OM 2

Operations Management Part 4 Supplementary Chapter A

Pr C

im, it takes 2,000 hours to build each electrical generating turbine, so if we have to build ten, it takes 20,000 hours. We should plan our budget and price per generator based on 20,000 hours, exclaimed Pete Jacobs, the vice president of nance. No, Pete. According to my calculations, it will take only 14,232 hours to build ten turbines and our total cost and budget will be much lower than you think, said Jim Conner, the vice president of operations. How do you get such crazy numbers? replied Jacobs.

Stefan Eduard/Plainpicture/Jupiter Images

What do you think?

Have you had an experience in which the time it took to perform some activity (such as solving a Sudoku puzzle or playing an X-box game) improved as your learning increased?

A2
62564_09_cA.indd 2

Part 4: OM2 Supplementary Chapters

en

ga

e r o ge f ty Le ar ni ng
learning outcomes
After studying this chapter you should be able to: measurement. methods.

op

WORK MEASUREmENT, LEARNING CURVES, AND STANDARDS

L01 Explain the purpose of work

L02 Describe how to apply time study

L03 Explain the principles and calculation

methods of work sampling.

L04 Explain the concept of a learning

curve and their application for estimating production times.

8/26/09 4:03 PM

A
Time standards represent reasonable estimates of the amount of time needed to perform a task based on an analysis of the work by a trained industrial engineer or other operations expert. Jacobs cannot understand the discrepancy between his estimate of 20,000 hours and Conners value of 14,232 hours to produce a batch of turbines. The assembly of electrical powergenerating turbines is a complex job with labor costs for engineers and production employees approaching $100 per hour. Obviously, a difference of 5,768 hours can be significant in terms of cost, budgets, and pricing decisions. Where did Conner get his figure? Moreover, why should the total time to produce ten turbines be less than 10 times the time to produce the first? Many work tasks show increased performance over time because of learning and improvement. Failure to recognize this can lead to poor budgeting, erroneous promises for delivery, and other bad management decisions. In this supplementary chapter we introduce work measurement, standards, and learning curves, and how they are used in business. Most large corporations develop standard times for routine work tasks using work measurement. They are used in setting job performance standards, establishing recognition and reward programs, and for compensation incentives. Valid standard times are vital to accomplishing most of the process design and operations analysis methods described in this text. Smaller businesses, especially service businesses, usually do not have such standard times for their work activities and tasks. However, if one seeks to improve operations, analyzing work and determining standard times for key work activities and processes is a crucial first step.

Mason Mort/Getty Images

en

ga

OM2 Supplementary Chapter A: Work Measurement, Learning Curves, and Standards

62564_09_cA.indd 3

e r o ge f ty Le ar ni ng
A3

C H A P TER

Pr

op

OM
8/26/09 4:03 PM

1 Work Measurement

perations managers are interested in how long it takes to create an output or outcome, or equivalently, how much can be produced over a certain length of time. Work measurement is a systematic procedure for the analysis of work and determination of times required to perform key tasks in processes.Work measurement leads to the development of labor and equipment time standards that are used for

operating conditions. The prescribed method to perform a task is usually developed by industrial engineers who identify the most efficient and safest procedure. However, not everyone works at the same pace, and people may either slow down or speed up their effort when they are being observed. Thus, observed times must be adjusted by a factor that accounts for the workers effort. Normal times are calculated using the following equation: Normal time 5 Observed time 3 Performance rating factor 5 OT 3 PRF

[A.1]

estimating work-force and equipment capacity, establishing budgets, determining what new work procedures will cost,

evaluating time and cost trade-offs among process design alternatives, establishing wage-incentive systems, monitoring and evaluating employee performance and productivity, and

providing accurate information for scheduling and sequencing.

Without accurate time standards it is impossible to perform these tasks. For example, the process of assemblyline balancing, discussed in Chapter 8, requires accurate estimates of the standard time required to perform each task or work activity. Standard times are managements anchor in an uncertain operating environment. To establish usable standards, work tasks and activities must be carefully defined and studied. Thus, job and process analysis should precede Work measurement work measurement. How is a systematic procedure long it takes to perform a for the analysis of work and task depends on the workdetermination of times required to perform key tasks ers pace, operating condiin processes. tions, and work method. Normal time is the exNormal time is the expected pected time required to pertime required to perform form some work activity at a some work activity at a normal pace, under normal operating conditions, and normal pace, under normal using a prescribed method. operating conditions, and Allowances include time using a prescribed method. for labor fatigue and personal By a normal pace, we mean needs, equipment breaka pace that can be consisdowns, rest periods, information delays, and so on. tently performed by the average employee without Standard time is normal time adjusted for allowances. undue fatigue under normal

ga

The performance rating factor (PRF) is a judgment made by the person doing the time study as to whether the employee is working at the normal pace (that is, 1.0 or 100 percent), below the normal pace (that is, less than 1.0 or 100 percent), or above the normal pace (that is, greater than 1.0 or 100 percent). For example, a PRF of 115 percent indicates that work is being performed at a pace that is 15 per- cent above normal. Typically, three or more highly trained work study analysts make these judgments independently and then the average PRF is used in equation A.1. For example, if work study analyst A rates an employee at PRF 5 1.2 and an observed time of 2.5 minutes per unit, B rates the same employee at PRF 5 1.0 and an observed time of 2.2 minutes per unit, and C rates the same employee at PRF 5 0.9 and an observed time of 2.1 minutes per unit, then, using equation A.1, the normal time is 2.363 minutes per unit (1.2 3 2.5 1 1.0 3 2.2 1 0.9 3 2.1)/3. Normal times must also be adjusted for personal time and unavoidable delays. Allowances include time for labor fatigue and personal needs, equipment breakdowns, rest periods, information delays, and so on. Most allowance factors are in the range of 10 to 20 percent. Standard time is normal time adjusted for allowances. It is computed using the following equation: Standard time 5 Normal time (1 1 Allowance factor) [A.2]

en

A4
62564_09_cA.indd 4

Part 4: OM2 Supplementary Chapters

e r o ge f ty Le ar ni ng
For example, if the normal time is 2.363 and an allowance factor of 1.2 is used, the standard time is 2.836 minutes per unit. The performance rating factor and allowance factor are based on human judgments, and therefore, the procedure for establishing standard times must include audits, third-party reviews, and extensive training and retraining of work-study analysts to keep them properly calibrated. Videotapes of employee work activities are often used to train work-study analysts.
8/26/09 4:03 PM

Pr

op

Work Measurement Activities in Roller Coaster Maintenance

A popular roller coaster called the Runaway Mine Train (RMT) requires extensive inspections, maintenance, and repair to keep it running and earning revenue. During peak season, the RMT is expected to operate 16 hours per day. Each part of the RMT, from structural steel uprights to the bearings in the wheels, must be inspected and well maintained. All RMT daily work tasks are grouped into work activities such as complete train inspection, track inspection, electrical inspection, cleaning the trains, and vehicle inspection. Each work activity is assigned a craft employee such as a track and vehicle machinist, electrician, sound engineer, custodian, oiler, software and computer operator, and so

