College and Research Libraries ' . FRED J. HEINRITZ Quantitative Management in Libraries Based on a position paper originally presented at the Institute on Quantitative M·ethods in Librarianship at Ohio State University Li- braries in August, 1969, the author discusses some of the elements of management: motion, time and cost studies, operations research and other mathematical techniques, and data processing equipment. INTRODUCTION M ODERN MANAGEMENT is a fusion of various pioneer approaches to the sub- ject.1 The operational school, popular- ized through the work of the French industrialist Henri Fayol ( 1841-1925), identified the activities or functions of managers and from these derived his principles of management. There is gen- eral agreement that the four major func- tions of ·all managers are planning, or- ganizing, staffing, and control. The traditional approach of the op- erational school has not been quantita- tive. Mathematics, or operations re- search, has been utilized· extensiv~ly by business and government administrators in planning and control only since World War II. In the planning propess numerical methods have proven par- ticularly useful in evaluating alternative courses of action-that is, in providing a rational basis for decision making. The mathematics associated with oper- ations research can be quite sophisticat- ed, including calculus and linear pro- gramming, and with a heavy emphasis on mathematical models and optimal solutions. The quantitative approach to control was developed by a second major Dr. Heinritz is Professor of Library Sci- ence at Southern Connecticut State Col- lege, New Haven, Connecticut. 232/ management school. This approach, known as scientific management, achieved prominence through the work of the American engineer Frederick W. Taylor (1856-1915). 2 This school was concerned primarily with developing fair performance standards. Motion and time study techniques were developed and utilized. Although the approach from the beginning has been quantita- tive, the mathematics utilized tends to be much simpler than that of operations research. Often only addition, subtrac- tion, multiplication, and division are needed. When equations are needed, basic algebra will usually suffice. Although the library profession has not entirely ignored quantitative man- agement, neither has it adopted it with any enthusiasm. Only one library school, Rutgers, has consistently emphasized this subject and made it an integral part of the curriculum. This was due primarily to the impact of Professor Ralph Shaw of Rutgers, long a leader in this area. The library literature de- voted to quantitative management also is not extensive. In 1954 an issue of Li- brary Trends (edited by Shaw) was devoted to "Scientific Management in Libraries."3 But it was not until 1966 that the first book devoted entirely to applying quantitative management techniques to libraries appeared, writ- ten by former students of Shaw. 4 There is, however, gradually increasing library interest in the subject. Quantitative Management in Libraries I 233 MoTION STUDY The logical and necessary first step in the control process is to establish stan- dards. Thus the development and utili- zation of performance standards is at the heart of scientific management. Since it makes no sense to perpetuate an inefficient procedure, and since a standard time has no meaning except in relation to standard methods, motion study is the prerequisite to time study. The chief technique for motion study is the chart, of which there are many varieties. Each chart has its strengths and weaknesses, and for any given management situation there will be one or more which is appropriate. The two major types are process and decision charts. The process group includes the How process chart, form process chart, gang chart (for group efforts), and op- erations chart (for hand and foot mo- tions at one work station). 5 They are appropriate for charting r~petitive pro- cedures, that do not involve too many yes-no alternatives. Decision charts, on the other hand, are appropriate for re- cording operations with many possible subprocedures-a common case in li- braries.6 Activities charted may range over an entire building, or take place at a single desk. The basic procedure is to record the present method, to analyze it, and from this analysis to develop an im- proved method. The improved method may be timed in order to develop per- formance standards and/ or used ( in chart form) as a procedure manual for staff reference and the training of new employees. As may be seen from the examples listed in the references, as well as many oth~rs which have ap- peared in the literature, charting has been employed to describe and