2015 in the Association of Research Libraries (ARL) statistics clearly shows that ILL has grown from a niche service to an expected one (ARL, 2016). Yet, our library processes for providing this service have not kept pace with technological development. Thus, the provision of ILL is less effective than it could be because it is predicated upon library processes and systems rather than on most effectively meeting users’ needs. This article explores the development of ILL as a service, suggests areas in need of improvement, provides a framework for redesigning this service in a user-centered way, and finally outlines efforts to create such a user-centered ILL to meet those needs.

Interlibrary loan holds a unique place within the suite of services libraries provide. ILL is entirely user initiated and driven by demonstrated user need. It provides a mechanism for users to acquire materials they have discovered and determined to be worthy of additional investigation but for which local copy is not available. ILL expands the resources available to users to that which can be delivered, not just the contents of the local collection.

The modern research library offers a range of services under the ‘Resource Sharing’ umbrella, including consortial sharing of returnables, interlibrary loan of returnables and non-returnables, and local document delivery operations. The ILL process discussed in this article is restricted to ILL as a brokered process whereby a library requests and arranges the loan of a physical item for use by an affiliated user. ILL practitioners refer to this process as traditional ILL of returnables, as the item will be returned to the owning library. Scans or reproductions of articles or portions of a work provided from a local collection or by another library fall outside this article’s scope because the workflows for sourcing and providing those items are quite different. This article primarily concentrates on ILL between academic libraries, though its recommendations are generalizable to public, medical, and other libraries.

Historical Development

ILL has a long history as a library service but for most of that history, it was a niche service provided to only a select group of library users, most often faculty members and perhaps graduate students. ILL was difficult, time consuming, and required a great deal of staff effort. Simply identifying an owning library was a challenge before the introduction of shared computerized catalogs. Citations needed careful verification to ensure accuracy, particularly for items created prior to the introduction of the International Standard Book Number (ISBN) system in 1968. Identifying holdings and ownership represented huge challenges. While tools like the Pre-1956 Union Catalog existed, these were out of date as soon as they were printed. Requests were made via mailed paper request forms. The library that owned the item would likely know nothing of the requesting library so the trusted relationships we take for granted had not yet developed. A library might send an item or it might not. An owning library might respond in the negative or it might not. It was at best an arduous process analogous to weaving cloth and sewing garments by hand rather than purchasing ready-made off the rack clothing.

The creation of the OCLC cooperative in 1967, specifically its shared index of items, provided the opportunity to vastly improve ILL processes and workflows. The OCLC database, eventually to be known as WorldCat, contained one record for a work and libraries that could indicate who owned a copy of that item. It was now possible to identify ownership easily. Moreover, this identification could be done in one place with simultaneous citation verification. OCLC introduced the first of its interlibrary loan subsystems in 1979 (Goldner, Birch, 2012, p. 5) because there were now enough item records and holding records in the shared OCLC index to support ILL processing. Over time, additional axillary ILL services for library staff were introduced by OCLC. For example, a library can provide contact information, address information, and explain what it will and will not lend with any associated costs for these services in the ILL policies directory. The OCLC ILL Fee Management (IFM) system provides an automated billing system as part of the transaction process. ILL became markedly easier to do, or at least portions of the process did.

The development of WorldCat and other union catalogs made the process of identifying owning libraries and placing requests much easier but these were closed systems with limited functionality. These systems did one thing: placed a request. Yet, ILL is a multi-part process consisting of many disparate steps that library staff perform. Files of request forms require maintenance. Users need to be contacted when items arrive or need to be returned. Circulating necessitates tracking over time. Physical items require packing and shipping. Invoices require payment.

For the library user, ILL is just one of many tools to acquire materials and the user’s interest is accessing the materials, not how the library chooses to source the requested item. Users once filled out a paper form which staff keyed into the requesting system. Then the user patiently waited until they received a phone call or postcard alerting them that the item had arrived. To be sure, verification and ordering had become easier but the process still involved many handoffs between different systems with minimal communication.

Easier ordering allowed ILL request volumes to increase markedly (Goldner, Birch, 2012, p. 5). ILL management systems were developed to automate the management and tracking of requests over their lifespan in addition to handling communication with users and to circulate the items. ILLiad is the most common ILL management system used today in academic libraries. Both owning libraries and requesting libraries came to rely upon these systems to manage requests over their lifespan. Request databases replaced file folders. Data could be pushed from one system into another. Routine tasks, such as sending overdue notices, could be automated. ILL had become a standard mainstream expected service rather than the niche one.