Solved Problem

Observation Cycles (cumulative, in minutes) Work Task 1 2 3 4 5 6 7 8 9 Average Count 2 $20s Difference Count 10 $10s Difference Count 12 $5s Difference Count 100 $1s Difference 0.11 0.11 0.25 0.25 0.37 0.37 1.02 1.02 0.15 0.21 0.66 0.28 0.79 0.14 2.44 1.26 0.36 0.14 0.94 0.33 0.93 0.51 3.70 1.31 0.50 0.15 1.27 0.52 1.44 0.47 5.01 1.38 0.65 0.13 1.79 0.26 1.91 0.39 6.39 1.72 0.78 0.11 2.05 0.31 2.30 0.58 8.11 1.32 0.89 1.02 1.22 0.13 0.20 2.36 2.77 3.20 0.41 0.43 2.88 3.61 4.22 0.73 0.61 9.43 10.55 11.55 1.12 1.00 0.15 0.35 0.48 1.27

ga

Assume your first job out of school is as a branch manager for a major bank. Your branch also has a back-office operation where bank packs of $300 are assembled for retail businesses. Each bank pack contains two $20 bills, ten $10 bills, twelve $5 bills, and one hundred $1 bills. Retail customers become upset when these bank packs are not available since they usually pick them up before their retail store opens. A continuous stopwatch study collected the information shown in the table below. The policy of the bank is to use a 20 percent allowance factor for branch bank operations. (a) What is the normal and standard time for a bank pack? (b) How long would it take to pack

en

Place Bank Packs in Tray (once every ten cycles) 5 2.50 Standard time per bank pack 5 Normal time 3 (1 1 Allowances) 5 2.8426 Standard time per ten bank packs with setup 5 31.426

OM2 Supplementary Chapter A: Work Measurement, Learning Curves, and Standards

62564_09_cA.indd 5

e r o ge f ty Le ar ni ng
200 packs? Now you must decide how many branch bank staff to assign to this work activity and when they will do the work.

on. Sixty-one distinct jobs existed for this attraction but only the jobs related to inspection, maintenance, and repair were to be time studied. A sampling plan was established and the data collected. Once time standards were determined for all maintenance-related tasks, the number of employees required was computed. Preventive maintenance was done each Wednesday and required additional changeover or setup time. Two teams were established, each with four people. One team worked Sunday to Wednesday and the second team from Wednesday to Saturday. The two teams overlapped on Wednesday for one-half day to get all preventive maintenance done. Work measurement information and analysis plays a major role in achieving the objectives of operating safety, efficiency, and profitability.1

Pr

op

Solution:

a. To determine the normal and standard times, note that the continuous stopwatch study timed each denomination separately and the cumulative times are shown across each row. The Difference rows show the individual times.

b. (200 bank packs)(31.426 minutes)/10 bank packs 5 (20)(31.426 minutes) 5 628.5 minutes or 10.5 hours

Performance Rating 1.15 0.95 1.00 1.10 1.00

Normal Time 0.170 0.331 0.475 1.393

Normal processing (run) time per bank pack 5 2.369 Normal time per ten bank packs with setup 5 26.188

A5
8/26/09 4:03 PM

2 Time Study Methods

ime study is the development of a

measurements, and then grouping or separating tasks accordingly. 2. Measure and record the time needed to perform each task or activity over a number of cycles. A trained observer with a stopwatch usually does this. A number of observations should be taken to account for variability in performance. Assuming that the distribution of task times is normally distributed for each task, the number of cycles that should be observed is determined statistically by the sample size (n) formula, equation A.3:

standard time by observing a task and analyzing it with the use of a stopwatch (see the feature below: Job DescriptionCity of Phoenix, Arizona). The general approach to time study can be described as follows.
1. Define and evaluate each task and activity. This includes determining what level of detail is best for time study

Time study is the development of a standard time by observing a task and analyzing it with the use of a stopwatch.

n $ (za/2)22/E2

[A.3]

Job DescriptionCity of Phoenix, Arizona

The city of Phoenix, Arizona posted this job description for an operations analyst. The job is to design, conduct, and participate in major work standards and systems analyses covering a wide variety of government functions. Considerable flexibility is allowed in this job for designing and conducting each study. Assignments are comprehensive and entail interactions between major government and civic organizational units. The duties also involve substantial contact with high-level government officials, so writing and presentation skills are essential.

Operations Analyst for City of Phoenix, Arizona

Essential Job Requirements:

en

Designs control reporting systems for use in unit measurement for evaluation of performance and for determination of staffing levels and recommends staffing levels to section chief;

Studies operational problems such as office space utilization, equipment utilization, management reporting systems, staffing patterns, process efficiency, and prepares written recommendations for changes and/or improvements; Develops project plans to achieve established objectives and time schedules; Writes and/or edits manuals for uniform use of new or revised procedures and policies;

A6
62564_09_cA.indd 6

Part 4: OM2 Supplementary Chapters

ga

Designs systems, procedures, forms, and work measurements to effect methods improvement, work simplification, improvement of manual processing, or for adaption to computer processing;

e r o ge f ty Le ar ni ng
Evaluates office machines and office or heavy operations equipment relative to quality, price, and determination of best equipment; Organizes, authors, and presents oral and written research reports; Identifies work elements in detail and develops complex flow charts, work standards, and work method improvements; Demonstrates continuous effort to improve operations, decrease turnaround times, streamline work processes, and work cooperatively and jointly to provide quality seamless customer service.

where za/2 is the value of the standard normal distribution having an area of /2 in the upper tail, is an estimate of the standard deviation, and E is the

Pr

op

Required Knowledge, Skills and Abilities:

Principles of work measurement and activity analysis. Principles of statistical methods and techniques. Employ work measurement techniques, i.e., stopwatch, pre-determined data, and time ladders.

Understand and carry out oral and written instruction provided in the English language.

Conduct studies and research with minimal supervision. Complete assignments with independent thought and action within the scope of specific assignments. Work cooperatively with other City employees, outside regulatory agencies, and the public. Enter data or information into a terminal, PC, or other keyboard device using various software packages.

Communicate orally with customers, co-workers, and the public in face-to-face one-on-one settings, in group settings, or using a telephone. Produce written documents with clearly organized thoughts using proper English sentence construction, punctuation, and grammar.2

8/26/09 4:04 PM

desired sampling error. When timing a work activity with multiple tasks, the general rule is to take the largest sample size estimate from equation A.3 for all tasks. 3. Rate the employees performance of each task or activity. As noted, rating human performance accurately requires considerable training. 4. Use the performance rating and equation A.1 to determine the normal task time. The sum of those task times is the normal time for the entire work activity. 5. Determine the allowance factor for the work activity. 6. Determine the standard time using equation A.2 .