analyze a wide variety of library activities, both in technical and public services, and has even been used to analyze the ref- erence process. Motion study is the most useful of f .. . scientific management activities in li- braries at the present time, and the sav- ings it can produce are substantial. The fact that the mathematics involved is minimal is important because the math- ematical ability of some librarians is al- so minimal. Charting is a possible un- dertaking where time study is not fea- sible-all persons experienced in mak- ing management studies know the re- sistance encountered when a time study is made. The typical employee finds a chart much less offensive than a watch. TIME STUDY Time study is the most reliable meth- od for determining performance stan- dards. However, unless it is used with discrimination, it may result in more harm than good. Although industrial employees do not like the idea of being timed,- they are at least used to the idea, as industry has timed operations for fifty years. In libraries, however, the idea is new, and employee resist- ance can be expected to be substantial- ly greater. Professional librarians. some- times particularly resent the idea of their work being quantified. There is no question,. that it is easier to develop time standards for clerical than profes- sional .-tasks. Noneth~less, professional library work does involve the expendi- tqre of time, and the possibility of time Jstudf should not be ruled out. - Even when time standards are impractical it may be possible to develop verifiable qualitative objectives and standards. Another major obstacle in developing time standards in libraries is the neces- sity for rating the worker. The timer must decide whether the employee be- ing observed is working faster or slower than "normal" so that he can adjust the observed time accordingly with a quan- titative rating factor. Clearly, consider- able judgment and experience are nec- essary - bere. Even professional study analysts find rating a difficult task. Thus it is advisable that libraries apply time analysis with considerable caution, and 234 I College & Research Libraries • July 1970 certainly not until thoroughgoing motion study analysis has been carried out. Another technique that can be used to collect time data, which avoids the use of a stop watcht is work sampling. Random observations are made to de- termine the proportions of productive and idle time, and the average rating factor for a job. These data, along with the number of units produced and the worker allowances for personal needs, fatigue and delay, provide the data needed to compute standard times. 7 CosT Cost standards are usually not as use- ful as time standards. Since cost is a function of time, the times must in any case be known before the costs can be computed. A time standard based on a standard method should be valid in any area so long as the work is the same and the method is adhered to. Costs, in contrast, will vary according to the labor and other rates for a given re- gion or city, as well as with time. Gross salary figures are useful in es- timation, but for analytical purposes it is necessary to relate cost to production. Many librarians would be · aghast (if they knew them ) at their actual labor costs per productive hour as compared to the hourly wage. With the decreas- ing length of the work week, and the increase in employer-paid benefits, this differential can only increase. If in ad- dition to worker allowances, vacation, paid holidays , and sick leave, we add space costs, equipment depreciation and maintenance, supplies, supervisory costs, and . time spent on conferences with su- pervisors or in staff meetings, the actual cost per productive hour may be close to twice the hourly wage. In estimating costs, it is sometimes sensible to ignore overhead and/ or de- preciation and/ or supply costs. How- ever, they should not be omitted out . of ignorance, for to do so is to delude one-· self that the cost of an operation is less than it really is. Sometimes these costs are of real utility. For example, the idea of depreciation is most helpful in but- tressing arguments for an expensive piece of equipment in those cases where a library system considers equipment purchases as current expenditures. CoMPARISON OF TIMES AND CosTs BE1WEEN LIBRARIES Once time data and/or costs based on them are collected by a library it is natural for it to want to compare them with data for similar operations in oth- er libraries. This is not easy to do. Even when such data have been published, there are many variations in the meth- odology used for collecting, organizing and disseminating them. In addition, the tasks timed are seldom defined in sufficient