Ariel—Document Delivery on the Internet

  • Cost-Effectiveness of Ariel for Interlibrary Loan Requests

Cost-Effectiveness of Ariel® for Interlibrary Loan Requests

This is the final report of a completed RLG project.

Summary of a report to RLG SHARES participants, March 6, 1996

I. Project goals
II. How the project was conducted
III. Anticipated results
IV. Principal project findings
V. Other findings
VI. Conclusions

At the end of 1993, the RLG Shared Resources Executive Group (SEG), an oversight committee for RLG's Shared Resources (SHARES) activities, approved a set of recommendations for a study of alternative methods to US mail for filling interlibrary loan copy requests.

The study would be conducted by institutions participating in SHARES and using the RLIN interlibrary loan system, which provides data on the monthly and annual lending and borrowing transactions among SHARES participants.

A SHARES Working Group on Rush Requests and Turnaround Time had recommended this project, to demonstrate whether or not turnaround time could be reduced through the use of Ariel (RLG's document-transmission system for the Internet) or fax. The group also wanted to gather data on the ease and cost-effectiveness of using Ariel and/or fax to delivery all copy requests.

They felt that this data would be useful to library directors making equipment investments to enhance access services within their libraries; and that it would support recommending Ariel as the routine method of delivery in RLG libraries.

A SHARES Ariel/Fax Pilot Project Task Force designed the project in 1994. Its members—

Anne Beaubien, University of Michigan, Chair
Sue Lorimer, Yale University
Anne Cerstvik Nolan, Brown University
Keith Rageth, University of Iowa

—originally envisioned a six-month trial to be conducted from August 1994 through January 1995, during which participants would send all ILL photocopies via Ariel or fax. In the end, the trial period was compressed to one month, March 1995, to obtain the information needed with minimal disruption for the participants' work flow.

The test was moved to the early spring to give participants an opportunity to obtain the new Ariel for Windows software (released in August 1994); become comfortable with its use; avoid the fall and winter holidays, the American Library Association's midwinter meeting, and academic-calendar spring breaks; and have a full calendar month of RLIN-generated ILL statistics on which to base turnaround calculations.

Suzanne Lorimer designed the reporting forms with feedback from the rest of the task force. RLG staff provided the initial data coding and analysis to help ensure that calculations were done in a uniform manner.

The task force calculated average cost figuresfor telecommunications costs to send a 10-page article via fax at $1.23; and the average cost to send the same article via US mail at $.78. RLIN statistics track data for copy response time, enabling the task force to use these in measuring turnaround time during the trial.

The trial was conducted March 1-31, 1995. 26 ILL units representing 20 SHARES institutions participated. 22 ILL units provided complete turnaround-time data used in Table II below.

I. Project goals

  • To improve document delivery turnaround time among SHARES participants by demonstrating the reduced turnaround time possible with Ariel and fax;
  • To gather data on the ease and cost-effectiveness of using Ariel 2.0, Ariel for Windows, and/or fax for delivery of all copy requests.

II. How the project was conducted

  1. Copy requests were sent via the RLIN ILL system and filled in standard SHARES turnaround time. No requests were accepted over Ariel or fax.
  2. This project studied only transmission of photocopies via Ariel or fax. During the project, participants tried to fill all copy requests via Ariel or fax; however, documents over 50 pages long were sent via US mail.
  3. Transmission of Copies:
    • Libraries with Ariel first tried to send copies via Ariel. The automatic default (28 re-sends) was the maximum required before an alternative method of delivery was used. If Ariel delivery attempts were unsuccessful, fax was used.
    • If a fax site was down for more than 24 hours (either sending or receiving), the request was sent via the US mail.*
    • US mail was used when Ariel and/or fax either did not work or were unavailable at a particular site.

      *Note: RLG libraries who are also members of the Committee on Institutional Cooperation (CIC) were to use CICfaxnet when the fax option was utilized. (CIC includes all the "Big 10" institutions plus the University of Illinois at Chicago, the University of Chicago, and Pennsylvania State University.) CICfaxnet was a parallel technology to Ariel which scanned and sent articles via the Internet using a fax machine. However, CICfaxnet was discontinued in January 1995. All CIC libraries now use Ariel.

