When I was reporting on the customized TexShare OCLC report last June, commenting on which libraries were overworked and which ones seemed under-utilized, it occurred to me that whether a library was overworked or under-utilized depended a lot on how much staff or equipment was available and how many transactions a library did for all ILL, not just for TexShare. The purpose of the survey we did in fall 1996 was to determine the staff to transaction ratio for each library. This is a simple ratio which describes how many ILL requests are received for each full-time equivalent staff member in the unit. There is comparable data in the world of library literature, and in fact, just last year, we at UT Austin replicated a 1989 survey of academic libraries.
Last year, I heard an interlibrary loan librarian say: "How many requests we process is important; how well we do is irrelevant." Those of us who are committed to interlibrary loan hope this isn't true in most libraries. But any way you look at it, productivity is an important issue as money gets tighter and ILL requests increase. ILL is a labor-intensive activity, and many ILL operations have had to deal with static staffing levels even though there is more work than ever before. We like to believe that automation has helped fill in the widening gap between increasing numbers of requests to process and the number of staff members handling them. Even so, most of us are well aware of our limitations every day as we measure our workload against our ability to complete it in a timely and accurate way.
Even though the newly revised National Interlibrary Loan Code states that "interlibrary borrowing is an integral element of collection development . . . not an ancillary option," libraries do not always support interlibrary loan at a satisfactory level, and library staff is expected to process a steadily increasing number of transactions with no more staff. The interplay among staffing, service quality, and workload are important in determining whether TexShare ILL goals can be satisfactorily met in the interlibrary loan operation.
For the success of a resource sharing project such as TexShare, borrowing and lending must viewed as equally important, although it is not hard to understand why the needs of the library's own clientele might take priority over the needs of other libraries' patrons. However, without the availability of excellent lending service, borrowing fails miserably. Excellence on both sides of the ILL coin are needed to ensure success for the project as a whole.
Staffing is a resource of primary importance. How often do you hear ILL supervisors wishing they had enough staff to process the volume of requests received? How often are ILL staff frustrated by having too much to do? A former staff member in UT Austin's ILS used to joke that ILS meant "I Love Stress" instead of Inter-Library Service.
On the other hand, how do we know when we are understaffed versus when we are not as efficient as we might be? How can we objectively show our library administrators that we are understaffed? One way is to compare our staffing levels to staffing levels in other libraries.
Studies of academic libraries have found a wide range in ILL staffing levels, leading to the conclusion that what is regarded as sufficient staff in one library might not be sufficient in another. Workload -- which includes factors such as volume of transactions, level of automation, and range of tasks performed -- is often used to allocate an appropriate level and size of staff. The relationship between workload and staffing affects productivity and service quality.
Productivity may be defined as output divided by input; that is, the quantity of services produced divided by the resources expended to accomplish the task (Williams, p. 12). Or in ILL terms, productivity is the number of ILL requests processed divided by the number of full-time equivalent staff. As the number of ILL requests increases per ILL staff member, productivity rises. We are doing more with less. This is the staff to transaction ratio.
Staff to transaction ratios provide an objective evaluation of how much we are doing and what human resources are expended in the process. This measure permits comparison among libraries.
As we all know, achieving the highest productivity possible does not necessarily mean that we are doing the best job possible. Productivity must be evaluated in conjunction with our goals and service quality. In ILL, goals such as high fill rates, fast turnaround time, user satisfaction, legible photocopy quality, and accuracy are very important. Statistics may prove that we are highly productive, that we do a lot of work with few resources, but if we send the wrong book or deliver the article long after the patron no longer needs it, we have not met the goals of our operation. Reaching the highest level of productivity possiblewhile still maintaining high service levels should be the desired goal. Determining what the optimal productivity level is is not an easy task and may take considerable experimentation and study.
Some data is available. In 1989, Pat Weaver-Meyers at the University of Oklahoma surveyed members of the Association of Research Libraries and published the results in an ARL paper called Interlibrary Loan in Academic and Research Libraries: Workload and Staffing. She compared staffing and workload in interlibrary loan, using the staff to transaction ratio as a measure of efficiency or productivity. The study found that the average number of transactions processed per lending staff member was 5905; the average number of transactions processed per borrowing staff member was 2439 transactions.