To illustrate time studAn operations activity ies, we will consider a simple chart is a detailed analysis manual assembly process. of work motions performed Exhibits A.1 and A.2 show for a manual task. a faucet stem assembly and an operations activity chart, which provide the basis for developing the time study. An operations activity chart is a detailed analysis of work motions performed for a manual task. Because these micromotions are typically so small that it would be difficult to measure them easily, we usually combine several smaller work tasks into larger activities. For instance, the tasks get washer

Exhibit A.1

Exhibit A.2

Operation chart Present method

Get Housing

ga

Hold Housing

en C
Place assembly in tray

OM2 Supplementary Chapter A: Work Measurement, Learning Curves, and Standards

62564_09_cA.indd 7

e r o ge f ty Le ar ni ng
Faucet Stem Assembly
Housing Stem Washer Screw

Pr

op
Left hand

Operations Activity Chart of Faucet Stem Assembly

Faucet stem assembly

Right hand

Get stem

Screw in stem Get washer Insert washer Get screw Insert screw

Get screwdriver Tighten screw Release screwdriver

A7
8/26/09 4:04 PM

and insert washer might be combined. This leads to the following set of work activities:
1. Get housing and stem. 2. Screw in stem. 3. Get and insert washer. 4. Get and insert screw. 5. Tighten screw. 6. Place completed assembly in tray.

n 5 (1.645)2(.019)2/(.01)2 5 9.8 10 observations

Exhibit A.3

A8
62564_09_cA.indd 8

Part 4: OM2 Supplementary Chapters

en

ga

Pr

A Spreadsheet for Time Study (This spreadsheet is available on the Premium Website.)

op

A sample size of 10 or more will provide the required precision. (Fractional values of n should always be rounded upward to ensure that the precision is at least as good as desired.)

e r o ge f ty Le ar ni ng
8/26/09 4:04 PM

To determine the sample size needed for a time study, suppose we desire a 90 percent probability that the value of the sample mean provides a sampling error of .01 minute or less. Further assume that is estimated from historical experience to be .019. Therefore, 5 .10, z.05 5 1.645, and E 5.01. Using equation A.3, we compute

A spreadsheet that can be used for a typical time study when continuous timing is used is shown in Exhibit A.3. Continuous timing involves starting the clock at the beginning of each task and recording the cumulative time at the completion of each work task. The task times are found by subtracting successive cumulative times. These are added and averaged to obtain the mean time for each work task. Performance ratings are given in the next-to-last column. By multiplying the performance-rating factor by the average observed time, we obtain the normal time for each work task and add them. Next the allowances are determined to compute the standard time. For the faucet-stem assembly, we assume a 5 percent personal allowance, 5 percent fatigue allowance, and 10 percent delay of materials allowance. Therefore, the total allowance factor is 20 percent. The standard time for the faucet stem assembly job is then computed, using equation A.2, as: Standard time 5 (0.550)(1 1 .2) 5 0.660 minutes per faucet assembly. Thus, an assembler of faucet-stem assemblies can be expected to produce at a standard rate of 1/0.660 parts per minute, or about 91 parts per hour. In a

7-hour workday with 1 hour off for lunch and breaks, an assembler can produce (7)(91) 5 637 faucet-stem assemblies per workday.

standard time-estimating formula for installing power lines. A good formula would help it plan capacity and staffing needs. The following data are collected:
Total Time Number (hours) of Poles 8.0 1 14.0 2 17 .5 3 7 .0 1 16.0 2 37 .5 4 39.5 4 10.5 1 17 .0 2 23.5 3 16.5 2 22.0 3 8.5 1 28.5 4 Wire (100 feet) 4 10 6 2.5 10 24 33 3 8 12 12 18 5 12 No. of Cross No. of Arms Insulators 1 2 2 4 3 6 2 3 4 6 8 12 7 11 2 4 4 8 6 12 2 4 3 6 2 3 8 12 No. of Guy wires 1 0 1 0 0 2 1 2 1 0 1 0 0 0

2.1 Using Regression Analysis to Determine Standard Time

Regression analysis provides an alternative method to estimating the time required to do a particular job or work activity. Regression analysis is used to predict times based on different attributes of the work, rather than by adding up individual task times. Using regression to estimate standard times can be advantageous because it avoids the assumption of additive task times when this might not hold; statistically significant variables can be determined; confidence intervals for the prediction can be developed; and finally, it may cost less than a detailed work study. Consider the following problem on developing standard time estimates for installing electrical power lines. An electric power company wishes to determine a

Exhibit A.4

Results of Regression Analysis for Electric Power Line Installation

en

ga

OM2 Supplementary Chapter A: Work Measurement, Learning Curves, and Standards

62564_09_cA.indd 9

e r o ge f ty Le ar ni ng
Exhibit A.4 shows the results using Excels Regression tool. The model obtained from this analysis is Time 5 0  .237 1 2.804 Poles 1 0.514 Wire 1 1.09 Cross arms 1 0.170 Insulators 1 1.50 Guy wires

Pr

op

A9
8/26/09 4:04 PM

The regression analysis shows a high R2 value, showing a strong fit to the data. Moreover, the p values for the regression coefficients are significant, meaning that each of the variables contributes to predicting time. If the utility faces a situation in which it estimates installation of 4 poles, 1,500 feet of wire, 7 cross arms, 12 insulators, and no guy wires, the predicted time for the job would be Time 5  0.237 1 2.804 3 4 1 0.514 3 1500 1 1.09 3 7 1 0.170 3 12 1 1.50 3 0 5 792 hours

2.3 The Debate Over Work Standards

Work standards evolved at the turn of the twentieth century, and although they have supported significant gains in productivity, they have been the subjects of debate since the quality revolution began in the United States. Critics such as W. Edwards Deming have condemned work standards on the basis that they destroy intrinsic motivation in jobs and rob workers of the creativity necessary for continuous improvement. That is certainly true when managers dictate standards in an effort to meet numerical goals set up by their superiors. However, the real culprit in that case is not the standards themselves, but managerial style. The old style of managing reflects Taylors philosophy: Managers and engineers think, and workers do what they are told. A total quality approach suggests that empowered workers can manage their own processes with help from managers and professional staff. Experience at GMs NUMMI plant has shown that work standards can have very positive results when they are not imposed by dictum, but designed by the workers themselves in a continuous effort to improve productivity, quality, and skills.3 At the GM-Fremont plant, industrial engineers performed all of the methods analysis and work-measurement activities, designing jobs as they saw fit. When the industrial engineers were performing motion studies, workers would naturally slow down and make the work look harder. At NUMMI, team members learned techniques of work analysis and improvement, then timed one another with stop-watches, looking for the safest, most efficient way to do each task at a sustainable pace. They picked the best performance, broke it down to its fundamental elements, and then explored ways to improve the task. The team compared the analyses with those from other shifts at the same workstation, and wrote detailed specifications that became the work standards. Results were excellent. From a total quality perspective, this was simply an approach to reduce variability. In addition, safety and quality improved, job rotation became more effective, and flexibility increased.

Predetermined time standards describe the amount of time necessary to accomplish specific movements (called micromotions), such as moving a human hand a certain distance or lifting a 1-pound part. These small time estimates have been documented and are available in books and electronic tables. If a job, work activity, or task can be broken down into such elemental tasks, an estimate of the normal time is made by adding up these predetermined times. This approach is especially appealing for developing standard times for new manufactured goods and some service tasks. An electronics manufacturer, for example, may have much experience with assembling small electronic components using human labor and keep a record of past micromotion and normal time analyses. For similar new tasks and electronic component parts, predetermined time standards can be used to estimate new normal times. Predetermined time standards were originally developed for labor-intensive human tasks but data sets exist for machine micromovements, such as those involving an automated drill press. Predetermined time standards are advantageous since they cost less than a stopwatch time study, avoid needing multiple performance ratings, and are best for new goods and services. However, this type of micromotion-based time standard is not justified for small order sizes or infrequent production runs. In addition, once the new good or service and its associated process are stable and running well, stop-watch or worksampling time studies still need to be done. Finally, the assumption of additivity is sometimes questionable, since a process or assembly sequence of difficult versus simple micromotions may or may not be additive.

