Process Analysis Note
Operations problems are a part of every management career. An important first step to
managing an operation is analyzing the operation's process. A process is any conversion
activity where inputs are transformed into outputs. In a factory, materials, labor, and capital
inputs are used to create finished products. In a retail store, the output could be a satisfied
customer, while the inputs could be the customers, goods for sale, and an effective service
delivery system. In a law office or accounting firm, both the inputs and the outputs are
largely information: professionals convert raw data into information for the specific needs
of the client.
Process analysis involves understanding what the process does, how and why it does it, how
effectively it works, and how it might be improved. It is a diagnostic process that involves
technical analysis, observation, and judgment, much like the diagnostic work of a physician.
Process analysis assesses the work or conversion activity performed by some working unit.
In general, four activities are involved:
1. Creating a process flow diagram,
2. Analyzing the operating unit structure,
3. Analyzing the work flow, and
4. Evaluating the overall process.
THE PROCESS FLOW DIAGRAM
For a concrete example of how to apply process analysis, consider some of the problems
faced by a copy center in a School of Management, where packets of cases and other
materials are put together for selected courses and sold to students. All items at the copy
center are either copied or purchased from an outside supplier.
This note was originally prepared by D.C. McDougall, J.G. Miller, and T.E. Vollmann. It was revised in 1992 by Peter
Arnold, Janelle Heineke, and Jay Kim, revised and expanded in 1995 by Janelle Heineke, Peter Arnold, and J. Robb
Dixon, and updated in 2013 by Janelle Heineke. Copyright Boston University and the authors.
The four components always specified in every process analysis are tasks (or operations),
flows, decision points, and storages (or queues). Drawing a diagram of the process is a
useful first step. In a typical process flow diagram, tasks are depicted as circles or
rectangles, flows as arrows, decision points as diamonds and storages as inverted triangles.
Figure 1 on the next page is a simplified process flow diagram that illustrates the
preparation process for course packets. The process begins when professors select
collections of cases and readings for their courses and send them to the copy center.
Packets waiting to be processed are kept in storage.
Purchase
Materials
Packet Complete Receive and
Review Collate
Order Purchase Inspect Sell
Packet Order Materials
Order Materials
yes
Copy,
Complete Good no
Staple, Inspect Rework
Work Order ?
Drill
Detail of Copy Operation
Copy Staple Drill
Figure 1: Process Flow Diagram for Course Packet Preparation
The first operation is review of the packet. The second operation is preparation of a
purchase order. At the same time the purchase order is prepared, a work order can also be
prepared. Neither task must be performed before the other, therefore it can be said that
there is no precedence requirement. The work order specifies the number of units to be
produced at a time (the batch size), and the sequence of steps required (the flow or
routing).
For example, for copying done in the copy center the steps might be copy, staple, and three-
hole drill. A queue of work orders is shown in front of the copying operation. After copying,
each work order goes into another storage. When the purchased items have arrived and have
2
been checked in, the next operation, collating, can take place. The result is a storage or
inventory of finished packets that is depleted as the packets are sold to students.
Constructing a process flow diagram is something of an art. For a given process, there may
be more than one "correct" process flow diagram. The degree of detail incorporated in the
diagram depends on why it is being drawn. For an overall picture of a complex process, a
fairly simple diagram might be drawn, whereas if the diagram is being constructed to analyze
and improve a particular series of process steps, each step might be broken into very minute
detail.
THE OPERATING UNIT
The focus of analysis is the operating unit, in this case, the copy center in the School of
Management. The operating unit is enclosed in the box of Figure 1. In some analyses, this
focus might be a department of a company, an entire factory, or even an entire company.
Work orders (packet requests in our example) are one input to the operating unit. Another
input is purchased material (cases and readings). The inputs flow through the operating unit
and are transformed by the tasks performed into outputs.
WORK CENTERS
Within an operating unit, particular kinds of transformation operations are performed by
people or pieces of equipment, called work centers. All steps in a routing can be associated
with an identifiable work center. In our example, the routing specifies that three-hole drilling
in stacks of paper occurs at the drilling work center.
