College and Research Libraries H. WILLIAM AXFORD Perforn1ance Measuren1ent Revisited Performance measurement, through unit-cost study programs, can be a beginning step toward achieving the goal of evaluating the value of systems. This article reports the application of such unit-cost studies in .the technical service functions at a large state university, and pre- sents three rnafor tables for labor costs in terms of minutes and dol- lars per volume. IN THE SPRING OF 1971, the author pub- lished the results of a unit-cost study of the technical services division at Florida Atlantic University. As initially con- ceived, this project had two major pur- poses: ( 1) to determine the direct labor costs of acquiring and fully processing a volume in terms of both minutes and dollars per function performed by level of employee; and ( 2) to test the efficien- cy of processing techniques which, at the time, were considered by some in the profession to be radical departures from standard library practices.1 In the more than eighty responses to the arti- cle, some maintained that what was pos- sible at a small, relatively insignificant institution such as FA U would be sheer folly to attempt in a large research li- brary supporting a broad-spectrum grad- uate program and a research-oriented faculty. The critics' fears turned out to be un- founded. At the time the article was published, the author was implementing both the unit-cost study program and the processing systems developed at Flor- ida Atlantic University at Arizona State University, an institution five times the size of FAU, with 27,000 students, 1,200 faculty members, and a library collec- tion in excess of 1,000,000 volumes, Dr. Axford is director of libraries, U n.i- versity of Oregon, Eugene. growing at the rate of over 100,000 vol- umes a year. Within a year, the bulk of an unprocessed backlog, which exceeded 50,000 volumes, had been eliminated, a huge logjam of faculty book requests had been cleared, and the library had ceased to be a major source of irrita- tion to the faculty, students, and the university administration. Concrete evi- dence of a rising confidence in the li- brary's ability to perform manifested itself in the increasingly genial and pro- ductive meetings of the University Li- brary Committee, a group which had formerly been torn by administrative problems not within its technical capa- bilities. Processing costs had been cut by 36 percent without tampering with the bibliographic integrity of the public catalog. During the past five years, an increas- ing number of academic librarians have become painfully aware of what Earl F. Cheit has called the "New Depres- sion in Higher Education."2 Book budg- ets have plateaued or have been dras- tically cut, new positions have not been forthcoming, and freezes on filling existing and future vacancies have been common. Magnifying the impact of re- duced or static budgets is the fact that there has been no discernable leveling of demands for library services. Real progress in making the library a vital and dynamic center for inspiration and information, and an intellectually in- / 249 250 I College & Research Libraries • September 1973 spiring place in which to work, cannot be gained during a severe budget crisis unless our labor-intensive organizations can achieve a higher level of manpower utilization than is now generally the case. Public service programs needed in the near future will have to be created through more efficient utilization of the manpower already at our disposal. As Peter Drucker has forcefully pointed out, the present budget crisis is in es- sence a productivity crisis affecting the nonmanual worker and the knowledge professions. "The only way out of it," he notes, "is for the nonmanual employ- ee, whether he is knowledge worker or policeman, to become more produc- t . "3 1ve. Concern with productivity does not, and should not imply that the knowl- edge worker in America is lazy. With re- spect to libraries, the worker has become a victim of an antiquated organization- al structure which is not only unproduc- tive