Improved staff processing was not the only driver for increased volume. OpenURL and other outgrowths of user-facing databases and the ubiquity of the internet made discovery easier (Musser, 2016, p. 646). The easy transfer of metadata via OpenURL increased request volume because users could request items by pressing a button instead of filling out a paper form. The request went into the request database for staff processing. Nonetheless, the improvements ILL management systems provided remained rooted in ILL’s traditional union catalog-based requesting workflows. They focused on making library staff processes to provide items more easily rather than user workflows or needs. Issues with the approach and workflows described above are explored below.

Problems with Our Current Approach

A number of issues limit ILL service’s usability which in turn limits its effectiveness for both users and library staff. To be sure, ILL services are valued by users and play an integral part in the suite of services libraries provide to source materials for users, but it can be improved upon by reconceptualizing the process whereby it is provided. Libraries can rethink how the individual parts of the process, be they software or workflow, are put together. Areas for reconceptualization fall into five broad categories, and these are discussed below.

First, existing systems are based on identifying libraries that own a requested item. But for the purposes of ILL, ownership is only the first step in the process. An on-shelf loanable copy must be located because only items that fit these criteria can fill the user’s need. WorldCat can tell us who owns an item but what we need is a library that can loan the item. Owning libraries, or lenders as ILL practitioners call them, still need to perform a search of their local catalog to determine if the item is on shelf and loanable. This involves a time-consuming antiquated manual workflow that fails to take advantage of tools such as Z39.50 for automated catalog lookup. Workflows have not kept up with technological advancements.

Consortial borrowing systems, such as Relais D2D or VDX, where a group of libraries share a discovery layer that displays availability, mitigate the issue described above but these systems also have a serious shortcoming: they force users to execute the same search in multiple discovery layers to find an available copy. Users, having identified an item, cannot simply submit a request and have the library source it for them. Rather, libraries expect users to navigate across disparate interfaces with unique request processes to request an item. Thus discovery and delivery become a fractured process for users as libraries push the work of finding a loanable copy onto their users.

Second, identifying owning libraries remains tied to the searching of union catalogs because metadata is not recycled efficiently. A user searches their local library’s discovery tool and finds that an item they want is checked out so they fill out an ILL request form populated with metadata from their local discovery tool. Library staff, or preferably automated systems, then re-execute a similar search using that same metadata against a larger database to identify potential lending libraries and the request is ported into a different system. Since the metadata populating the local discovery tool likely came from WorldCat in the first place and that metadata will be used to search against WorldCat again, said metadata should be trusted rather than assuming that the citation needs verification by library staff. This is again an antiquated workflow rooted in past practices.

Third, ILL is very much predicated on the terms imposed by the owning library. While the OCLC policies directory provides library staff with information about terms of use for borrowed items, the lack of consistent agreed-upon standards for loan periods between libraries creates a situation ripe for confusion on the part of users. Again, this harks back to an era where ILL was rare, difficult, and unique rather than the current situation where ILL is a standard service. Too much emphasis is placed on unique locally defined rules rather than on setting broadly agreed-upon standards or considering users’ needs for materials.

Fourth, the process uses siloed systems with weak integrations and poor interoperability. Discovery happens in one system. Requests are managed in a separate ILL management system which ties to an external ordering system for sourcing items. When the item arrives at the borrowing library, these respective systems must be updated but then the item needs to be handled as a circulation likely in a separate system again or in a system separate from the one that manages the user’s loans for locally owned materials. Yes, the systems can communicate between each other but this process is staff intensive and lacking in automation. Crosswalks, bridges, and information exchange protocols are not employed fully or efficiently.

Finally and most importantly, providing ILL services is predicated on library processes or library tools rather than user processes or needs. Users must learn and jump between disparate systems, often with jarring handoffs, to acquire materials. Depending on how the item is sourced by the library for the user, they need to find the system where the library has chosen to process that request. Communication is scant. It comes from different systems and mostly consists of silence until a pick up notification is sent. This confusing process is followed by inconsistent rules surrounding use based on the lending library’s terms of use. Usability studies have demonstrated how this confuses users (Foran, 2015, p. 6). Presented with multiple, often contradictory delivery options, and unclear explanations of the differences between them, users tend to place requests in each system in the hopes that one will work. Not only is this poor customer service, but it also increases staff workloads and costs for the library with duplicated work. Why? Because libraries define ILL success as having acquired a copy for the user. The user’s needs—required turnaround time, format, amount of time they will need the item or even its relative importance to them for intended use of it—are secondary, when even considered. Libraries need to gain a better understanding of how ILL fits into the user’s activities and how they can more effectively support those activities. ILL needs to be borrower-centered not lender-centered.