detail, if at all, for library journals are limited in space, and the desired -details are usually difficult to come by-if they were recorded in the first place. Full-text technical reports containing quantitative library data are increasingly available, but, without stan- dardized collection and reporting pro- cedures, such data are of limited utility for comparative purposes. Ideally, data collected by one library could be compiled into a catalog of standard motions or tasks with times as- signed to each operation. Another li- brary could then consult the catalog to learn how much time it would take it to perform a given task, just as a mathe- n1atician consults a table of trigono- metric functions, rather than work out the value of the sine, cosine or tangent .. of a particular angle each time he needs it. A small start has been made at de- termining standard time units for cer- tain repetitive library operations such as posting book pockets, both manual- ly and by machine, using an electric eraser, finding the "Secret" page, and the like, but a profession-wide effort is required for significant progress em- bracing broad areas of library activity. • ' ., .. ~ ' ~ · Quantitative Management in Libraries I 235 OPERATIONS RESEARCH Operations research was first used ex- tensively during the Second World War. Although it has since been used widely in industry, it has to date received little attention from librarians. A substantial portion of the literature applying op- erations research to libraries has been written by nonlibrarians: industrial en- gineers and other scientists. These sci- entists are too often either unable or un- willing to write in a style appropriate for librarians. Some librarians, on the other hand, are ignorant even of basic mathematics which they might reason- ably be expected to know. The result is a communications gap of substantial proportions. Operations research problems are generally classified as belonging to the following types: allocation, sequencing, inventory, replacement, queueing theo- ry, and competitive strategies. 9 ·In allo- cation the problem is to distribute a limited resource, such as space or labor or dollars, in the most effective way. For example, mathematical models have been developed for optimal shelving of books by size;10 for allocating personnel ·in a technical services operation so as to e~minate backlogs and achieve a minimal overall per-volume processing time; 11 and for creating standards for evaluating budget proposals of a uni- versity library.1 2 In sequencing we are concerned with the order of operations. There are, for example, procedures for determining the minimal time or cost for searching a group of bibliographic tools and for visiting service points in the routing of centrally processed ma- terial.13 Inventory theory is concerned with balancing the economies and costs as- sociated with different levels of goods on hand. It has been applied by a pub- lisher in determining how many copies of a back list title to reprint.14 Queueing theory, which analyzes the interrela- tions between customer arrival, waiting, and servicing, has been used in libraries to help predict the effect of providing duplicate services, usage restrictions, and reserves.15 Little has been done by libraries to date in the areas of replace- ment and competitive strategies.16 OTHER MATHEMATICAL TECHNIQUES Mathematical techniques not leading to optimal solutions may still be useful for predictive purposes. Library collec- tions and catalogs are growing larger. If, for example, library records show a particular collection or catalog to be growing at a more or less constant rate, it is possible to utilize compound in- terest tables and equations to predict its size at some future time, or the point in time when a particular size will be achieved. 17 As books grow older, they are, on the average, used less. When this decrease in use occurs at a fairly constant rate, the same basic for- mulas and others derived from them help us to make rational judgments as to an appropriate age at which to re- tire books into depository storage for less active materials.1s Traditional statistical procedures have also proven helpful in quantita- tive library. management. The single most useful technique is sampling-es- pecially random sampling. Over and over again it happens that a librarian needs to take a sample, and does not have any clear notion of how to do it. This is a serious problem, for most de- tailed discussions of sampling require considerable mathematical sophistica- tion.19 DATA PROCESSING EQUIPMENT Machines are not a satisfactory sub- stitute for either mathematical ideas ·or- techniques, or the creative imagination required to see their application to