  4. Delivery cover sheets:
    • Ariel. The fully completed Ariel header (including patron name and RLIN ILL record ID) was sufficient for Ariel requests. A copy of the request form was not needed.
    • Fax. A cover sheet for fax transmissions was not required. However, a readable copy of the request form indicating the patron name and RLIN ILL record ID had to be used. If the order form was too light to copy, a cover sheet with the patron name and RLIN ILL record ID was used.
  5. Re-send requests:
    • Ariel was the method of choice for requesting Ariel re-sends if transmissions were not clearly legible for any reason.
    • Fax was the method of choice for requesting fax re-sends if transmissions were not clearly legible.
  6. Data tracked daily:
    • Total number of re-sends, with a notation as to why the re- send was necessary.
    • Total number of articles received/sent during the project via Ariel and total number received/sent via fax.
    • Reasons for defaulting to US mail delivery.
    • Elapsed time from request made by ILL staff until document arrival at borrower's ILL office.
    • Total number of Ariel scans from original document.
    • Total number of Ariel scans from a photocopy of the document.
  7. Background information collected:
    • Kind of equipment used for photocopying, scanning, and transmission.
    • Staffing configuration.
    • Total number of requests processed for other pilot project participants.
  8. Data sampled during one-week period of the March trial:
    • Average length of articles sent.
    • Staff time required to:
      1. Scan and send articles via Ariel (separate data was collected for scans directly from bound material and for scans from photocopied material);
      2. Photocopy and fax articles; and
      3. Photocopy and send articles via US mail.
    • Whether staff were doing multiple tasks, e.g., entering requests while scanning and/or sending photocopies.

III. Anticipated results

Based on previous knowledge and experience, the task force expected project findings to include:

  1. Improvement of turnaround time for delivery of copies among SHARES partners.
  2. Increased information about the impact on staffing of sending all copy requests via Ariel or fax.
  3. Increased information on how well Ariel works in a high-production environment.
  4. Increased information on the cost differences among various delivery options.
  5. That sending articles via Ariel or fax would take the same as or less staff time than using US mail and would cost the same as or less than sending via US mail, while delivering photocopies more rapidly.
  6. That the use of high technology such as high-speed photocopiers and scanners in access services is a cost-effective investment.

IV. Principal project findings

The actual results of the completed project were in line with these expectations:

1. Sending articles via Ariel or fax took less staff time than via US mail, cost less than US mail, and decreased receipt time for both borrowing and lending.

The data demonstrate that on a cost basis alone, the use of Ariel is more cost-efficient than either fax or mail for the delivery of materials (See Table I below.)

Of the 12 possible combinations of level of staff (professional, paraprofessional, clerical, and student) and delivery method (Ariel, fax,and mail) the five most cost-efficient were:

a. Ariel with student staff (average cost for 10-page article = $.69).
b. Mail with student staff (average cost for 10-page article = $1.11).
c. Mail with clerical staff (average cost for 10-page article = $1.30).
d. Ariel with paraprofessional staff (average cost for a 10-page article = $1.32).
e. Ariel with clerical staff (average cost for 10-page article = $1.82).

The study did request each participant to indicate the cost of its monthly network charge for the Ariel-required Ethernet connection. Thirteen respondents did not report a charge. The eight reporting sites averaged a cost of $19.15 per month (ranging from $0 to $25). Since the study did not request the monthly cost of a telephone line for a fax machine, the average Ethernet cost figure reported was not included in the computation of amounts reported in the Ariel columns of the "Cost to Send" section of Table I. However, the amortized per-transaction cost for an Ethernet connection would have to rise to at least $1.52 for Ariel to be a more costly delivery method than fax.

All computations disregarded startup costs.

