Roaming Reference: Reinvigorating Reference through Point of Need Service

Kealin M. McCabe
Research and Learning Services Librarian
University of Northern British Columbia

James R.W. MacDonald
Digital Initiatives Librarian
University of Northern British Columbia


Roaming reference service was pursued as a way to address declining reference statistics. Librarians armed with iPads staffed the service over a period of six months during the 2010/2011 academic year. Transactional statistics were collected in relation to query type (research, facilitative or technology), location and approach (librarian to patron, patron to librarian or via chat widget). Overall, roaming reference resulted in an additional 228 reference questions, 67% (n=153) of which were research related. Two iterations of the service were implemented: roaming reference as a standalone service (Fall 2010) and roaming reference integrated with traditional reference desk duties (Winter 2011). The results demonstrate that although the Weller Library’s reference transactions are declining annually, they are not disappearing. For a roaming reference service to succeed, it must be a standalone service provided in addition to traditional reference services. The integration of the two reference models (roaming reference and reference desk) resulted in a 56% decline in the total number of roaming reference questions from the previous term. The simple act of roaming has the potential to reinvigorate reference services as a whole, by forcing librarians out of their comfort zones, and allowing them to reach patrons at their point of need.


reference services; roaming reference; academic libraries; point of need; iPads


The University of Northern British Columbia is a small, research intensive university located in Prince George, British Columbia. The University’s 3,500 FTE students and faculty are supported by the Geoffrey R. Weller Library, a four storey building and sole library on campus. The library’s reference services consist of a standard reference desk and virtual reference (VR), a service provided in conjunction with post-secondary institutions throughout the province.

Like many libraries, the number of reference transactions handled at the Weller Library’s reference desk has been declining annually, at an average rate of 10% per year between 2005 and 2010. While the traditional reference desk may be losing its prominence as the Weller Library’s research help mainstay, it has been supplemented by VR. Patron use of chat reference has increased by a staggering 500% annually since its implementation in 2008. Faced with these numbers, the library decided to investigate alternative modes of reference to reach their patrons. Use of reference services is largely dependent upon the patron approaching the desk, a form of passive librarianship. If the library’s reference services were to become relevant again, an active approach was needed.

Roaming Throughout the Ages: What the Literature Says

The term "roaming" reference has never been clearly defined. Generally, it has been used to describe services provided in a non-traditional manner: roving, outpost, offsite and point of need reference services. In essence, it is anything occurring away from the confines of the reference desk. Although this type of service has been reported in the literature throughout the last three decades, we treat its existence like a covert operation, mainly because we are still struggling with its meaning, provision and effectiveness (Kramer 67).

Roaming and point of need reference services have been implemented with varying degrees of success. The technology used to support such services has varied from tablet computers, microcomputers, desktops and laptops to cell phones. Brenda J. Ford noted in 1986 that fast-paced technological changes would require changes in the way we provide reference service (493). Technology has changed the way librarians interact with the patron (Hibner 19-20), but whether or not it could eventually change the meaning of traditional reference remains to be seen. Could technological changes bring the idea of roaming, a reference model libraries have flirted with for over thirty years, into the mainstream?

The implementation and execution of roaming services require extensive planning, taking service hours, location, approach, evaluation, technology, training and staffing into consideration. Generally, staff attitudes towards the implementation of a roaming service have been underwhelming. There is the view that the provision of point of need services has the potential to debase a librarian’s role as a professional, possibly lessening how faculty and students view their capabilities as research experts (Kramer 69). Others have expressed concern about patron privacy and the impact of staff shortages (Reynolds 62). There is also the impression that providing roaming services in addition to regular reference duties will significantly add to existing workloads (Smith 254). All agree, however, that the success of a new service such as roaming is dependent upon the willingness of staff in its rollout and subsequent provision. It is vital to staff this type of service with individuals who are service oriented and outgoing (Schmel-Hines 12).