and self-defeating, but intellectual- ly and psychologically debilitating as well. It is for this reason that Drucker, taking the larger view, believes that the present budget crisis is the best thing to happen to higher education in recent years: it provides a golden opportunity for the university to evaluate critically every system and procedure in terms of goals and priorities, and to weed out those which have become obsolete due to technological change and/ or the changing needs of society. In the library we need to evaluate every system and procedure in terms of what it conb·ib- utes to the user, and jettison those which only serve to make the profession comfortable. Performance measurement can be a beginning step toward achieving this goal. It is particularly effective when ap- plied to any operation which is process oriented and in which data can be easily quantified. The technical services oper- ation meets these criteria ideally. In ad- dition, it is this area of the library which clings most stubbornly to anti- quated procedures, and whose only an- swer to low output is a loud wail for more and more people. Little thought is ever given to whether or not those already on hand are being used effective- ly. Even less is given to a critical exam- ination of library systems in terms of how well they serve the user. It is for these reasons that technical services will often absorb 50 percent or more of the total personnel budget of a large li- brary. And still, the unprocessed orders pile up and the cataloging backlog con- tinues to grow and oppress those who labor within its ominous shadow. To librarians who have been forced to trim budgets to the bone at a time when libraries are besieged with rising demands for services, the demand to trim off even more seems insulting. N ev- ertheless, there is no such thing as an or- ganization in which manpower utiliza- tion is 100 percent effective. Moreover, budget trimming by itself is not an ade- quate response to a long-term budget crisis. In its initial stages it affects only marginal things. When the process goes deeper the result is an across-the-board reduction of program quality rather than a rational decking of priorities. Budget cutting alone seldom, if ever, al- ters the basic internal systems and proce- dures of libaries, particularly with re- spect to the acquisition and processing of library materials. The programs needed to streamline library operations and make them more productive will require willpower to ini- tiate and discipline to carry out. Agen- cies outside of the library are capable of providing both if librarians do not take the initiative. It is clearly in the in- terest of the profession and its users that the motivation for change be in- ternal, and that the necessary discipline be self-imposed. Departmental self-measurement was the methodology utilized by the unit- cost study developed at Florida Atlantic , : J Performance Measurement Revisited I 251 University and is now operational at Arizona State University. The individu- al departments and the Technical Ser- vices Division as a whole define their functions, establish their own produc- tivity goals, and measure achievements in terms of them. The measurement of the program's success ultimately lies in the extent to which this concept is un- derstood by all concerned. The unit-cost program has now been operational for two years at ASU, and it has been applied to three production years, 1969 I 10, 1970 I 71, and 1971172. In order for the program to have any meaning, it must be reiterated every year; since there are no national stan- dards to measure against, the library must establish a base year against which the productivity of all subsequent years can be measured. At ASU the base year is 1970171, as that was the year in which new processing systems and procedures were implemented. Table 1 shows the minutes per volume by level of employ- ee by department for each of the three production years subjected to the pro- gram. Table 2 shows the costs in terms of dollars. During the base year, the Technical Services