Table I: Costs

Professional
Paraprofessional
Clerical
Student
Staff Costs
Avg hourly cost
$27.290
$15.850
$15.560
$5.860
Avg per-min cost
0.455
0.264
0.26
0.098
Ariel per-page
Time (mins)
0.500
0.500
0.700
0.700
Cost
0.228
0.132
0.182
0.069
Fax per-page
Time (mins)
1.200
0.900
0.900
1.000
Cost
0.546
0.238
0.234
0.098
Mail per-page
Time (mins)
N/A
N/A
0.300
0.600
Cost
N/A
N/A
0.078
0.059
Cost to Create
Ariel
2.280
1.320
1.820
0.690
Fax
5.460
2.380
2.340
0.980
Mail
N/A
N/A
0.780
0.590
Cost to Send
Ariel
0.000
0.000
0.000
0.000
Fax
1.230
1.230
1.230
1.230
Mail
N/A
N/A
0.520
0.520
Total Cost
Ariel
2.280
1.320
1.820
0.690
Fax
6.690
3.610
3.570
2.210
Mail
N/A
N/A
1.300
1.110

2. Ariel's ease of use, whether in scanning directly from the volume or scanning from a photocopy, was apparent.

The average amount of staff time used per page—.76 minutes with direct scanning and .66 minutes when scanning from photocopy—demonstrated the ease of use and efficiency of both the Ariel and fax machines.

While high-speed scanners are not necessarily any easier to use, they can take 3 seconds per page compared to 10 seconds. Since this allows a higher volume of material to be copied using less staff time, the high-speed scanner can pay for itself in staff time saved and devoted to other work. (It should be noted that during the project, only two participants were using high-speed scanners, though several sites have subsequently installed them.)

3. The statistics collected for turnaround time during the study reflect positively on the potential use of technology to increase receipt time and to better serve library users' needs.

Of the 22 ILL units that provided data for the turnaround-time portion of the study, eight experienced a decrease in both borrowing and lending receipt time. Thirteen libraries experienced a decrease in receipt time for either borrowing or lending, and only one library experienced no decrease in receipt time. (See Table II below.)

Quicker turnaround time was accomplished in spite of the fact that the data shows an overall increase of 4% in total copy requests filled in March 1995 compared to March 1994. The increase underestimates the true growth of ILL traffic because calculations were limited to traffic among project trial ILL units during one calendar month.

Note: "Receipt time" in these findings is based on the date that the borrowing library updated the RLIN ILL record to "received". During this study it became clear that differences in institutional updating practices skew the statistics considerably. For example, receipt times of 54.3 days and other prolonged periods generally indicate that some records for filled requests were not being updated promptly. It is possible that the study results would have been even more positive had timeliness of updating not been an issue.

Table II: Receipt time

Library
Total
Filled
Total
Filled
Received
Time
(Days)
Received
Time
(Days)
Received
Time
Difference
(Days)
Mar-94
Mar-95
Mar-94
Mar-95
A: Borrowing
40
22
5.7
6.1
0.4
A: Lending
29
33
54.3
11.9
-42.4
B: Borrowing
93
44
6.8
23.0
16.2
B: Lending
12
7
7.8
4.1
-3.7
C: Borrowing
86
122
6.5
5.5
-1.0
C: Lending
143
198
23.5
6.1
-17.4
D: Borrowing
143
180
10.1
8.0
-2.1
D: Lending
187
182
13.7
28.5
14.8
E: Borrowing
37
44
0
3.6
3.6
E: Lending
50
70
9.1
15.4
6.3
F: Borrowing
29
18
3.8
11.8
8.0
F: Lending
42
90
4.6
3.6
-1.0
G: Borrowing
2
1
11.0
1.0
-10.0
G: Lending
6
31
1.2
19.3
18.1
H: Borrowing
18
37
7.6
14.5
6.9
H: Lending
24
14
9.5
3.9
-5.6
I: Borrowing
89
34
7.4
6.1
-1.3
I: Lending
122
156
32.7
9.4
-23.3
J: Borrowing
366
140
7.1
5.8
-1.3
J: Lending
468
582
8.0
5.9
-2.1
K: Borrowing
116
106
10.3
9.3
-1.0
K: Lending
346
369
7.4
5.6
-1.8
L: Borrowing
347
358
7.6
21.5
13.9
L: Lending
94
173
7.9
4.5
-3.4
M: Borrowing
139
137
10.9
8.0
-2.9
M: Lending
291
361
15.9
6.7
-9.2
N: Borrowing
79
51
6.9
7.5
0.6
N: Lending
143
176
11.8
10.1
-1.7
O: Borrowing
285
308
11.4
8.1
-3.3
O: Lending
150
163
6.2
7.0
0.8
P: Borrowing
35
58
15.5
19.4
3.9
P: Lending
1
5
7.5
0.6
-6.9
Q: Borrowing
425
461
9.2
6.5
-2.7
Q: Lending
147
166
7.0
7.5
0.5
R: Borrowing
650
529
6.5
4.5
-2.0
R: Lending
288
447
7.8
7.2
-0.6
S: Borrowing
9
16
7.3
5.3
-2.0
S: Lending
62
62
24.8
7.1
-17.7
T: Borrowing
445
228
6.0
4.3
-1.7
T: Lending
86
128
7.7
4.8
-2.9
U: Borrowing
34
46
3.7
2.7
-1.0
U: Lending
52
80
6.3
22.2
15.9