While many are cautious in their expressed attitudes towards roaming services, the overall benefits are undeniable. The studies of Davis and Schmehl-Hines found that roaming services increased the visibility of the library whereby "…the focus changes from collection based to a more user centred, teaching and training focus" (Davis 57) (Schmehl-Hines 12). Successful offsite programs have the potential to increase patron-librarian interactions (Holmes and Woznicki 584-585). Taking reference services on the ‘road’ gives librarians ample opportunity to reach the elusive patron who avoids the reference desk and its services. Kramer noted that there was no overlap between questions asked at the reference desk and those asked during roaming shifts, rightly concluding that "…roving reference could significantly increase the quantity of reference that reaches users" (71).

Schmehl-Hines noted that, although the total number of transactions in her study was low, the transactions were of greater difficulty and required expertise to answer. This observation matches Kramer’s assertion that among the benefits of such a service, aside from increased patron contact, are the enhancement of staff skills, the creation of new library policies and the acquisition of resources that better reflect the patrons’ needs (68-74). Roaming has the potential to increase the efficiency and the effectiveness of reference services (Hibner 21) while at the same time instilling patrons with a sense of confidence and comfort.

The location or "path" of a roaming service is very important. Schmehl-Hines noted the importance of identifying and differentiating between academic and social spaces. She found that social spaces alone were not conducive to reference service, but spaces that were academic or academic/social worked best (Schmehl-Hines 12). Nims echoed these sentiments stating that the wrong location can result in poor use of the service. The roaming service at Bowling Green State University was situated in a student computer lab. Nims found that the students were unsure of the library’s presence within ‘their’ space, which left participating librarians feeling like "…intruders in their domain" (87). With that being said, the presence of a roaming librarian can also act as a reassurance to patrons because they know the main role of the roaming librarian is to help (Lorenzen 35).

The overarching goals of this type of service are consistent. "Roaming" services allow librarians to meet patrons at their point of need while increasing the visibility of the library, its staff and resources (Davis 55). This type of service allows the library to become an integral part of the wider institutional culture, building stronger relationships outside the library (Holmes and Woznicki 584). It has the potential to reinvigorate the profession and reference services as a whole by forcing librarians to work outside their comfort zones (Smith 251). Librarians have been debating the issue of change for decades, recognizing the fact that patrons’ needs are changing as fast as the technology they use, yet they seem to balk at the opportunity to change reference services to mirror these changing needs. Ford is correct in her assertion that "the reference desk appears to be a sacred library tradition that many librarians are unwilling or unable to relinquish" (491).

Current Reference Services

Prior to roaming reference services, two reference service points existed at the Weller Library: the traditional reference desk and virtual reference. The reference desk is staffed by 7 librarians and 6 library assistants an average of 36 hours a week. Virtual reference (VR) is staffed by a provincial consortium 67 hours a week. In recent years, UNBC librarians have taken the initiative to create and successfully implement new and innovative outreach programs, which have included office relocation pilots (Fyfe) and outpost reference (McCabe and Fuchs 81-84). In fact, while reference desk statistics were dipping, outreach reference program statistics continued to grow. It was the success of these projects that indicated the potential for roaming reference at UNBC.

As the literature review demonstrates, attempts at defining roaming reference services have been ambiguous at best. For the purpose of this study, roaming reference is defined as the elimination of the physical barriers that the traditional reference desk represents. It uses a combination of virtual reference (chat) and advances in mobile technology (the iPad) to offer face-to-face reference services at a patron’s point of need.

UNBC’s Aims for Roaming Service

The purpose of this service was to explore whether or not roaming was a viable model to implement at the Weller Library. The goal of providing such a service was to continue to develop, and potentially expand, our ability to reach patrons who would not normally approach the desk. Roaming also promised an enhancement to the already existing chat service by ensuring UNBC students were interacting with UNBC librarians during roaming service hours. Normally, patrons using VR would be chatting with a librarian from another institution. Integrating roaming and chat during the proposed service hours could take chat reference one step further. Thus any chat questions from UNBC patrons would be answered by a UNBC librarian during roaming service hours. The proposed change would allow on-campus users to request research assistance through chat at their point of need anywhere on campus. This would also increase the number of opportunities librarians have to reach students and provide instruction. With these aims in mind, the project was conducted between September 2010 and April 2011.