Division processed 154,437 vol- umes, approximately 52,000 more than the year before, and used 7,270 fewer man hours. The total minutes required rr to order and fully process a volume dropped from 101.46 to 64.52. In other words, in 1970/71 the Technical Services Division achieved a 36 percent increase in productivity in spite of a 4 percent decline in the number of man hours as- signed to it. If the same processing sys- tems and procedures used in 1969 I 10 had been used in 1970 I 71, it would have required approximately 46 additional F.T.E.'s to produce the 154,437 volumes actually processed in 1970/71. In the following year, 1971172, the cost in minutes to acquire and process a volume rose from 64.52 to 77.08 min- utes. Since the total cost in minutes per volume was 12.2 minutes higher in 19711 72 than in 1970 I 71, it would appear at first glance that there was a significant drop in productivity in 1971172. This, however, is not the case. The catalog department bears the burden of the unit-cost study program, as it is the end of the processing pipe- line where the number of work units completed is counted. The catalog de- partment catalogs titles, not volumes, but the unit-cost study measures produc- tivity in terms of volumes. Therefore, in order to compare productivity from one year to the next, the number of volumes produced in a given year must be adjusted to conform to the ratio of volumes per title cataloged which pre- vailed during the base year. In 1970/71, the base year at ASU, the ratio of volumes to titles cataloged was 2:1. This dropped to 1:1.71 in 1971172. Had the ratio of the base year prevailed in 1971172, the 65,754 titles cataloged would have produced 131,508 volumes, or 18,009 more than the actual produc- tion count. When this factor is taken into account, the loss in productivity drops to an insignificant 2 percent. Ta- ble 3 shows the adjusted costs in minutes per volume. As can be seen in Table 3, the num- ber of hours worked in the Technical Services Division dropped by 27,516, or 15 percent, between 1969170 and 19711 72. This is the equivalent of 14.46 F.T.E.'s. The distribution by level of employee is shown on Table 4. The increased productivity achieved at ASU was the result of a number of minor and several major procedural and system changes. Minor changes included not under lining the first letter of the main entry on the title page, not pencil- ing the call number on the Title Page, and keypunching only the title on the machine readable book card. If these labor savings seem insignificant, it should be noted that for a library proc- essing 100,000 volumes a year, a one- TABLE 1 ~ LABOR COSTS (MINUTES PER VOLUME) TECHNICAL SERVICES DIVISION too .......... 1969/70--1971/72 0 0 1969/70 ( 102,308 vols.) 1970/71 ( 154,437 vols.) 1971/72 ( 113,499 vols.) ~ ~ Total Minutes Total Minutes Total Minutes ~ Hours Per Volume Hours Per Volume Hours Per Volume (';) c;- Acquisitions ~ Professional 4,590 2.69 2,700 1.05 3,666 1.94 (';) Subprofessional 13,125 7.70 8,126 3.16 8,704 4.60 "-> (';) Clerical 14"062 8.25 13,765 5.35 19,021 10.06 ~ Student Assistants 8,402 4.93 5,979 2.32 3,461 1.83 Total 40,179 23.57 30,570 11.88 34,852 18.43 ~ ~ Bibliographic Search ~. Professional 1,800 1.06 1,800 .70 Functions transferred to ~ Subprofessional 18,750 11.00 13,819 5.37 · Acquisitions and Catalog a ""'\ Clerical 3,750 2.20 -- Departments ~. (';) Student Assistants 11,726 6.88 5,887 2.29 "-> Total 36,026 21.14 21,506 8.36 • ~ Cataloging (';) Professional 19,170 11.24 19,666 7.64 18,638 9.85 ~ ~ Subprofessional 28,751 16.86 26,809 10.42 25,183 13.31 (';) Clerical 24,375 14.30 30,691 11.92 29,292 15.48 ~ ~ Student Assistants 8,271 4.85 8,979 3.49 12,517 6.62 (';) Total 80,567 47.25 86,145 33.47 85,630 45.26 ""'\ Physical Preparation "- ~ Professional 0 0 0 0 0 0 Cj Subprofessional 3,375 1.98 1,688 .66 0 0 Clerical 4,967 2.91 5,156 2.00 8,201 4.34 Student Assistants 710 .42 2,465 .96 1,925 1.02 Total 9,052 5.31 9,309 3.62 10,126 5.36 Serials Professional 0 0 1,800 .70 1,438 .76 Subprofessional 3,907 2.29 8,756 3.40 0 0 Clerical 2,295 1.35 