V. Other findings

  1. Borrowers reported that they received a total of 1,530 filled requests during the project. The average length of an article transmitted during this project was 10.07 pages.
  2. 977 requests (64% of all requests filled) were filled via Ariel. Borrowers encountered problems with 17.8% of these requests. The most frequently reported problems were "Text Cut Off" (24%) and "Missing Pages" (24%).
  3. 301 requests (20% of all filled requests) were filled via fax during the project. Borrowers encountered problems with 14% of these requests, the most frequent of which were "Text Cut Off" (24%) and "Missing Pages (24%).
  4. 252 (16% of all filled requests) were filled via US mail. One borrower problem was reported: "Missing Pages."
  5. Lenders reported problems with 15% of their Ariel transmissions. The most common difficulty (46%) was "Re-send limit reached." Some site-specific reception difficulties were encountered in the early stages of implementing Ariel for Windows that have since been resolved. That makes it likely that this problem would be less significant if another pilot were to be conducted in March 1996. Twelve problems were reported by lenders using fax, the most common of which (33.3%) was "original too large."

The number of problems reported overall was minimal when compared to the total number of requests and the total number of pages sent. The most commonly reported receipt problems could be related to human error as well as to possible transmission problems. In the case of direct scanning as well as in photocopying to scan with Ariel, incorrectly positioning the volume, slightly moving the volume, or skipping a page while scanning/copying could account for the most often reported borrower problems.

Over one-half of the lender problems were related to the inability of Ariel to connect to the borrower's machine, a problem which could be ameliorated by leaving Ariel "on" after work hours to ease reception across time-zones.

VI. Conclusions

The task force considers the project goals to have been met.

Overall, the study demonstrated that using electronic technologies such as Ariel and fax improves turnaround time for library users. Timely service becomes even more important as library staff search for quick access to articles that are not available through commercial document delivery vendors. SHARES participants rely heavily on each other's collections, and as a consortium RLG and its members should use their resources to get materials to each other in a timely fashion using technology that already exists.

The ease of using Ariel and fax was also demonstrated. Many sites use students successfully in the operation of Ariel, and this should be expanded on throughout the SHARES community. The task force also felt that the use of high-speed photocopiers and scanners could make further contributions to the cost-effectiveness of both Ariel and fax. Leaving Ariel and fax machines in operational mode overnight can ease problems caused for lenders that are located in distant time zones but wish to supply items in a timely manner.

Human behavior is still an important part of the puzzle. Timely updating of ILL transaction records is essential if ILL units are to be accountable for maintaining responsiveness to each other's users. Careful photocopying is necessary to ensure that the complete, legible text is supplied.

When the Task Force began its work in 1994, Ariel had not yet been as widely adopted as it is today. It is no longer necessary to demonstrate its usefulness as a delivery mechanism. However, Ariel can not achieve its full potential in isolation. As electronic document delivery from paper-based collections matures into a primary strategy for providing information to users, access services operations must invest in all the appropriate tools to support efficient operations. High-speed scanners, photocopiers, and high-quality delivery mechanisms are all essential for cost-effective operations. With adequate investment in the right tools, access services can become a cost-effective and reliable arm of library services.

We are a worldwide library cooperative, owned, governed and sustained by members since 1967. Our public purpose is a statement of commitment to each other—that we will work together to improve access to the information held in libraries around the globe, and find ways to reduce costs for libraries through collaboration.