A10
62564_09_cA.indd 10

Part 4: OM2 Supplementary Chapters

en

ga

e r o ge f ty Le ar ni ng
8/26/09 4:04 PM

2.2 Predetermined Time Standard Methods

Pr

op

3 Work Sampling
is a method of randomly observing work over a period of time to obtain a distribution of the activities that an individual or a group of employees perform. Work sampling determines the proportion of time spent doing certain activities on a job. It can be used to determine the percentage of idle time and also as a means of assessing nonproductive time to determine performance ratings or to establish allowances. Work sampling is based on the binomial probability distribution, because it is concerned with the proportion of time that a certain activity occurs. Thus, the sample size (n) for a work-sampling study is found by using equation A.4:
ork sampling

where p is an estimate of the population proportion of the binomial distribution. Obviously, p will never be known exactly, since it is the population parameter we are trying to estimate. We can choose a value for p from past data, a preliminary sample, or a subjective estimate. If p is difficult to determine in those ways, we can select p 5 0.5, since it gives us the largest value for p(1 2 p) and therefore provides the largest and most conservative sample size. To illustrate work sampling, consider the secretarial staff in a college department office. The secretaries spend their time in various ways, such as

op

n $ (z/2)2p(1 2 p)/E2

answering the telephone revising technical papers talking to students duplicating class handouts other productive activities personal time idle periods

ga

typing drafts of technical papers

en

OM2 Supplementary Chapter A: Work Measurement, Learning Curves, and Standards

62564_09_cA.indd 11

e r o ge f ty Le ar ni ng
[A.4] n$ (1.96)20.4(1 2 0.4)/0.052 5 368.8 or at least 369 observations If that is to be done over a 1-week (40-hour) period, it represents approximately 9 observations per hour (9.225, to be exact). The observations should be taken randomly when work is at a normal level (not during the spring break!).
Exhibit A.5

Suppose that a new voiceWork sampling is a methactivated word processod of randomly observing ing software could greatly work over a period of time to increase productivity obtain a distribution of the activities that an individual in typing, spell checkor a group of employees ing, and revising techniperform. cal papers. However, the purchase of this product is not justified unless it is used a significant percentage of the time. To determine the percentage of time secretaries spend performing the relevant work activities, we could observe them at random times and record their activities. If 100 observations are taken, we might get the results in Exhibit A.5. The percentage of time spent typing or revising is 21 percent 1 7 percent 5 28 percent. To determine the needed sample size (that is, the number of observations), suppose we want to estimate the proportion of time spent typing to 65 percent, with a 95 percent probability. We use equation A.4, with E 5 0.05 and z/2 5 1.96. Suppose the head secretary estimates that 40 percent of the time is spent typing. This provides a value for p of 0.4. Then the needed sample size using equation A.4 is

Pr

Work Sampling Activity Frequency Data

14 21 7 10 15 25 6 2 100

Activity Answering the telephone Typing drafts Revising papers Talking to students Duplicating Other productive activity Personal Idle TOTAL

Frequency

A11
8/26/09 4:04 PM

time of the xth unit.

Solved Problem

Solution:

Effective number of hours worked 5 2,700(240)/3,000 5 216.

A12
62564_09_cA.indd 12

Part 4: OM2 Supplementary Chapters

In a work-sampling study an administrative assistant was found to be working 2,700 times in a total of 3,000 observations made over a time span of 240 working hours. The employees output was 1,800 forms. If a performance rating of 1.05 and an allowance of 15 percent are given, what is the standard output for this task?

en

ga

e r o ge f ty Le ar ni ng
Output during this period 5 1,800 forms. Actual time per form 5 216(60)/1,800 5 7.2 minutes, or 8.33 forms per hour. Normal time 5 Actual observed time 3 Performance rating 5 7.2(1.05) 5 7.56 minutes. Standard time 5 7.56(1.15) 5 8.694 minutes per form. Standard output 5 60/8.694 5 6.9 forms per hour, or 55 forms per 8-hour day.
8/26/09 4:04 PM

To take a random sample, we can use the table of random digits in Appendix C. There are several ways to use random digits in deciding when to take observations. For this example, an average of 9.225 observations per hour requires the observations to be spaced, on the average (60/9.225) 5 6.5 minutes apart. We should not take observations exactly 6.5 minutes apart, however, for then the sample would not be random. Suppose observations are between 3 and 10 minutes apart. If they are random, the average is 6.5 minutes. We can use the random digits as follows. Suppose the first observation is taken at 9:00. We choose numbers from the first row of Appendix C to find how many minutes later we should take the next observation (0 represents 10 minutes, and we discard any 1s or 2s). For instance, the first number is 6; thus we take the next observation at 9:06. The next number is 3, so the third observation is made at 9:09. We discard the 2 and take the next observation 7 minutes later, at 9:16. We see that the time and cost required to take a random sample can be significant; that is one of the disadvantages of random sampling. Work sampling is based on statistics, and like all statistical procedures, it can suffer from sampling The learning curve error and lead to erroconcept is that direct neous conclusions simlabor unit cost decreases in a predictable manner as the ply by chance. Also, as experience in producing the the famous Hawthorne unit increases. experiments showed, A p -percent learning people often change curve characterizes a protheir behavior when cess in which the time of the 2xth unit is p percent of the being observed, and this

can influence the results. Thus, work sampling should be used with caution.

he learning curve concept is that direct labor unit cost decreases in a predictable manner as the experience in producing the unit increases. For most people, for example, the longer they play a musical instrument or a video game, the better and faster they become. The same is true in assembly operations, which was recognized in the 1920s at Wright-Patterson Air Force Base in the assembly of aircraft. Studies showed that the number of labor hours required to produce the fourth plane was about 80 percent of the amount of time spent on the second; the eighth plane took only 80 percent as much time as the fourth; the sixteenth plane 80 percent of the time of the eighth, and so on. The decrease in production time as the number produced increases is illustrated in Exhibit A.6. As production doubles from x units to 2x units, the time per unit of the 2xth unit is 80 percent of the time of the xth unit. This is called an 80 percent learning curve. Such a curve exhibits a steep initial decline and then levels off as employees become more proficient in their tasks. In general, a p-percent learning curve characterizes a process in which the time of the 2xth unit is p percent of the time of the xth unit. Defense industries (for example, the aircraft and electronics industries), which introduce many new and complex products, use learning curves to estimate labor requirements and capacity, determine costs and

4 Learning Curves

Pr

op

Exhibit A.6

An 80 Percent Learning Curve

100

Time per unit as a percentage of first unit

80 60 40 20

budget requirements, and plan and schedule production. Eighty-percent learning curves are generally accepted as a standard, although the ratio of machine work to manual assembly affects the curve percentage. Obviously, no learning takes place if all assembly is done by machine. As a rule of thumb, if the ratio of manual to machine work is 3 to 1 (three-fourths manual), then 80 percent is a good value; if the ratio is 1 to 3, then 90 percent is often used. An even split of manual and machine work would suggest the use of an 85 percent learning curve. The learning factor may also be estimated from past histories of similar parts or products. Mathematically, the learning curve is represented by the function

y 5 ax2b

Thus, for an 80 percent learning curve, p 5 0.8 and

For a 90 percent curve, p 5 0.9 and b 5 2ln 0.9/ln 2 5 2(2 0.105)/0.693 5 0.152

Although the learning curve theory implies that improvement will continue forever, in actual practice

b 5 2ln 0.8/ln 2 5 2(2 0.223)/0.693 5 0.322

en

where x 5 number of units produced, a 5 hours required to produce the first unit, y 5 time to produce the xth unit, and b5 constant equal to 2ln p/ln 2 for a 100p percent learning curve.