Planning the Service

The timing of the proposed project coincided with Apple’s launch of the iPad. Deciding what technology to use for the project was easy. We were looking for something that was small, light, mobile and user-friendly. The iPad was a natural fit. Using iPads made marketing the service to students very easy. Using new technology that the students wanted to see and use was perfect. It made the recruitment of librarians for the service an easy task too.

Prior to this project, library server logs and reference desk statistics were examined in order to determine hours of service. As described by Smith, reference desk statistics were used to determine peak service hours and the types of questions asked during those times (254). The statistics (2005-2009) revealed that the reference desk received its highest levels of traffic Monday to Wednesday, between 1 and 4 pm. Ideally, the IP addresses of chat patrons could have been used to determine the number of patrons using VR while on campus, but this information is not collected by the consortial chat service due to concerns about patron privacy. As a result, another approach was used to determine patron locations when using library resources.

Unlike previous studies, the library’s server logs were used to determine the geographic location of patrons making page requests to the library homepage. Analysis of the Internet Protocol (IP) addresses indicated whether or not there would be a demand for roaming services. On average, 61% of library page requests originated from within the university. While this number does not represent patrons requiring research or reference help, it is indicative of where they are when they are using library resources.


5 librarians staffed the service during the Fall 2010 term with the understanding that roaming hours were being provided on top of their regular duties. Unlike the experience described by Reynolds and Kramer, UNBC librarians eagerly agreed to participate, embracing the principles behind the pilot, and with the understanding that their involvement was strictly voluntary. Overall, they were expected to be visible and available: wandering stacks, going into computer labs and meeting and greeting students as they roamed. The project leads acted as back-ups during shifts, monitoring the chat queues should the service get too busy and overwhelm the roaming librarian. Project leads were individuals involved in the design and implementation of the service and its technology. These individuals were also responsible for troubleshooting any issues that arose.

A typical iRoam shift lasted 90 minutes. iRoam librarians were asked to spend upwards of 30 minutes roaming the 1st through 3rd floors of the library, with a minimum of 2 roams per shift. Librarians were also encouraged to roam outside the library as well. After each roam, librarians could return to their offices for downtime during which they were still expected to respond to chat requests via the iPad and roam to the student if required.

The Technology

iPads were assigned to librarians two months prior to the commencement of the service to allow them to familiarize themselves with the technology and personalize the iPad to their needs. This approach was taken in order to take the "awkwardness" of new technology out of the equation when the service was launched. Training on how to use the chat application (BeeJive and LibraryH3lp) and the browser (Perfect Browser) was provided (MacDonald and McCabe). The Web Services Librarian was responsible for troubleshooting any technological glitches occurring with the iPads and their applications.

Service Scenarios

iRoam librarians had three service scenarios to contend with. First, as the title of the project indicates, participating librarians could encounter students while roaming, whereby the student approached them or vice versa. The second scenario involved students chatting with the roaming librarian through a chat widget placed centrally on the library’s homepage. A specific chat widget was created for the roaming project to indicate to students that a roaming librarian was on duty. The iRoam widget and regular chat widget were timed to change on the library homepage according to service hours (MacDonald and McCabe). This was done to ensure a seamless transition between iRoam and AskAway chat services for students. Depending on the complexity or appropriateness of the question, service providers were encouraged to ask patrons if they were in the library or on campus. If patrons answered in the affirmative, the provider would then ask the patrons whether they would like the librarian to come to them. Finally, roaming librarians might be paged through the chat widget by staff at the circulation or reference desks to come to the first floor to provide assistance.

Collection of Data

The library uses LimeSurvey open source survey software to capture all of its reference statistics. This software is housed on university servers, thus ensuring that any of the information captured relating to patrons and their behaviour is secure. The reference e-form was adjusted to record additional statistics related to roaming reference. The reference e-form allowed roaming librarians to record each encounter immediately, via the browser on their iPad.

Roaming librarians recorded statistics relating to the time the reference transaction occurred, the type of question they answered (research, facilitative or technology), location of the transaction and approach (librarian to patron, patron to librarian or via chat widget). Patrons were also asked to fill out an e-questionnaire at the end of a roaming transaction. The questions related to their past use of reference services, their thoughts on roaming reference and whether or not the use of this service increased their interest in contacting a librarian for research help. Filling out the e-questionnaire was optional.