The Texas State Library also uses the staff to transaction ratio as a measure of productivity in the ten major resource centers in Texas. Data collected annually by TexNet show that the staff to transaction ratio is steadily increasing. For fiscal year 1988, an average 6969borrowing and lending transactions were processed per ILL staff member in TexNet libraries. By 1995, this had jumped to 10,317. ILL transactions in public libraries may differ from those in academic libraries, since requests are primarily for books and include referrals from many small public libraries. Therefore, direct comparisons with academic libraries may not necessarily be appropriate but they are interesting nonetheless.
In 1995, we at UT Austin replicated the 1989 survey among Greater Midwest Research Library Consortium members -- for non-football fans, that's the Big 12 plus six who are involved in a cooperative library consortium. The results will be published in a special issue of theJournal of Library Administration -- hopefully in December. This survey found the average lending staff to transaction ratio was 7454 lending requests for every staff member, and 3670 borrowing requests for every staff member. Combining borrowing and lending operations to compare with the TexNet data produced a staff to transaction ratio of 5757, a figure much lower than the 10,317 found in public libraries.
Comparing Weaver-Meyers' 1989 findings with those of the 1995 survey of academic libraries shows an increase of 1549 transactions per FTE in lending (26.23 percent) and an increase of 1231 transactions in borrowing (50.47 percent). This illustrates that ILL requests increased over the past five years at a more rapid rate than ILL staff did. Depending on how you look at it, over the past few years, ILL operations have become more productive or more over-worked.
Both these studies have shown that fewer borrowing requests can be processed per staff member than lending requests. Yet just the other day, a librarian in a very small library -- just three staff members for the whole library -- said she thought thelending requests took a lot more time than theborrowing requests. At UT Austin, a very large library, we have just the opposite situation. We know that the number of lending requests per staff member is lot higher than the number of borrowing requests. We decided that it was because it is more efficient to process a lot of requests than just a few, and when we send a page to the stacks to retrieve 30 journals, it is less time-consuming per journal volume than to send someone to the stacks to retrieve just two journals.
Determining the optimal level of productivity for your ILL operation depends on a number of factors. For example, service quality goals are important. So is the workload, which includes things like the level of automation, the volume of requests received, the variety of tasks donein the unit instead ofby support units, and the availability of equipment. Training and the relative expertise of employees may also affect the optimal staffing level.
A level of productivity or a staff to transaction ratio that produces an efficient service provider in one library may yield an operation that provides substandard service with overworked staff in another.
Everyone wants to be as productive as possible. But is that a desired goal if it negatively affects service quality? Too few staff handling too many requests can lead to poor service, as measured by lower fill rates, slower turnaround time and inaccuracy. There can also be low morale and stress, too much staff turnover and the added expense when new staff must be hired and trained.
Although no survey can define an optimal level of staffing for every interlibrary loan operation, staff to transaction ratios can serve as guides. Libraries below the average can seek ways to streamline procedures and perhaps cut costs by reducing staff, while libraries above average can examine service quality to determine if institutional goals are being met and whether additional staff is necessary.
One place to begin is by comparing your ILL operation to ILL in other libraries. If your staff to transaction ratio is considerably higher than the TexShare average, look for ways to increase your staff or reduce your workload.
To help TexShare libraries compare their ILL operations to those of other TexShare libraries, we sent out the survey you are now hopefully all familiar with. As a reminder, if you haven't returned your survey, it's not too late. Your data will be added to what's already been analyzed, and we'll update the results on the TexShare and LoanStar lists. There are some copies on the table if you need a copy. Non-TexShare libraries are invited to participate too.
We divided responding libraries into three groups by the size of the client base. Group One serves populations under 5000; Group Two serves populations from 5000 to 15,000; Group Three serves populations over 15,000. Group one included ten libraries, group two had 14 libraries, and group three had nine libraries. The statistical table, Borrowing, Lending and Staff to Transaction Ratios in TexShare Libraries, shows which libraries were included in each group.