ga

OM2 Supplementary Chapter A: Work Measurement, Learning Curves, and Standards

62564_09_cA.indd 13

e r o ge f ty Le ar ni ng
Number produced

10

15

[A.5]

the learning curve flattens out. As management interest in the initial creation of a new good or service decreases, employees may reach a level of production that is expected of them and hold that rate. Another way to view the theory of learning is that early on extraordinary new practices and methods are found to dramatically improve performance, such as substituting plastic for steel parts. Later in the life of the learning curve, the focus shifts to incremental improvements. Learning curves can apply to individual employees or, in an aggregate sense, to the big-picture initiatives such as pricing strategy. For example, learning curves are used to monitor employees typing and encoding checks in a banks operations. Each employee must reach a certain threshold-learning rate within 6 months or more training is required. In some cases, the bankencoding employee is transferred to another bank job because the employee is just not suited to the encoding job. Learning curves help managers make such decisions. From an aggregate and strategic perspective, a firm may use the learning-curve concept to establish a pricing schedule that does not initially cover cost in order to gain increased market share. Managers should realize that improvement along a learning curve does not take place automatically. Learning-curve theory is most applicable to new products or processes that have a high potential for improvement and when the benefits will be realized only when appropriate incentives and effective motivational tools are used. Organizational changes may also have

Pr

op

A13
8/26/09 4:04 PM

The experience curve states that the cost of doing any repetitive task, work activity, or project decreases as the accumulated experience of doing the job increases.

significant effects on learning. Changes in technology or work methods will affect the learning curve, as will the institution of productivity and qualityimprovement programs. As an illustration of learning curves, suppose a manufacturing firm is introducing a new and complex machine and has determined that a 90 percent learning curve is applicable. Estimates of demand for the next 3 years are 50, 75, and 100 units. The time to produce the first unit is estimated to be 3,500 hours. Therefore, the learning-curve function is y 5 3,500x20.152

Consequently, the time to manufacture the second unit will be 3,500 (2)20.152 5 3,150 hours

112,497/1,920 5 59 employees

to produce this machine. In the second year, the total number of hours required will be the difference between the cumulative requirements for the first two years production (246,160 hours) and the first years production (112,497 hours), or 246,160 2 112,497 5 133,663 hours

Pr

Exhibit A.7 gives the cumulative number of hours required to produce the 3-year demand in increments of 25 units. Thus, to produce the 50 units in the first year, the firm will require 112,497 hours. If we assume that each employee works 160 hours per month, or 1,920 hours per year, we find that for the first year the firm will need

So the labor requirements for the second year are 133,663/1,920 5 70 employees

ga

Similarly, for the third year, the labor requirements are

en

406,112 2 246,160 __________________ 5 83 employees 1,920

These are aggregate numbers; at a more detailed planning level, they will vary according to how production is actually scheduled over the year. Also, note that the number of employees required to produce these units increases in a nonlinear way, reflecting the learning that is taking place among employees. For example, we might need 59 employees to produce 50 units, 70

A14
62564_09_cA.indd 14

Part 4: OM2 Supplementary Chapters

e r o ge f ty Le ar ni ng
Exhibit A.7

employees for 125, and 83 for 225. Clearly, understanding learning is important for aggregate planning. Values for learning-curve functions can be easily computed and summarized through the use of tables. Exhibits A.8 and A.9 present unit values and cumulative values, respectively, for learning curves from 60 percent through 95 percent. To find the time to produce a specific unit, multiply the time for the first unit by the appropriate factor in Exhibit A.8. For the 90 percent learning-curve example presented earlier, the time for the second unit is 3,500(0.9000) 5 3,150. The time for the third unit is 3,500(0.8462) 5 2,961.7, and so on. To find the time for a cumulative number of units, we can use Exhibit A.9. Thus, for a 90 percent learning curve, if the time for the first unit is 3,500, the time for the first 25 units is 3,500(17.7132) 5 61,996. Similarly, the time for the first 100 units is 3,500(58.1410) 5 203,494. The values in Exhibit A.7 were found using this table. A broader extension of the learning curve is the experience curve. The experience curve states that the cost of doing any repetitive task, work activity, or project decreases as the accumulated experience of doing the job increases. The terms improvement curve, experience curve, and manufacturing progress function are often used to describe the learning phenomenon in the aggregate context. Marketing research, software design, developing engineering specifications for a water plant, accounting and financial auditing of the same client, implementing a software integration project, and so on are examples of this broader view. The idea is that each time experience doubles, costs decline by 10 percent to 30 percent. Costs must always be translated into

op

Cumulative Time RequiredUsing Learning Curves

25 50 75 100 125 150 175 200 225

Cumulative Units

Cumulative Hours Required 61,996 112,497 159,164 203,494 246,160 287 ,545 327 ,894 367 ,374 406,112

8/26/09 4:04 PM

Solved Problem
A yacht manufacturer has been commissioned to build five sailboats for a Florida resort. The first boat took 6,000 labor-hours to build. How many labor-hours will it take to complete the order, assuming that a 90 percent learning curve is applicable?
Unit 1 2 3 4 5 Total Time Required 6,000 5,400 5,077 4,860 4,698 26,035 hours

Solution:
Using a 90 percent learning curve, we have the times listed here:

Or, using Exhibit A.9, 6,000(4.3392) 5 26,035 hours.

constant dollars to eliminate the inflation effect. Of course, the learning or experience curve does not continue this dramatic decrease in time and costs indefinitely, and at some point begins to flatten out.

The following ten factors can affect the applicability of the learning or experience curve and/or the amount of learning that occurs. Good management judgment is required to recognize these factors and take appropriate action, including stopping the current learningcurve analysis, beginning a new learning-curve analysis, and/or using other planning methods for the remainder of the work.

op

4.1 Practical Issues in Using Learning Curves

1. The learning curve does not usually apply to supervisory personnel, some skilled craftspeople, or jobs that have nonrepetitive job tasks.

3. The institution of incentive systems, bonus plans, quality initiatives, empowerment, and the like may increase learning.

ga

2. A change in the ratio of indirect labor or supervisory talent to direct labor can alter the rate of learning.

en

OM2 Supplementary Chapter A: Work Measurement, Learning Curves, and Standards

62564_09_cA.indd 15

e r o ge f ty Le ar ni ng
4. Changes in product design, raw material usage, technology, and/or the process may significantly alter the learning curve. 5. Humans learn simple task(s) quickly and reach a limit on learning for the task(s), but for complex intellectual task(s) such as software programming, learning is less limited and may continue. The first type of learning is described with an exponential curve; the more complex learning is sometimes described by an S-shaped curve. 6. A contract phaseout may result in a lengthening of processing times for the last units produced, since employees want to prolong their income period. 7. The lack of proper maintenance of tools and equipment, the nonreplacement of tools, or the aging of equipment can have a negative impact on learning. 8. Keeping groups of employees together, such as highly specialized consulting groups, reaps a productivity benefit but may stifle innovation and new experiences. 9. The transfer of employees may result in an interruption or a regression to an earlier stage of the learning curve or may necessitate a new learning curve. 10. Learning curves focus on direct labor and ignore indirect labor that also contributes to efficiency and effectiveness.