Results and Conclusions

Term I: Fall 2010 Results

The first phase of roaming reference was conducted between September and November 2010. Over a period of 36 days and 108 service hours, a total of 158 roaming reference transactions occurred. This number is significant in that, in one term, roaming reference was able to capture 9% of total reference transactions for the library (Fig. 1).

Fig. 1 Total reference questions by mode demonstrating impact of roaming reference.

Overall, the majority of questions were of a research nature. Of the 158 questions, 99 were classified as research, 18 as technology and 41 as facilitative (Fig 2). Technology questions related mainly to printing issues, whereas facilitative questions were both directional and procedural (circulation-related).

Fig.2 Roaming reference questions by type demonstrating breakdown by percentage.

Research questions (n=99) were further classified by the amount of time taken to answer the query. These reference transactions were classified as: basic (1-5 minutes), comprehensive (6-15 minutes) or complex (16 minutes or more) (Fig 3).

Fig. 3 Total research questions according to the amount of time taken to answer query.

The majority (54%) of research questions were received via the chat widget, while the remaining 46% were the result of face-to-face interaction between the librarian and the patron (Fig. 4).

Fig. 4 Research questions received according to approach.

The location of face-to-face research questions was also of interest. Thirty-five percent (n=16) of these transactions occurred on the first floor of the library – the same location as our reference desk (Fig. 5). Given the large number of librarians paged to the reference area, it begs the question, why did students sitting twenty feet from the reference desk prefer to request assistance via chat rather than approaching the reference staff? Further exploration of existing literature would be required to answer this.

Fig. 5 Number of research questions answered by location.

The importance of using a widget as a direct line to the roaming librarian is undeniable. The majority of all roaming reference transactions, 46% (n=72), occurred via chat widget, 31% (n=22) of which resulted in the librarian going to the patron’s location (Fig. 6).

Fig. 6 Location of patrons requesting face-to-face assistance from a librarian via widget.

Of the questions asked during these sessions, 19 were classified as research related and 3 were facilitative. Of these research questions, 8 were basic (1-5 minutes), 8 were comprehensive (6-15 minutes) and 3 were of a more complex nature requiring 16 minutes or more to complete.

Fall 2010 Conclusions

Overall, roaming reference was a success. The results demonstrate the relevance of the service especially when considering the number of research questions received away from the reference desk. Table 1 demonstrates the impact roaming reference had on reference services as a whole. Patron use of roaming reference almost matched UNBC patron use of the provincial chat service, despite being open fewer hours and days per week.

Table 1 Comparing reference services: hours and transactions.


Number of Days Service Provided

Hours per Week (Average)

Total Hours Service Open (Fall 2010)

Total Number of Reference Transactions

Provincial Chat Service










Reference Desk





1 Open 7 days a week closed for Thanksgiving and Remembrance Day.
2 Open 11 hours/day 5 days a week, 6 hours/day Friday and Saturday (exception weeks of Thanksgiving and Remembrance Day)
3 Total calculated by number of reference transcripts

While roaming statistics demonstrate the success of the pilot project, our request for feedback from patrons via the e-questionnaire was disappointing. Overall, there was a 4% response rate to the e-survey. One contributing factor was inconsistency in providing the survey link to users. Service providers admitted that they either forgot or it was just too awkward to provide the user with the link. The few comments that were received (n=5) provided some insight into the potential value of the service. Patrons said: "Answered my cry for help right away and solved my issue!" and "Helpful when they come to you because you don't have to pack up all your millions of stuff to go see them." Despite the lack of formal feedback, these statements indicated that the service was worthwhile and should be tweaked and provided again.

Term II: Winter 2011 Results

For the Winter term, an attempt was made to integrate roaming reference with regular reference desk duties, in hopes of increasing the efficiency of reference services. The assumption being that marrying the two services was possible, as it meant only answering a potential 158 additional questions over a three month period. The goal was to find the right fit for the service and to see if it would continue to thrive when provided in conjunction with the reference desk.