4,650 1.81 11,758 6.22 Student Assistants 1,280 .75 3,300 1.28 1,995 1.05 Total 7,482 4.39 18,506 7.19 15,191 8.03 Grand Total 173,306 101.66 166,036 64.52 145,799 77.08 TABLE 2 LABOR CosTs (DoLLARs PER VoLUME) 1969/70--1971/72 1969/70 ( 102,308 vols.) 1970/71 ( 154,437 vols.) 1971/72 ( 113,499 vols.) Total Dollars Total Dollars Total Dollars Dollars Per Volume Dollars Per Volume Dollars Per Volume Acquisitions Professional 23,540 .23 13,941 .09 20,639 .18 Subprofessional 31,500 .31 22,649 .15 25,770 .23 Clerical 27,400 .27 31,107 .20 48,226 .42 Student Assistants 12,971 .13 9,206 .06 5,678 .05 Total 95,411 .94 76,903 .50 100,313 .88 Bibliographic Search Professional 7,800 .08 8,900 .06 Subprofessional 39,500 .39 31,622 .20 Functions transferred to ~ Clerical 7,400 .07 0 0 Acquisitions and Catalog Student Assistants 17,573 .17 9,408 .06 Deparbnents -4. Total 72,273 .71 49,930 .32 c '"t Cataloging ~ Professional 88,840 .87 101,192 .66 107,832 .95 Subprofessional 63,770 .62 66,459 .43 68,275 .60 ~ <:') Clerical 46,456 .45 65,359 .42 67,308 .59 ~ Student Assistants 12,590 .12 14,471 .09 20,598 .18 a:: Total 211,656 2.06 247,481 1.60 264,013 2.32 ~ ~ Physical Preparation ~ Professional 0 0 0 0 0 0 ~ '"t Subprofessional 7,425 .07 3,780 .03 0 0 ~ ~ Clerical 9,550 .09 11,593 .08 20,815 .18 ~ Student Assistants 996 .01 4,219 .03 3,309 .03 ~ Total 17,971 .17 19,682 .14 24,124 .21 ~ ~ Serials ~ Professional 0 0 8,600 .06 9,130 .08 0 ~- Subprofessional 8,220 .08 19,822 .13 0 0 ~- Clerical 4,450 .04 9,290 .06 29,504 .2.6 ~ ~ Student Assistants 1,904 .02 5,778 .04 3,435 .03 ~ Total 14,574 .14 43,490 .29 42,069 .37 ........... Grand Total 411,885 4.02 437,486 2.85 430,519 3.78 ~ TABLE 3 t1 ADJUSTED LABoR CosT (MINUTEs PER VoLUME) joj:::o. 1969/70--1971/72 .......... CJ 0 1969/70 (102,308 vols.) 1970/71 ( 154,437 vols.) 1971/72 ( 131,508 vols.) .,....,. Total Minutes Total Minutes Total Minutes ~ Hours Per Volume Hours Per Volume Hours Per Volume ~ ~ Acquisitions G- Professional 4,590 2.69 2,700 1.05 3,666 1.67 ~ ~ Subprofessional 13,125 7.70 8,126 3.16 8,704 3.97 c, ~ Clerical 14,062 8.25 13,765 5.35 19,021 8.67 ~ Student Assistants 8,402 4.93 5,979 2.32 3,461 1.57 d Total 40,179 23.57 30,570 11.88 34,852 15.88 ;:r- Bibliographic Search t""4 ~. Professional 1,800 1.06 1,800 .70 ~ ""t Subprofessional 18,750 11.00 13,819 5.37 Functions transferred to ~ Clerical 3,750 2.20 0 0 Acquisitions and Catalog ""t ~. ~ Student Assistants 11,726 6.88 5,887 2.29 Departments c, Total 36,026 21.14 21,506 8.36 • Cataloging C/) ~ Professional 19,170 11.24 19,666 7.64 18,638 8.50 ~ ~ Subprofessional 28,751 16.86 26,809 10.42 25,183 11.48 ~ Clerical 24,375 14.30 30,691 11.92 29,292 13.36 ~ Student Assistants 8,271 4.85 8,979 3.49 12,517 5.71 ~ ~ Total 80,567 47.25 86,145 33.47 85,630 39.05 ""t ,....,. Physical Preparation <:o Professional 0 0 0 0 0 0 2:} Subprofessional 3,375 1.98 1,688 .66 0 0 Clerical 4,967 2.91 5,156 2.00 8,201 3.74 Student Assistants 710 .42 2,465 .96 1,925 .87 Total 9,052 5.31 9,309 3.62 10,126 4.61 Serials Professional 0 0 1,800 .70 1,438 .65 Subprofessional 3,907 2.29 8,756 3.40 0 0 Clerical 2,295 1.35 4,650 1.81 11,758 5.36 Student Assistants 1,280 .75 3,300 1.28 1,995 .91 Total 7,482 4.39 18,506 7.19 15,191 6.92 Grand Total 173,306 101.66 166,036 64.52 145,799 66.46 -r - C::: Performance Measurement Revisited I 255 TABLE 4 HoURS ExPENDED BY LEvEL OF EMPLOYEE 1969/70-1971/72 Total Hours Professional Subprofessional Clerical Student Assistants TOTAL 1969!70 25,560 67,908 49,449 30,389 173,306 minute reduction in the time it takes to process a volume is the equivalent of .91 F.T.E.'s in terms of labor savings. 