Pr

A15
8/26/09 4:04 PM

Exhibit A.8
p x b
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 125 150 175 200 225 250 275 300 350 400 450 500 550 600 650 700 750 800 850 900

Unit Values for Learning Curves

.60 .737
1.0000 0.6000 0.4450 0.3600 0.3054 0.2670 0.2383 0.2160 0.1980 0.1832 0.1708 0.1602 0.1510 0.1430 0.1359 0.1296 0.1239 0.1188 0.1142 0.1099 0.1061 0.1025 0.0992 0.0961 0.0933 0.0815 0.0728 0.0660 0.0605 0.0560 0.0522 0.0489 0.0461 0.0437 0.0415 0.0396 0.0379 0.0363 0.0349 0.0336 0.0285 0.0249 0.0222 0.0201 0.0185 0.0171 0.0159 0.0149 0.0133 0.0121 0.0111 0.0103 0.0096 0.0090 0.0085 0.0080 0.0076 0.0073 0.0069 0.0067

.65 .621
1.0000 0.6500 0.5052 0.4225 0.3678 0.3284 0.2984 0.2746 0.2552 0.2391 0.2253 0.2135 0.2031 0.1940 0.1858 0.1785 0.1719 0.1659 0.1604 0.1554 0.1507 0.1465 0.1425 0.1387 0.1353 0.1208 0.1097 0.1010 0.0939 0.0879 0.0829 0.0785 0.0747 0.0713 0.0683 0.0657 0.0632 0.0610 0.0590 0.0572 0.0498 0.0444 0.0404 0.0371 0.0345 0.0323 0.0305 0.0289 0.0262 0.0241 0.0224 0.0210 0.0198 0.0188 0.0179 0.0171 0.0163 0.0157 0.0151 0.0146

.70 .515
1.0000 0.7000 0.5682 0.4900 0.4368 0.3977 0.3674 0.3430 0.3228 0.3058 0.2912 0.2784 0.2672 0.2572 0.2482 0.2401 0.2327 0.2260 0.2198 0.2141 0.2087 0.2038 0.1992 0.1949 0.1908 0.1737 0.1605 0.1498 0.1410 0.1336 0.1272 0.1216 0.1167 0.1123 0.1084 0.1049 0.1017 0.0987 0.0960 0.0935 0.0834 0.0759 0.0701 0.0655 0.0616 0.0584 0.0556 0.0531 0.0491 0.0458 0.0431 0.0408 0.0389 0.0372 0.0357 0.0344 0.0332 0.0321 0.0311 0.0302

.75 .415
1.0000 0.7500 0.6338 0.5625 0.5127 0.4754 0.4459 0.4219 0.4017 0.3846 0.3696 0.3565 0.3449 0.3344 0.3250 0.3164 0.3085 0.3013 0.2946 0.2884 0.2826 0.2772 0.2722 0.2674 0.2629 0.2437 0.2286 0.2163 0.2060 0.1972 0.1895 0.1828 0.1768 0.1715 0.1666 0.1622 0.1582 0.1545 0.1511 0.1479 0.1348 0.1250 0.1172 0.1109 0.1056 0.1011 0.0972 0.0937 0.0879 0.0832 0.0792 0.0758 0.0729 0.0703 0.0680 0.0659 0.0641 0.0624 0.0508 0.0594

.80 .322
1.0000 0.8000 0.7021 0.6400 0.5956 0.5617 0.5345 0.5120 0.4929 0.4765 0.4621 0.4493 0.4379 0.4276 0.4182 0.4096 0.4017 0.3944 0.3876 0.3812 0.3753 0.3697 0.3644 0.3995 0.3548 0.3346 0.3184 0.3050 0.2936 0.2838 0.2753 0.2676 0.2608 0.2547 0.2491 0.2440 0.2393 0.2349 0.2308 0.2271 0.2113 0.1993 0.1896 0.1816 0.1749 0.1691 0.1639 0.1594 0.1517 0.1453 0.1399 0.1352 0.1312 0.1275 0.1243 0.1214 0.1187 0.1163 0.1140 0.1119

.85 .234
1.0000 0.8500 0.7729 0.7225 0.6857 0.6570 0.6337 0.6141 0.5974 0.5828 0.5699 0.5584 0.5480 0.5386 0.5300 0.5220 0.5146 0.5078 0.5014 0.4954 0.4898 0.4844 0.4794 0.4747 0.4701 0.4505 0.4345 0.4211 0.4096 0.3996 0.3908 0.3829 0.3758 0.3693 0.3634 0.3579 0.3529 0.3482 0.3438 0.3397 0.3224 0.3089 0.2979 0.2887 0.2809 0.2740 0.2680 0.2625 0.2532 0.2454 0.2387 0.2329 0.2278 0.2232 0.2190 0.2152 0.2118 0.2086 0.2057 0.2029

.90 .152
1.0000 0.9000 0.8462 0.8100 0.7830 0.7616 0.7439 0.7290 0.7161 0.7047 0.6946 0.6854 0.6771 0.6696 0.6626 0.6561 0.6501 0.6445 0.6392 0.6342 0.6295 0.6251 0.6209 0.6169 0.6131 0.5963 0.5825 0.5708 0.5607 0.5518 0.5438 0.5367 0.5302 0.5243 0.5188 0.5137 0.5090 0.5046 0.5005 0.4966 0.4800 0.4669 0.4561 0.4469 0.4390 0.4320 0.4258 0.4202 0.4105 0.4022 0.3951 0.3888 0.3832 0.3782 0.3736 0.3694 0.3656 0.3620 0.3587 0.3556

.95 .074
1.0000 0.9500 0.9219 0.9025 0.8877 0.8758 0.8659 0.8574 0.8499 0.8433 0.8374 0.8320 0.8271 0.8226 0.8184 0.8145 0.8109 0.8074 0.8042 0.8012 0.7983 0.7955 0.7929 0.7904 0.7880 0.7775 0.7687 0.7611 0.7545 0.7486 0.7434 0.7386 0.7342 0.7302 0.7265 0.7231 0.7198 0.7168 0.7139 0.7112 0.6996 0.6902 0.6824 0.6757 0.6698 0.6646 0.6599 0.6557 0.6482 0.6419 0.6363 0.6314 0.6269 0.6229 0.6192 0.6158 0.6127 0.6098 0.6070 0.6045

A16
62564_09_cA.indd 16

Part 4: OM2 Supplementary Chapters

en

ga

e r o ge f ty Le ar ni ng
8/26/09 4:04 PM

Pr

op

Exhibit A.9
x 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 125 150 175 200 225 250 275 300 350 400 450 500 550 600 650 700 750 800 850 900