Service days and hours remained the same, Monday to Wednesday 1 to 4pm. Reference desk staff was asked to monitor the chat queue using the reference desk computer. While the chat software used the previous term (BeeJive and LibraryH3lp) worked well, it was believed that software with the ability to create canned messages, push links and surveys, create a personalized widget, and collect IP information would be beneficial to the project. LivePerson, proprietary chat software, fulfilled these requirements.

As in the previous term, service providers were encouraged to roam to chatting patrons, a process made easier by the presence of the patron’s IP address. The fact that LivePerson had the ability to push user surveys to the chatting patron was of great benefit. While there was still an issue with distributing survey links to patrons face-to-face, it was believed this would be negated by the fact that chat was the initial point of contact in the majority of instances.

Project leads continued to provide extended support during the Winter term. The integration of the two reference points meant that there would be an increasing reliance on the back-up service. Project leads were prepared to staff the chat service and to perform roaming duties as there would be times when reference desk staff would be too busy to do either.

Service scenarios for the Winter term changed greatly. Reference staff was asked to notify project leads when leaving the desk to roam the library. During this time, project leads were solely responsible for answering chat requests. The downfall of LivePerson was that it complicated the roaming process greatly. LivePerson’s iPad app turned out to be lacking key features, which meant that chat requests could no longer be reliably pushed to the roaming librarian’s iPad. This process was further complicated by the addition of a separate widget. Should a patron approach the reference desk when the librarian was roaming, a sign instructed them to type their request into a Meebo widget which was left open on the reference desk monitor. This message was then pushed to the roaming librarian’s iPad.

Very early in the term it became apparent that integrating the two services was not going to work. The technological changes were complicating the processes and both services were too busy to be staffed as one. In the end, the integration had drastic effects on roaming services as a whole (Fig. 8).

Fig. 8 Monthly totals for roaming reference showing how the change in service cut the number of questions received via roaming by more than 50%.

The total number of roaming reference transactions dropped from 158 in the Fall term to 70 in the 2011 Winter term. Again, the patterns, although reduced, are generally the same, with research questions making up the majority of transactions (Fig. 9).

Fig. 9 Impact of merging reference desk and roaming service.

It is Fig.10, which truly demonstrates the negative impact of the merged services. The reference desk was too busy to allow the librarian to roam effectively. This is demonstrated by their diminished capacity to interact with patrons in person. During the Fall term, 29% of total roaming reference transactions were the result of librarian- or patron-initiated approach. This number dropped to 8% in the Winter term.

Fig. 10 Research questions by approach and the dramatic shift in results after merging the services.

Although the number of roaming transactions was disappointing, we achieved a 76% response rate to our survey. Sadly, due to the changes made to the service, the feedback received is not reflective of roaming specifically. Instead, the responses relate mainly to the availability of chat. However, 75% of chat requests during this term originated from within the university, 10% of which were in the library.

Winter 2011 Conclusions

The second iteration of roaming reference was not successful. Merging roaming services with the reference desk did not work because conducting two busy services during peak service hours proved impractical. The reference desk was too busy to allow for proper staffing of chat and did not allow librarians to roam the library. If a student paged a librarian to their location, the librarian was too busy to go because she or he had students at the desk. In the end, backup staff members became the main providers: a task which interfered with their own work loads.

The technological changes during the second term were disastrous due to the sheer complexity of the processes. Roaming reference went from an iPad, one app and a survey form, to an iPad, new chat software, a new chat app, two widgets, and a survey form. Not exactly a get up and go service. In addition to these changes, the costs of providing the service exploded. While using LivePerson offered a host of excellent features, it lacked crucial functionality in its iPad app; thus there was no way the costs could be justified on an on-going basis.

The Future of Roaming Reference at UNBC

As important as statistical analysis and user feedback are in determining the importance of a service, the feedback and willingness of service providers truly decides its future. The mistakes made in the second term tested the patience of the original group of service providers and brought the idea of roaming service into question for the newly recruited. The overall consensus was that the original service model was the most efficient and effective way to provide roaming reference. Roaming reference will continue to be provided in Fall 2011, using the initial model with a few alterations and technological simplifications. Roaming reference will be conducted as a standalone service as a result of the 56% decline seen in the number of roaming reference transactions between the Fall and Winter terms.