In many ways the issues outlined above are a natural outcome of a service’s evolution over time and the result of a fairly stable ecosystem that expanded gradually over time. The foundational systems which undergird the service were able to absorb the increased request volume and processes simply continued without redesign or rethinking. Yet the environment in which the service exists is evolving rapidly and the time for a radical rethinking of the technology used to support the service workflows and metrics for success is here.

Recommendations for Developing an Alternative Framework

At the International ILLiad Conference in March 2016, Katie Birch of OCLC announced that OCLC intended to “move ILLiad to the cloud”. Far more than any other change in ILL processing or systems, including the introduction of Worldshare ILL, this announcement shook the foundations of academic library ILL in the United States. We were presented the opportunity to reimagine how we provide ILL services. We began to ask the question “what should the ILL workflows be?” How could we make them more user-centered rather than continuing the historic workflows mandated by vendor-supplied platforms? Concurrently and partially in response to this announcement the Big Ten Academic Alliance (BTAA), previously known as the Center for Institutional Cooperation (CIC), embarked on a project to explore, redefine, document, and share a user-centered discovery to delivery process. The project’s goal was to describe an easy-to-understand user experience that shielded them from the disparate library staff systems and provided a more linear discovery to resource delivery process. Usability studies confirmed library staff members’ impression that the process was confusing and disjointed to users (Big, 2016, pp. 19-22; Big, 2017b, pp. 19-21). Cooperatively with Ivies Plus Libraries and the Greater Western Library Association (GWLA), we defined base requirements and system functionalities for a new user centered vision of ILL. A one page summary document entitled “Next Generation Discovery to Delivery: A Vision” was released in February of 2017. Staff from BTAA libraries, including the author of this article, wrote two reports entitled “A Vision for Next Generation Resource Delivery” and “Next Generation Resource Delivery: Management System and UX Functional Requirements”. These works, in part, inform the three broad recommendations outlined below, described as: user process, technological, and cultural.

To start, the library tools that support the users’ processes must be based upon their workflows rather than the processes library systems staff use to manage that work. Where in the past a user interface was tacked onto a library staff system, this should no longer be the case. Users deserve a simple universal request mechanism, a “get it” button (Foran, 2015, p. 5) that connects to a smart fulfillment system (Big, 2017b, p. 9). Requests should display in a single dashboard-like interface that allows users to manage all their library interactions in one place (Big, 2017b, p. 9). No longer should users be expected to hunt across disparate library system interfaces to locate their request for that specific item. Achieving this requires that we rethink how we, library staff, present library systems to users. Since the primary local discovery layer is the user’s primary entry point into the library and the place where they manage their library interactions, this interface needs to be the place where we display all request information to them. Thus, vendors who provide discovery layer tools must make them open and capable of incorporating data from external sources so we can provide users a unified display. They should be shielded from systems libraries use to perform their work of fulfilling requests. Users need items and which library staff process is invoked is immaterial to them. Getting the item is paramount. This notion must inform how libraries design, combine, and present their backroom systems to our customers.

Second, delivery of an available on-shelf loanable copy to the user who needs it and made the effort to ask for it is what matters, not identifying owning libraries. ILL loans are simply more complicated circulations. Discovery tools should be separated from discovery options as these two do not need to be interconnected. The metadata from discovery is all that is needed to initiate delivery. Request should be managed via a lightweight system specifically designed around the efficient and timely fulfillment of that user’s request with user satisfaction serving as the primary metric for defining success. The BTAA reports named this new idea “Resource Delivery Management System” (RDMS) (Big, 2017b, p. 12). Working off a list of potential partner libraries maintained and defined in the RDMS, a simple Z39.50 search using that recycled metadata should identify a potential lending partner and when a loanable copy is found, a request should be placed via NCIP with routing and courier tracking/shipping information included in the RDMS’s request record. Circulations of ILL items should occur in the local Library Services Platform (LSP) so users can managed all loans regardless of how they are sourced in one place.