Cumulative Unit Values for Learning Curves

p 5.65 1.0000 1.6500 2.1552 2.5777 2.9455 3.2739 3.5723 3.8469 4.1021 4.3412 4.5665 4.7800 4.9831 5.1770 5.3628 5.5413 5.7132 5.8791 6.0396 6.1950 6.3457 6.4922 6.6346 6.7734 6.9086 7 .5398 8.1095 8.6312 9.1143 9.5654 9.9896 10.3906 10.7715 11.1347 11.4823 11.8158 12.1367 12.4461 12.7451 13.0345 14.3614 15.5326 16.5883 17 .5541 18.4477 19.2816 20.0653 20.8059 22.1796 23.4362 24.5987 25.6835 26.7028 27 .6662 28.5808 29.4528 30.2869 31.0871 31.8568 32.5989 p 5.70 1.0000 1.7000 2.2682 2.7582 3.1950 3.5928 3.9601 4.3031 4.6260 4.9318 5.2229 5.5013 5.7685 6.0257 6.2739 6.5140 6.7467 6.9727 7 .1925 7 .4065 7 .6153 7 .8191 8.0183 8.2132 8.4040 9.3050 10.1328 10.9024 11.6245 12.3069 12.9553 13.5742 14.1674 14.7376 15.2874 15.8188 16.3335 16.8329 17 .3182 17 .7907 19.9894 21.9722 23.7917 25.4820 27 .0669 28.5638 29.9855 31.3423 33.8916 36.2596 38.4799 40.5766 42.5680 44.4684 46.2889 48.0387 49.7254 51.3552 52.9333 54.4644 p 5.75 1.0000 1.7500 2.3838 2.9463 3.4591 3.9345 4.3804 4.8022 5.2040 5.5886 5.9582 6.3147 6.6596 6.9940 7 .3190 7 .6355 7 .9440 8.2453 8.5399 8.8284 9.1110 9.3882 9.6601 9.9278 10.1907 11.4458 12.6179 13.7232 14.7731 15.7761 16.7386 17 .6658 18.5617 19.4296 20.2722 21.0921 21.8910 22.6708 23.4329 24.1786 27 .6971 30.9342 33.9545 36.8007 39.5029 42.0833 44.5588 46.9427 51.4760 55.7477 59.8030 63.6753 67 .3900 70.9671 74.4225 77 .7693 81.0183 84.1786 87 .2580 90.2631 p 5.80 1.0000 1.8000 2.5021 3.1421 3.7377 4.2994 4.8339 5.3459 5.8389 6.3154 6.7775 7 .2268 7 .6647 8.0923 8.5105 8.9201 9.3218 9.7162 10.1037 10.4849 10.8602 11.2299 11.5943 11.9538 12.3086 14.0199 15.6428 17 .1935 18.6835 20.1217 21.5147 22.8678 24.1853 25.4708 26.7273 27 .9572 29.1628 30.3459 31.5081 32.6508 38.1131 43.2335 48.0859 52.7000 57 .1712 61.4659 65.6246 69.6634 77 .4311 84.8487 91.9733 98.8473 105.5032 111.9671 118.2598 124.3985 130.3976 136.2693 142.0242 147 .6709 p 5.85 1.0000 1.8500 2.6229 3.3454 4.0311 4.6881 5.3217 5.9358 6.5332 7 .1161 7 .6860 8.2444 8.7925 9.3311 9.8611 10.3831 10.8977 11.4055 11.9069 12.4023 12.8920 13.3765 13.8559 14.3306 14.8007 17 .0907 19.2938 21.4252 23.4955 25.5131 27 .4843 29.4143 31.3071 33.1664 34.9949 36.7953 38.5696 40.3198 42.0474 43.7539 52.0109 59.8883 67 .4633 74.7885 81.9021 88.8328 95.6028 102.2301 115.1123 127 .5691 139.6656 151.4506 162.9622 174.2309 185.2815 196.1346 206.8073 217 .3144 227 .6687 237 .8811 p 5.90 1.0000 1.9000 2.7462 3.5562 4.3392 5.1008 5.8447 6.5737 7 .2898 7 .9945 8.6890 9.3745 10.0516 10.7212 11.3837 12.0398 12.6899 13.3344 13.9735 14.6078 15.2373 15.8624 16.4833 17 .1002 17 .7132 20.7269 23.6660 26.5427 29.3658 32.1420 34.8766 37 .5740 40.2377 42.8706 45.4753 48.0539 50.6082 53.1399 55.6504 58.1410 70.3315 82.1558 93.6839 104.9641 116.0319 126.9144 137 .6327 148.2040 168.9596 189.2677 209.1935 228.7851 248.0809 267 .1118 285.9030 304.4757 322.8479 341.0347 359.0497 376.9043 p 5.95 1.0000 1.9500 2.8719 3.7744 4.6621 5.5380 6.4039 7 .2612 8.1112 8.9545 9.7919 10.6239 11.4511 12.2736 13.0921 13.9066 14.7174 15.5249 16.3291 17 .1302 17 .9285 18.7241 19.5170 20.3074 21.0955 25.0032 28.8636 32.6838 36.4692 40.2239 43.9511 47 .6535 51.3333 54.9924 58.6323 62.2544 65.8599 69.4498 73.0250 76.5864 94.2095 111.5730 128.7232 145.6931 162.5066 179.1824 195.7354 212.1774 244.7667 277 .0124 308.9609 340.6475 372.1002 403.3421 434.3918 465.2653 495.9759 526.5353 556.9538 587 .2402

p 5 .60 1.0000 1.6000 2.0450 2.4050 2.7104 2.9774 3.2158 3.4318 3.6298 3.8131 3.9839 4.1441 4.2951 4.4381 4.5740 4.7036 4.8276 4.9464 5.0606 5.1705 5.2766 5.3791 5.4783 5.5744 5.6677 6.0974 6.4779 6.8208 7 .1337 7 .4222 7 .6904 7 .9413 8.1774 8.4006 8.6123 8.8140 9.0067 9.1912 9.3683 9.5388 10.3079 10.9712 11.5576 12.0853 12.5665 13.0098 13.4216 13.8068 14.5112 15.1451 15.7230 16.2555 16.7500 17 .2125 17 .6474 18.0583 18.4482 18.8193 19.1737 19.5131

en

ga

OM2 Supplementary Chapter A: Work Measurement, Learning Curves, and Standards

62564_09_cA.indd 17

e r o ge f ty Le ar ni ng
A17
8/26/09 4:04 PM

Pr

op

Problems, Activities, and Discussions

1. Do you think the following jobs require standard times? Explain your reasoning. a. Carpet installers b. Software programmers c. Cable T.V. installers d. Hotel maids e. Bank tellers f. Airline flight attendants g. Dentists h. Medical doctors i. Restaurant reservations j. Telephone call center representatives 2. What sample sizes should be used for these time studies? 3. Compute the number of observations required in a work-sampling study if the standard deviation is 0.2 minute and there should be a 90 percent chance that the sample mean has an error of (a) 0.15 minute, (b) 0.10 minute, and (c) 0.005 minute. 4. Exhibit A.10 shows a partially completed time-study worksheet. Determine the standard time for this operation. 5. Using a fatigue allowance of 20 percent, and given the following time-study data obtained by continuous time measurement, compute the standard time.
Cycle of Observation Performance Activity 1 2 3 4 5 Rating 0.21 0.48 1.52 1.73 1.98 2.10 2.31 2.59 3.65 3.83 4.09 4.20 4.41 4.66 5.66 5.91 6.15 6.25 6.45 8.59 6.70 8.86 7 .74 9.90 7 .96 10.10 8.21 10.30 8.34 10.42 0.95 0.90 1.00 0.95 0.80 1.10

a. There should be a .95 probability that the value of the sample mean is within 2 minutes, given that the standard deviation is 4 minutes.