The service will continue to make use of iPads for roaming. iPads were useful in terms of marketing, in that they were new, sleek and a novelty, but the overall success of the project was not based on their use. What made the service a success was roaming itself. The iPads made service providers visible, but the technology was not always used when answering patrons’ questions. Service providers reported that they were more likely to use the patron’s laptop, lab computer or OPAC station than the iPad when providing assistance. Therefore, all a roaming reference service needs to be a success is staff members who are willing to roam and any mobile device that can tell them where to go.

Overall, 228 roaming reference transactions were conducted, 67% of which were research, 10% technology and 23% facilitative. Twenty four percent of all roaming reference transactions were the result of either librarian or patron approach, a number which would certainly have been higher had roaming reference and the reference desk not been integrated during the service’s second iteration. The combination of the two services did not give service providers the opportunity to roam, demonstrating that it must be a standalone service provided in addition to traditional reference services.

Roaming reference has the potential to reinvigorate reference services, allowing librarians the ability to reach patrons at their point of need and making underutilized research and reference services more relevant. Librarians have flirted with variations of roaming reference for over 30 years. This study demonstrates that with new technologies and dedicated librarians, this type of reference model has great potential to meet the changing information seeking behaviours of library patrons.

Institutions investigating the possibilities of roaming reference services will likely continue to struggle with issues of staffing and workload. With the current doom-saying around the future of the reference desk, librarians will welcome evidence that its presence is still far from irrelevant. Rather, roaming and virtual reference services demonstrate that reference services are needed now more than ever.

Works Cited

Davis, Stephanie R. and Linda Weber. "High Tech, High Touch: Providing Personalized Service on Users’ Turf." Behavioral and Social Sciences Librarian 21.1 (2002):53-60. Library, Information Science & Technology Abstracts with Full Text. Web. 27 Jun. 2011.

Ford, Brenda J. "Reference Beyond (and Without) the Reference Desk." College and Research Libraries 47.5 (1986): 491-494. Print.

Fyfe, Trina. "Librarian Office Relocation Pilot: does office location impact student-librarian interaction?" CHLA/ABSC 2011. Web. 27 Jun. 2011.

Hibner, Holly. "The Wireless Librarian: Using Tablet PCs for Ultimate Reference and Customer Service: A Case Study." Library Hi Tech News 5 (2005): 19-22. Print.

Holmes, Claire and Lisa Woznicki. "Librarians at your Doorstep: Roving Reference at Towson University." College and Research Libraries News 71.11 (2010): 582-585. Web. 27 Jun. 2011.

Kramer, Eileen. "Roving Reference: A Case Study in a Small Academic Library." Reference Services Review 24.3 (1996):67-80. Print.

Lorenzen, Michael. "Management by Wandering Around." The Reference Librarian 28.59 (1997): 51-57. Print.

MacDonald, James, and Kealin M. McCabe. "iRoam: Leveraging Mobile Technology to Provide Innovative Point of Need Reference Services." Code4Lib Journal 13 (2011): n. pag. Web. 27 Jun. 2011.

McCabe, Kealin M., and Willow E. Fuchs. "Point of Need Reference Service." Evidence Based Library and Information Practice 6.2 (2011): 81-84. Web. 27 Jun. 2011.

Nims, Julia K. "Meeting Students on their own Turf." Research Strategies 16.1 (1998): 85-89. Science Direct. Web. 27 Jun. 2011.

Reynolds, Marianne. "Operation Rover." Library Journal 130.7 (2005):62. Academic Search Premier. Web. 27 Jun. 2011.

Schmehl-Hines, Samantha. "Outpost Reference: Meeting Patrons on Their Own Ground." PNLA Quarterly 72.1 (2007): 12-13. Web. 27 Jun. 2011.

Smith, Michael. "Enabling the Roving Reference Librarian: Wireless Access With Tablet PCs." Reference Services Review 32.3 (2004): 249-255. Print.