li- brari~. The advent of data processing equipment, ., however, has stimulated in- terest in scientific management in li- braries. The expense of automating an operation, and of making major changes 236 I College & Research Libraries • July 1970 once the procedure is operational, en- courage careful planning. One of the serious faults with some of the current efforts to automate has been the failure to examine and evaluate adequately present practices. As a result the auto- mated system, based on an inefficient manual system, is cumbersome and overly expensive. The computing power of the digital computer, which in some cases allows problem solving approaches that were not practical with hand calculation, has been used extensively in operations research. Sometimes, in lieu of an avail- able analytical technique, a problem can only be solved by evaluating a very large number of possibilities, and then selecting the desired one (say minimal cost) from the group. The cost and tedium of doing this by hand is pro- hibitive. t5'ifficult equations may some- times be solved by iteration-that is, by successive approximations. Where the number of approximations is large, the computer may still be able to achieve a solution. It is possible to use the com- puter to select a random sample. This technique can be useful in connection with simulated, or Monte Carlo, sam- pling: when it is too expensive or even impossible to take an actual sample, a s.ample described by some appropriate probability distribution is substituted. There has to date been little use of these new quantitative possibilities in libraries. CoNCLUSION The fact is that quantitative manage- ment is not used extensively in libraries at the present time. It is not difficult to find reasons for this, the basic one being that the majority of librarians are hu- manities-trained and oriented. They are · products of an educational system that allows students to acquire graduate de- grees in English literature, for example, without having any clear idea even of elementary mathematical techniques and applications. It is unrealistic to ex- pect the library profession to modify the American system of education. It is, however, quite possible for library schools to pay at least passing attention to quantitative ability in assessing ad- missions standards, to make clear to stu- dents at the outset that some work with numbers is necessary and to be expect- ed, and to insert a reasonable amount of quantitative management into the curriculum, preferably within the broad- er general framework of administration. The need is not to make industrial en- gineers or mathematicians out of library school students, but rather to inculcate the idea that the utilization of num- bers (the language of size) to solve li- brary problems is not incompatible with loving books and people. The real value of quantitative management comes not so much from special management studies as from its absorption into the everyday-thinking and activities of the average librarian. A second problem is that most li- braries are still quite small units of op- eration. Most of the quantitative man- agement done has been in connection with larger library systems. Librarians are busy people, and things do not get done unless there is somebody to do them. Larger libraries often have a spe- cial person, or even several, who devote their full time to work improvement, standards, and other research. A few university libraries have even made use of their local Department of Industrial Engineering. 20 Sn1all libraries seldom have such specialized talent at hand. Nor do they usually have access to high- ly specialized equipment, such as high speed computers. The trend toward larger units of li- brary service is gradually alleviating the problem of minuscule library sys- tems. Attention to quantitative matters in recruiting and training librarians is another matter. This will require a con- scious, coordinated effort on the part of library leaders and educators. Books ~ I J • ... t I I ~· Quantitative Management in Libraries I 237 appropriate as texts are now available, 21 and many library schools have at least one faculty n1ember with some numer- ical interest and facility. Bibliography is the foundation stone of librarianship; but a few more stones are required to complete the building. REFERENCES 1. For a history of management and an overview of the various schools of management see: Harold Koontz and Cyril O'Donnell, Principles of Man- agement (4th ed.; New York: Mc- Graw-Hill, 1968), Chapters 1 and 2. 