Exhibit A.10

A18
62564_09_cA.indd 18

Part 4: OM2 Supplementary Chapters

en

ga

Pr

Time Study Worksheet for Problem 4

op

b. There should be a 90 percent chance that the sample mean has an error of 0.10 minutes or less when the variance is estimated as 0.50 minutes.

e r o ge f ty Le ar ni ng
Get casting Fix into fixture Drilling operation Unload Inspect Replace
8/26/09 4:04 PM

 6. Provide the data missing from the following information. Time is in minutes.
Actual Time 10.6 7 .8 6.5 2 Normal Time _____ 7 .2 _____ _____ Standard Time _____ _____ 7 .98 _____ Performance Rating 1.06 _____ 1.05 1.10 Fatigue Allowance 20% 15% _____ 15%

formed teams of employees who are re- sponsible for the entire job. She has contracted with a developer to build 20 homes of similar type and size. She has four teams of workers. The first homes were built in an average of 145 days. How long will it take to complete the contract if an 85 percent learning curve applies? 10. A manufacturer has committed to supply 16 units of a particular product in 4 months (that is, 16 weeks) at a price of $30,000 each. The first unit took 1,000 hours to produce. Even though the second unit took only 750 hours to produce, the manufacturer is anxious to know: a. if the delivery commitment of 16 weeks will be met, b. whether enough labor is available (currently 500 hours are available per week),

7. A part-time employee who rolls out dough balls at a pizza restaurant was observed over a 40-hour period for a work-sampling study. During that time, she prepared 550 pieces of pizza dough. The analyst made 50 observations and found the employee not working four times. The overall performance rating was 1.10. The allowance for the job is 15 percent. Based on these data, what is the standard time in minutes for preparing pizza dough? 8. How many observations should be made in a work-sampling study to obtain an estimate within 10percent of the proportion of time spent changing tools by a production worker with a 99 percent probability?

Pr C

The State Versus John Bracket Case Study

John Bracket has filed a lawsuit against us, George, stated Paul Cumin, the vice president of operations for the State Rehabilitation Services Commission (SRSC). George, you are Brackets manager. So what happened? He claims you raised his daily productivity quota for processing invoices from 200 to 300. Paul, I did raise his quota to more closely match the other employees. Bracket is always late for work, plays games on the computer, violates our dress code, and is generally disliked by his peer employees, responded George Davis, Brackets immediate supervisor. Is there any logic or numerical basis for your increasing his quota? Paul asked him. As he left the room, George responded, Paul, Ill get my work study data out, review it, and get back to you this afternoon. The data George Davis had for justifying raising Brackets quota are shown in Exhibit A.11. Samples #1 to #4 represent four different commission employees
Don Bayley/istockphoto.com

op

9. Linda Bryant recently started a small home-construction company. In an effort to foster high quality, rather than subcontracting individual work, she has

en

ga

OM2 Supplementary Chapter A: Work Measurement, Learning Curves, and Standards

62564_09_cA.indd 19

e r o ge f ty Le ar ni ng
c. whether or not the venture is profitable. Apply learning-curve theory to each of those issues. Assume the material cost per unit equals $22,000; labor equals $10 per labor-hour, and overhead is $2,000 per week.

doing the same job as Bracketprocessing commission invoices. These data result in an average normal time per invoice of 1.0915 minutes [(1.265 1 1.221 1 1.003 1 0.877)/4]. With an allowance factor of 20 percent, the standard time is 1.3098 (1.0915 3 1.2) minutes per invoice, or 45.81 per hour. During a typical 7-hour workday, an average employee could process 320.66 invoices per day, assuming a 1-hour lunch break. A similar study was performed for Brackets invoice-processing productivity; these results are shown in Exhibit A.12. These data result in an average normal time per invoice of 1.722 minutes [(1.808 1 1.452 1 2.032 1 1.595)/4]. With an allowance factor of 20 percent, the standard time is 2.066 (1.722 3 1.2) minutes per invoice, or 29.04 per hour. During a typical 7-hour workday, Bracket could process 203.29 invoices per day, assuming a 1-hour lunch break.4

A19
8/26/09 4:04 PM

Case Questions for Discussion

1. Whose case is justifiedDavis or Bracket? Explain. 2. What other issues should be considered?

3. Would you present these data in court? Why or why not? 4. What are your final recommendations?

Exhibit A.11

Work Measurement Time Study Original Data for Rehabilitation Services Commission

Process: Invoices Cumulative Time (ct) Select Time (t) All times in (minutes. hundredths of seconds)

Total Time Number Processed

op

Sorting/Matching Keying End-of-Day Activities

t ct t ct t ct

Process: Invoices Cumulative Time (ct) Select Time (t) All times in (minutes. hundredths of seconds) Major Work Elements Sorting/Matching Keying End-of-Day Activities t ct t ct t ct Sample #3 Rating Factor Normal Time

Pr

Normal Time/Invoice

Total Time Number Processed

ga

Normal Time/Invoice

en

A20
62564_09_cA.indd 20

Part 4: OM2 Supplementary Chapters

e r o ge f ty Le ar ni ng
13.09 1.00 13.09 13.09 28.45 1.00 28.45 41.54 24.23 1.00 24.23 65.77 65.77 t ct t ct t ct 22.60 1.10 22.60 23.40 1.10 46.00 13.94 1.10 59.94 59.94 54.00 24.86 25.74 15.33 65.93 65.93 65.77 52.00 65.77 52.00 54.00 1.265 1.221 Sample #4 Rating Factor Normal Time 8.92 6.66 8.44 1.10 9.28 8.44 14.26 1.10 15.69 22.70 .12 6.47 1.10 7 29.17 32.09 t ct t ct t ct .43 7 1.20 .43 7 5.55 1.20 12.98 15.52 1.20 28.50 28.50 39.00 18.62 34.20 34.20 39.00 0.877 29.17 32.00 32.09 32.00 1.003
8/26/09 4:04 PM

Major Work Elements

Sample #1

Rating Factor

Normal Time

Sample #2

Rating Factor

Normal Time

Exhibit A.12

Work Measurement Time Study Original Data for John Bracket

Process: Invoices Cumulative Time (ct) Select Time (t) All times in (minutes. hundredths of seconds)

Total Time Number Processed

op

Sorting/Matching Keying End-of-Day Activities

Process: Invoices Cumulative Time (ct) Select Time (t) All times in (minutes. hundredths of seconds) Major Work Elements Sorting/Matching Keying End-of-Day Activities Sample #3 Rating Factor Normal Time

Pr

Normal Time/Invoice

Total Time Number Processed

ga

Normal Time/Invoice

en

OM2 Supplementary Chapter A: Work Measurement, Learning Curves, and Standards

62564_09_cA.indd 21

e r o ge f ty Le ar ni ng
t ct t ct t ct 25.58 0.80 20.46 25.58 23.92 0.80 19.14 49.50 .00 7 0.80 5.60 56.50 45.20 t ct t ct t ct 15.69 0.85 15.69 .86 27 0.90 43.55 8.96 0.90 52.51 52.51 32.00 13.34 25.07 8.06 46.47 46.47 56.50 25.00 45.20 25.00 32.00 1.808 1.452 Sample #4 Rating Factor Normal Time 19.67 51.55 t ct t ct t ct 22.12 0.85 18.80 22.12 .74 34.68 0.80 27 56.80 19.22 0.75 14.42 76.02 60.96 76.02 t ct t ct t ct 19.67 1.00 19.67 51.55 1.00 71.22 26.05 1.00 .27 97 .27 97 61.00 26.05 .27 97 .27 97 61.00 1.595 60.96 30.00 30.00 2.032

Major Work Elements

Sample #1

Rating Factor

Normal Time

Sample #2

Rating Factor

Normal Time

A21
8/26/09 4:04 PM