0 The category of minor changes also includes all measures that resulted in a more efficient utilization of personnel, particularly at the professional level. Table 5, extracted from the technical services unit-cost studies, provides an ex- cellent example. Almost 40 percent fewer professional hours were expended in 1970171 as compared with the previous year, but these were far more efficiently utilized. For instance, in 1969170 slightly more than one-half F.T.E. professional was absorbed by subprofessional routines, i.e., assigning vendor and fund num- bers, revising typing, signing purchase orders, and bursting forms. With re- spect to the latter, more than two weeks of professional time was devoted to this simple function during the year. This is an example of how a small leak will, in time, result in a sizable puddle. In 1969170 only 29.4 percent of the pro- fessional hours in the department were expended on administration and super- vision; in 1970 I 11 the figure rose to 78.6 percent. The last category of minor procedur- al changes included eliminating obso- lete files. Two glaring instances were discovered in the catalog department which absorbed approximately the labor of a half-time person. ° Computed on the basis of the ASU stan- dard work year for a nonprofessional employee ( 1,830 hours). 1971/72 + or- F.T.E. 23,742 - 1,818 - 1.05 33,887 -34,021 - 18.74 68,272 +18,823 +10.37 19,889 -10,500 - 5.04 145,790 -27,516 - 14.46 The major changes which resulted in increased productivity were as follows: ( 1) Eliminating establishing the main entry prior to placing an or- der. ( 2) Splitting the catalog into its three component parts, author, ti- tle, and subject. ( 3) Color highlighting in lieu of typ- ing added entries at the top of the cards and filing behind head- ers in the subject catalog. ( 4) Leaving the call number in the lower left hand corner of the card. ( 5) Using the title catalog as the «on order" file. The savings in ( 1) can be substantial, as much of the information obtained in the "pre-cataloging" processing has to be revised after the material arrives. In addition, there is the perennial prob- lem of the catalog department accept- ing bibliographic information generat- ed by another department. A consider- able amount of wasted time can be avoided if the acquisitions department confines itself to determining if the li- brary has an item, if it is on order, or if it exists, and leaving the de1icriptive cataloging to be performed by the cata- log department after the item arrives. Following this procedure, cost in min- utes per volume for searching and veri- fying dropped from 7.78 minutes in 1969 I 10 to 4.21 minutes per volume processed in 1970171, a decrease of 45 percent. Since the total cost in rninutes per volume for the catalog department 256 I College & Research Libraries • September 1973 TABLE 5 ANALYSIS OF PROFESSIONAL HOURS ASU AcQUismoNs DEPARTMENT 1969/70-1970/71 Administration and supervision Review requests and selection aids Search and verify bibliographic information Assign vendor and fund number Revise typing, sign, and mail purchase orders Burst order forms Receiving routines Process faculty inquiries Order O.P. titles Miscellaneous activities also dropped significantly ( 29 percent), obviously changing procedures did not just shift costs from one department to another. Items ( 3) and ( 4) (color highlight- ing instead of typing added entries and leaving the call number in the lower left hand corner of the card) produced a labor savings of 53 percent in produc- ing card sets. In a test environment at FA U in 1969 (simultaneous production of 100 card sets by typing added entries at the top of the card, and 100 card sets by color highlighting) the labor savings amounted to 71 percent. The difference between the test figure and that derived from the ASU unit-cost study probably was due to the fact that supervision in an actual working environment cannot possibly approach the level that is pos- sible in a test environment. The FA U study showed what is possible. The ASU cost study showed what one library ac- tually achieved. Quantitative analysis can be a power- ful tool in the administrator's kit to in- crease productivity in those areas of the library which are process oriented, up- grade the quality of the work per- formed , and provide a better and more satisfying working environment as well. At ASU, increased productivity and more complete cataloging have gone hand-in-hand. As the unprocessed back- log melted away it was possible to pro- 