Flawless quality is one of the goals of operations management for many processes. This can
only be attained when the routing is precisely specified and there is firm control over how
work is done. In a large operating unit, such as a factory with thousands of workers and very
complex parts that require many routing steps, this detail will be an important part of the
factory information system.
PROCESSING TIME
Jobs or work orders that are routed through a particular work center will use some of its
capacity. In the copy center example, suppose a particular case order is produced in batches
of 20 pages and the machine can copy ten pages per minute. This job has a run time of two
minutes and would seem to consume two minutes of capacity. The run time is the sum of the
times required to complete the work for each task. However, for most tasks, it is necessary
to spend some time to prepare for the job and to clean up after the job is completed. This
time is called setup time. Setup time is independent of the number of items in the batch. In
our example, if it takes 0.7 minutes to load the item to be copied into the copier and 0.3
minutes to remove the item and the finished copies, then the setup time is one minute, and
the total time for the job is three minutes: two minutes of run time and one minute of setup
time.
CAPACITY
A key attribute of a work center or of an operating unit is its capacity. Capacity can be
defined as the ability to process units of work per unit of time and can be expressed in either
units of output (such as pages per minute or tons per day) or in units of time (such as
available hours per day), depending on which provides the most useful information for the
manager. The copy center's copying machine capacity, for example, could be expressed as
either 8 hours per day or 28,800 pages per day (60 pages/minute x 60 minutes/hour x 8
hours/day). The unit of time used depends on the analysis. Knowing the copy machine has
a capacity of 60 pages per minute is probably more useful for determining how long it will
take to copy a 20-page paper than knowing that the capacity is 7,200,000 pages per year or
that 2000 hours are available per year.
It is often more informative to express work center capacity in terms of available operating
hours per period of time rather than units produced per period of time because of the need to
consider setup. For example, it might be tempting to state the copy machine capacity in
terms of copies per hour, but it should be clear that at the end of a day the total number of
pages copied will be different if many small jobs are run as opposed to one very long job.
What if the jobs require switching colors of paper, print size reduction, or printing on both
sides of the pages? The desired output for the operating unit or work center is jobs.
Different jobs will have different attributes and, therefore, different capacity requirements.
The problem of jobs having different capacity requirements is much more significant in work
centers that perform work of great variety. In a restaurant, for example, a meal for a single
customer could be a cup of coffee and a salad or a five-course dinner. The capacity
requirements are clearly quite different.
Either machines or labor can constrain capacity at a work center. Whether capacity is
machine constrained or labor constrained can only be determined in particular situations
by analysis of the time requirements of the routing of the jobs for both the workers and the
machines. For example, if the same person performed three-hole drilling and stapling, the
only common denominator for capacity is hours of the worker. If both drilling and stapling
machines are faster than the worker's capacity to feed them, capacity is worker constrained
4
and not machine constrained. If, however, the pace of the machine determines how much
output can be produced, capacity is machine constrained.
THROUGHPUT TIME AND CYCLE TIME
The time required for a unit of output to pass through an operating unit — including waiting
time between operations — is called throughput time or elapsed time. Consider an
automated car wash with four distinct work centers connected by a conveyor belt: pay,
wash, rinse, and air dry. With no waiting time, the time it takes a single car to move through
the entire process is the throughput time, perhaps two minutes. However, on a sunny spring
Saturday it may take twenty minutes for a customer to complete the process. Throughput
time may vary widely because it includes waiting time.
The interval between completion of successive units in a process is called cycle time. The
interval between cars emerging from the car wash is the cycle time, perhaps every fifteen
seconds. Cycle time is used to determine the theoretical maximum capacity in units of
output per unit of time:
(time available)/(cycle time) = capacity in units
In the car wash example, if four cars can be washed per minute and the total time available in
a work day is eight hours, the capacity of the carwash in cars per day is:
(8 hours/day * 60 minutes/hour) * 4 cars/minute = 1920 cars per day
Throughput time and cycle time are both important concepts, but unfortunately, in practice
the labels for these concepts are not standardized. It is important to remember that
throughput time is associated with a particular job; it is the elapsed time from the start of a
job to its completion. Cycle time, on the other hand, is a characteristic of the system — the
time between completion of successive units.