The statistical summary shows several expected results. First the average number of borrowing requests increased substantially as the size of the ILL constituency increased. Borrowing fill rates were fairly constant regardless of the institution size, although the average fill rate decreased as the size of the constituency increased.
The smaller libraries had comparatively better staffing in borrowing than medium or large size libraries, as illustrated by the lower staff to transaction ratio. Lending staff to transaction ratios were comparable for small and medium sized libraries, although somewhat higher in small libraries. The largest libraries were the most productive in both borrowing and lending -- or should we say, overworked?
The statistical summary also shows that the number of requests received increased with the size of the library. Group three libraries received about three times as many borrowing requests as Group two libraries and about three times as many lending requests. The differences were not as great between Group one and two libraries. Lending fill rates decline as the size of the library increases. And the larger Group three libraries have a much higher staff to transaction ratio than the small and medium sized libraries in Groups one and two.
How do TexShare libraries compare with large academic libraries nationally? For lending, TexShare's average staff to transaction ratio of 5705 is below the Big 12 Plus average of 7454, with Group three libraries more comparable at 7846. For borrowing as well, TexShare's average staff to transaction ratio of 3391 is close to the 3670 average for the Big 12 Plus, except for the smallest TexShare libraries that averaged 2170.
What else did we learn through the survey? That. . .
Regardless of size, libraries indicated no more than two days to fill lending requests, with the majority filling requests in one day. Most libraries processed borrowing requests for faculty, staff and all categories of students. Nine libraries also processed borrowing requests for courtesy borrowers.
We were interested to know what percentage of all borrowed materials were from TexShare libraries. 27 percent of the filled borrowing requests for Group One libraries were TexShare requests; 47 percent of Group Two libraries and 36 percent of Group Three libraries requests were from TexShare. Within groups, however, there was a lot of variation. For most medical libraries, TexShare libraries filled a very small percentage of the total. Most other libraries received between 50 and 75 percent of the total borrowed from TexShare, although one library got 90 percent from TexShare and one got only 9 percent. The three largest libraries -- A&M, Houston and UT Austin -- filled only about one-third of their borrows from TexShare libraries.
Then we looked at the percentage of loaned materials that went to TexShare libraries. Medical libraries that did not charge other TexShare libraries loaned a greater percentage than they received by several times; TexShare lending represented between 10 and 20 percent of the total lending. For most libraries, TexShare borrowing represented a greater percentage of their totals than TexShare lending, although the percentages were usually fairly close.
We also looked at what range of tasks ILL units were responsible for. Most units wrapped some or all of the packages for mailing, and most retrieved the materials for lending from the stacks. Two-thirds prepared invoices, and about half received payments and made deposits.
Five libraries did not have an Ariel workstation; four libraries had two workstations each, and two libraries had three workstations. One library had 11 networked OCLC terminals; five libraries had more than one dedicated OCLC workstation; eleven libraries shared OCLC terminals with other departments.
Conclusions
Check for your library's staff to transaction ratio and see how it compares with the average for your group and for TexShare as a whole. If your staff to transaction ratio is less, you are very fortunate. If it is substantially higher, you need to investigate the quality of the service you provide and perhaps see if there is any way you can add new staff or reduce the your workload.
Statistics can be used very effectively when you need new equipment or additional staff. Compare your activity and staffing with other libraries in TexShare. Compare your current activity and staffing with the situation five and ten years ago. Are you overworked? Do you need to become more efficient? Are there ways to reduce workload? Ask yourself these and similar questions. There are no easy answers, but collecting quantitative data is a place to begin when seeking solutions.
Sources
Weaver-Meyers, Pat, Shelly Clement and Carolyn Mahin. Interlibrary Loan in Academic and Research Libraries: Workload and Staffing. Washington, DC: Office of Management Services, Association of Research Libraries, 1989.
Williams, Robin V. "Productivity Measurements in Special Libraries." Special Libraries 79 (Spring 1988): 101-114.