The ideas above, in many ways, represent a somewhat radical break from past processes or practices. They decouple sourcing of materials from a shared index. Instead, they are based on library-defined partnerships and the identification of a loanable copy at a partner. Moreover, this approach promotes interoperability across different systems as the request is not tied to any legacy or monolithic system. Multiple micro-systems each play a part to complete a multistep process. Finally, it limits the functional scope of the RDMS to just the management of delivery, avoiding the current problem of (often subpar) duplication of functionality across systems. While no such system as described above exists, potential development is under exploration by vendors.

The ideas outlined above further move us from the current siloed systems to one where integrations are central and key and where the best, most appropriate system, manages or provides the required information (Big, 2017a, p. 1). Thus, the local LSP handles all aspects of notification, circulation, and fines or blocks. Viewing this as a process consisting of many parts also allows us to reimagine it so that we can incorporate other previously excluded information such as shipping status derived from the UPS or FedEx APIs. Additional communications to users about the status of their request should be included too. Companies provide these updates on orders and shipping as a matter of course so libraries can also. Users reasonably expect them. Authoritative sources, rather than poorly duplicated ones, should be called upon to provide information as needed. Local address information sourced from that campus identity management system, for example. This system consists of many parts communicating with each other via protocols using APIs when needed. Binding their collective parts together with each assigned a specific task provides a new framework for the workaday provisioning of ILL services.

Technology is easy to change. Culture is more difficult, particularly entrenched library policies. These policies’ efficacy at guiding user behavior and promoting shared stewardship of materials is almost never tested. Yet, users and library staff are both equally engaged in the management of loaned items. Libraries need to embrace the early slogan of the Rethinking Resources Sharing Initiative, “throw down your policies and embrace your collections” and libraries need to manage this sharing efficiently in a data-driven way.

It is important to remember that users need materials to complete their work. The use of materials by users is predicated upon their need, associated timeline, and perceived value of the item. As the Big Ten Academic Alliance has stressed, “All that matters is format, time to delivery, loan period, and costs to the patron, if any” (Big, 2016, p. 9). These items have value to the user. They put effort into acquiring them. ILL is entirely user-driven unlike many other library processes. Arbitrary loan periods as set by any owning lending library may and in fact do come into conflict with users’ needs (Foran, 2015, p. 4). Libraries can resolve these conflicts easily by moving to standardized loan periods for ILL. Standards should replace the boutique exceptionality encouraged by the OCLC policies directory.

Stated differently, the emphasis needs to shift from lender-imposed restrictions to borrowing libraries having the ability to communicate standard policies. For example, the BTAA shared twelve week loan period, when complemented by the equivalent Northwestern University local loan period, coupled with user blocks and assessment of replacement cost fines after thirty days provide a consistent user experience that, in turn, encourages the timely return of items. For example, only 29 of 29,137 total ILL loans were lost by Northwestern University users in 2016. This example demonstrates how consistent policies promote compliance. Why? Because they are both easy to understand and failure to comply with communicated expectations has direct consequences, specifically the loss of library privileges. Further, research done by the Ivies Plus Libraries demonstrates that almost all items are returned to the owning library after the user has completed their use of said item. Only 70 items of roughly 750,000 over three years were truly lost by patrons or never returned. This data clearly demonstrates the need to rethink policies across libraries and reconsider shared assumptions. In other words, the emphasis needs to be on understanding user behavior based on their needs and developing effective ways to affect their behavior to achieve agreed upon reasonable outcomes.

Libraries must also shift from their historic lender-centric ILL system to one where an ILL user receives an item and national standards provide them a consistent easy-to-understand experience. This would promote an environment where borrowing libraries can more effectively manage their users. Appropriate effective tools, tested by data, are needed. Ineffective tools need to be discarded, like overdue notices via email from the lending library to the ILL borrowing staff. These will never affect user behavior. Making the process easier for users to understand in terms of policy is critical. The introduction of standardized loan periods, replacement costs, and the like across libraries would simplifying the management of ILL for both users and library staff. It would also greatly assist in achieving compliance and reducing (often pointless) staff work.

Rather than starting with the question of which library system can perform a specific job, we need to rethink this process and backfill the appropriate system, library or other, from the starting point: the initial discovery and request by the user. The BTAA phrased this as smart fulfillment. Smart fulfillment is a linear path for users to follow where effective automated handoffs between library systems source and manage requests from or in the most appropriate place.