1969/70 1,350 990 0 622.5 330 82.5 0 210.5 105 150 4,590 1970!71 2,124 115.2 165.2 79.2 0 0 18 0 0 0 2,700 vide, as a routine matter, analytics for all titles in series, something that was a hit or miss affair during the years when the catalog department was unable to cope with the volume of material com- ing into the library. The author catalog is being read, corrections made, and hundreds of new header cards are being added. Subject entries for new serials are being prepared for the first time in a number of years, and work is progress- ing on eliminating any gaps in this area. Pockets of "difficult" material which had been gathering dust in obscure nooks and crannies for months or even years have been cleaned up. Time has been found for thinking about and planning for the future, and the entire staff of the catalog department recent- ly participated in a complete depart- mental reorganization designed to make the best possible use of talent at all lev- els. Criticism aimed at the unit-cost study program is the accuracy of the data, particularly with respect to the distri- bution of an employee's time over the range of functions for which he is re- sponsible. The margin of error seems to be tolerable, as supported in an ar- ticle by R. K. MacLeod, who analysed an almost identical cost study program designed for the South Shore Mental Health Center in Chicago. On the sub- ject of the accuracy of the data, Mac- 1 Performance Measurement Revisited I 251 Leod concluded that "even a rough idea of the cost of a program is so useful that arguments about precision are re- duced to the level of quibbles."4 Because of its very nature, it is prob- ably impossible to initiate a unit-cost study program without producing a measurable amount of adverse reaction on the part of those working in the technical services division, particularly among the professional members of the staff. On the other hand, when correctly used, the unit-cost study program can enhance staff development and staff self- esteem. Through participating in an on- going unit-cost study program, each member of the technical services divi- sion has an opportunity to analyze both his own and his department's efforts and make suggestions for system streamlin- ing. Through this process, the creative energies of the group are channeled into significant contributions toward re- designing systems. Leadership may crop up in places which do not coincide with the power structure. In such cases, if the power structure listens and capitalizes on what it learns, it will recognize that "good employees" need not be "yes men," that creative discussion, not an intolerable level of interpersonal con- flict, has a chance to develop. Christopher Morley once observed that "there is no squabbling so violent as that between people who accepted an idea yesterday and those who will ac- cept the same idea tomorrow." Some form of performance measurement, or accountability, will become a standard administrative tool in libraries and a part of the normal working environ- ment of every librarian in the near fu- ture. If we can avoid expending pre- cious energies on useless and unproduc- tive intramural squabbling and divert them toward designing performance measurement procedures of, by, and for librarians, we will not only be working in the best interests of the profession, but those of the communities that sup- port and sustain libraries as well. REFERENCES 1. H. William Axford, "An Approach to Per- formance Budgeting at Florida Atlantic Uni- versity," CRL 32:87-104 (March 1971). See also H. William Axford ( ed.), Proceed- ings of the LARC Association Institute on Unit Cost Studies (Tempe, Ariz., LARC As- sociation, 1972). 2. Earl F. Cheit, The New Depression in High- er Education (New York: McGraw-Hill,. 1971 ). 3. Peter Drucker, "The Surprising Seventies," Harpers Magazine 243:35-39 (July 1971). 4. Roderick K. MacLeod, "Program Budgeting Works in Nonprofit Institutions," Harvard Business Review 49:52 (Sept./Oct. 1971). MacLeod is a partner in the prestigious ac- counting firm of Lybrand, Ross Bros., and Montgomery.