2. Taylor's major writings are still in print, and are of particular interest to students of quantitative management: Frederick W. Taylor, Scientific Man- agement, Comprising Shop Manage- ment, the Principles of Scientific Management, Testimony Before the Special House Committee, with a Foreword by Harlow S. Person (New York: Harper, 1947), 3v. in 1. Prin- ciples and Testimony are written for the layman. Shop Management is more technical. 3. Ralph Shaw, ed. "Scientific Manage- ment in Libraries," LibTary Trends, II, no. 2 (Jan. 1954). 4. Richard M. Dougherty and Fred J. Heinritz, Scientific Management of Li- brary Operations (Metuchen, N.J.: Scarecrow Press, 1966). 5. For library-oriented examples of these process charts see Dougherty and Heinritz, Scientific Management, Chap- ters 3 and 5. 6. Louis A. Schultheiss and others, Ad- vanced Data Processing in the U ni- versity Library (Metuchen, N.J.: Scarecrow Press, 1962); Dougherty and Heinritz, Scientific Management, Chapter 4; G. Carlson, Search Strat- egy by Reference Librarians; Part 3 of the Final Report on the Organiza- tion of Large Files. NSF Contract C- 280. 17 March 1964. Advance Infor- mation Systems Division, Hughes Dy- namics, Inc., cj o 3943 Woodfield Drive, Sherman Oaks, Calif.; F. S. Stych, "Teaching Reference Work- The Flow Chart Method," RQ', vol. 5, no. 4 (Summer 1966), p.14-17. 7. Ralph M. Barnes, W ark Sampling (2d ed.; New York: Wiley, 1957). See especially Chapter 9, "Determin- ing Time Standards by Work Sam- pling." For work sampling applied specifically to libraries see: Scott T. Poage, "Work Sampling in Library Administration," Library Quarterly, vol. 30, no. 3 (July 1960), p.213-18. 8. Henry Voos, "Standard Times for Cer- tain Clerical Activities in Technical Processing," Library Resources & Tech- nical Services, vol. 10, no. 2 (Spring 1966)' p.223-27. 9. There are many operations research texts in print. The best known is: C. West Churchman, Russell L. Ack- off, and E. Leonard Arnoff, Intro- duction to Operations Research (New York: Wiley, 1957). For a more re- cent treatment see: Russell L. Ackoff and Maurice W. Sasieni, Fundamen- tals of Operations Research (New York: Wiley, 1968). 10. Ferdinand F. Leimkuhler and J. Gra- dy Cox, "Compact Book Storage in Libraries," Opetations Reseatch, val. 12, no. 3 (May-June 1964), p.419-27.' 11. Fred J. Heinritz, "Optimum Alloca- tion of Technical Services Personnel," Library Resources & Technical Ser~ vices, val. 13, no. 1 (Winter 1969), p.99-101. 12. Powell Niland, "Developing Standards for Library Expenditures," Manage- ment Science. Series B-Managerial, vol. 13, no. 12 (August 1967), p.B797- B808. 13. Gerald J. Lazorick and Thomas L. Minder, "A Least-Cost Searching Se- quence," College & Research Libraries, vol. 25, no. 2 (March 1964), p.126-28. See also: Ashby J. Fristoe, "The Bit- ter End," Library Resources & Tech- nical Services, vol. 10, no. 1 (Winter 1966), p.91-95; Fred J. Heinritz and 238 I College & Research Libraries • July 1970 James C. Hsiao, "Optimum Distribu- tion of Centrally Processed Material," Library Resources & Technical Ser- vices, vol. 13, no. 2 (Spring 1969), p.206-08; James C. Hsiao and Fred J. Heinritz, "Optimum Distribution of Centrally Processed Material: Multi- ple Routing Solutions Utilizing the Lockset Method of Sequential Pro- gramming." Library Resources & Tech- nical Services, vol. 13, no. 4 (Fall 1969), p.537-44. 14. Leonard Shatzkin, "That Solid Back- list Title: How Many to Print?," Pub- lishers Weekly, vol. 191, no. 24 (12 June 1967), p .31-32. 15. Philip M. Morse, Library Effectiveness: a Systems Approach (Cambridge, Mass.: The Massachusetts Institute of Tech- nology Press, 1968). 16. The classic lay introduction to the analysis of competitive strategies is: J. D. Williams, The Compleat Strategyst (rev. ed.; New York: McGraw-Hill, 1966). 17. Fred J. Heinritz, "Predicting the Need for Catalog Expansion," Li- brary Resources & Technical Services, vol. 11, no. 2 (Spring 1967 ), p.247- 48. 18. Ferdinand F. Leimkuhler, "Systems Analysis in University Libraries," Col- lege & Research Libraries, vol. 27, no. 1 (January 1966) , p.13-18. 19. A good introductory article is: M. Carl Drott, "Random Sampling: a Tool for Library Research," College & Re- search Libraries, vol. 30, no. 2 (March 1969), p.119-25. See also: Dougher- ty and Heinritz, Scientific Manage- ment, Chapter 8. For a nontechnical overview of the subject read: Morris J. Slonim, Sampling: a Quick Reliable Guide to !Practical Statistics (former title: Sampling in a Nutshell) (New York: Simon and Schuster, 1966). 20. For example, Purdue University, the University of Michigan, and the Mas- sachusetts Institute of Technology. 21. In addition to those mentioned above, the library educator may also find useful: Barton R. Burkhalter, ed., Case Studies in Systems Analysis in a University Library (Metuchen, N.J.: Scarecrow Press, 1968). ~ 1 I I ~ I 1