ILL has grown from a niche service to an expected standard one, growing 129% between 1991 and 2015 in ARL libraries (ARL, 2016). Yet workflows and system integrations have not evolved as much as they should have in response to this growth. A confluence of announcements and work to redefine processes now presents libraries with a unique opportunity to rethink ILL, transition from legacy practices, and to unify the fractured discovery to delivery process we present to our users. If we integrate library systems and systems that support library systems differently, and effectively leverage each system’s strength, we can create an easy-to-use service that meets demonstrated user needs. We can provide a service that provides smart fulfilment of requests and improves both the user and staff experience. This should be our goal.

The author wishes to extend his deepest thanks to Heidi Nance, Director of Resource Sharing Initiatives for the Ivy Plus Libraries, for her willingness to review this article, apply her deep knowledge of ILL while doing so, and for the thoughtful comments and suggestions. Thank you, Heidi.


Association of Research Libraries. (2016). ARL Statistics 2014-15. Association of Research Libraries. Retrieved from http://www.arl.org/storage/documents/service-trends.pdf

Big Ten Academic Alliance. (2016). A Vision for Next Generation Resource Delivery. Retrieved from https://www.btaa.org/docs/default-source/library/d2dnov2016report.pdf?sfvrsn=4

Big Ten Academic Alliance. (2017a). Next Generation Discovery to Delivery System: a Vision. Retrieved from https://www.btaa.org/docs/default-source/library/discoverysystemsvisiononepage.pdf?sfvrsn=2

Big Ten Academic Alliance. (2017b). Next Generation Resource Delivery: Management System and UX Functional Requirements. Retrieved from http://www.btaa.org/docs/default-source/library/next-generation-resource-delivery–functional-requirements.pdf

Foran, K. (2015). “New Zealand Library Patron Expectations of an Interloan Service.” New Zealand Library & Information Management Journal. 55(3), 3-9. https://lianza.org.nz/nzlimj-volume-55-no-3-october-2015

Goldner, M., & Birch, K. (2012). “Resource Sharing in a Cloud Computing Age.” Interlending & Document Supply, 40(1), 4-11. https://doi.org/10.1108/02641611211214224

Musser, L., & Coopey, B. (2016). “Impact of a Discovery System on Interlibrary Loan.” College & Research Libraries. 77(5), 643-653. https://doi.org/10.5860/crl.77.5.643

Stapel, J. (2016). “Interlibrary Loan and Document Supply in the Netherlands.” Interlending & Document Supply. 44(3), 104-107., https://doi.org/10.1108/ILDS-03-2016-0015

4 Responses

  1. Luke M

    Hi Kurt,
    An interesting insight into how ILL operates in North American libraries! The thing that stood out the most as a difference between US and Australian ILL systems (if I have read your paper correctly), is the onus placed on users to find out which libraries hold the item they want and send separate requests to each institution or public library service. I think the Libraries Australia Document Delivery System (LADD), managed by the National Library of Australia could prove a good model for a more user friendly experience. Australian and New Zealander users can search for an item across the National Bibliographic Database. Once they identify that an item exists in Australia or New Zealand, they can simply ‘get it’. You can view the database for yourself here – https://librariesaustralia.nla.gov.au/search/simpleSearch?action=Login&mode=login&main=true&querystring=null . Depending on the rarity of an item or where it is held, there may be a fee charged for the item, but many public libraries will supply an item for free. The request is mediated by a librarian at the library they belong to, who double checks for availability of an item then forwards the request on to every holding library on the users behalf. This is managed by a rota, so the ILL officers at holding libraries will see the requests one at a time until someone can fulfill it. In this way a request can often be fulfilled withing about 5 – 10 days. This is how it works in a public library in Queensland anyway, the process could be a bit different for academic libraries. This paper from the National Library of Australia goes into some more detail about how LADD works – https://www.nla.gov.au/content/libraries-australia-document-delivery-a-system-for-a-variety-of-users
    Hope this is of interest!

    1. Luke M

      I have incorrectly addressed my comment to Heidi rather than the author. Sorry Kurt! Could the moderators please amend this?

  2. Kurt Munson

    The national bibliographic database is like to OCLC’s WorldCat for the US. Remember that there is no equivalent national database for the United States, a key difference. Our’s is a much more distributed system where we start with a local system, often for one library only, and then move out from that in concentric circles of fulfillment systems ending with OCLC’s WorldShareILL.