Capacity utilization is a measure of how much of the available capacity is being used; it
relates the demand for the work center and the capacity (or supply) at the work center.
Capacity utilization is calculated by dividing the capacity required by the capacity available:
(Capacity required)/(capacity available) = capacity utilization.
For example, if the demand on a work center is for 30 units in an hour, but the work center
could produce 60 units in an hour, the capacity utilization is:
(30 units/hour)/(60 units/hour) = 50%
What is the right capacity utilization for a work unit? The answer is “It depends.” When
capacity utilization is close to 100 percent, a work unit’s productive resources are fully
employed. Efficiency is high, but other measures of the work unit’s performance, such as
delivery speed and quality, may be low. A textile mill needs to maintain a capacity utilization
of close to 95 percent to be efficient and remain profitable. A fast food restaurant operating
at 95 percent capacity utilization might be efficient, but it would not remain profitable for
long because it would not be fast. Customers’ expectations for cost, quality, and speed and
reliability of delivery, along with the firm’s goals for efficiency, determine the appropriate
capacity utilization for a work unit. There is not, however, a universal formula for calculating
what it should be. This is a matter for managerial judgment.
If a work unit’s capacity utilization is too high, what can be done about it? Capacity
utilization is a way of expressing the relationship between the supply of capacity and the
demand for it. To decrease capacity utilization, either capacity needs to be increased, or
demand must be decreased. One way to increase capacity is to add productive resources;
buying another copy machine for the copy center, for example. Another way to increase
capacity is to use existing productive resources more effectively (decrease cycle time).
Good workers and managers are constantly developing better methods for doing their work.
It is also possible for capacity utilization to be too low. Capacity utilization can be increased
by decreasing capacity or increasing demand. Decreasing capacity is usually achieved by
reducing productive resources, rather than by purposely using productive resources less
efficiently.
There are three basic ways, then, to affect the relationship between the capacity supply and
demand (capacity utilization): change the resources available, change the cycle time, and
change capacity required (demand).
INVENTORIES
A different kind of capacity relates to queues or inventories. Inventories can be of three
basic types. Raw materials inventories are inputs stored before processing. Inventories
6
held at or between work centers are work-in-process (WIP) inventories. Storages of
completed products or outputs are finished goods inventories. Inventories have physical
capacity constraints, such as the number of working papers that will fit on a shelf or the
number of customers that can wait in an airport gate waiting area.
All inventories provide a buffer against uncertainty. Raw materials inventories protect the
operating unit from supply problems. WIP inventories "decouple" tasks in a process, making
each work center less dependent on the previous work center for supplying jobs to be done.
Finished goods inventories buffer against demand fluctuations by keeping output ready for
customers when they order. Of course, there are costs related to holding inventory, and
inventories can mask problems in the production process as well.
Letting the input items wait in inventories or queues allows the system to be operated at a
higher level of capacity utilization. If each work center has a backlog of work it can be
operated with fewer interruptions. When the queues are large, however, average throughput
time goes up, even though the throughput time for any individual unit, can be shortened if it
is given a higher priority at work centers.
It is increasingly recognized that measuring efficiency and capacity utilization for individual
work centers can lead to suboptimization of an overall operating unit. Historically,
departments or work centers in many factories have been evaluated that way. The result is
pressure to release large amounts of work into the factory so that queues are large enough
for high capacity utilization and efficiency levels. Applying this logic to a fast food restaurant
might result in someone making a two week supply of French fries in order to highly utilize
the fryer and maximize its efficiency.
WORK FLOW
The flow of work is a major management issue in an operating unit. Some operating units are
able to accommodate quite different patterns of work flow and others are set up to operate
only in tightly prescribed ways. For example, the kitchen in a fast food restaurant is
organized to produce only a few items that follow similar routings. On the other hand, a fine
restaurant has a much wider menu, resulting in a larger set of routings for the kitchen and
more complex operations. The fast food restaurant can have the work centers operate at
higher levels of capacity utilization than the fine restaurant because the flow of work and
resultant capacity requirements are much more predictable.
The ability of an operating unit to tolerate wide variances in work order requirements is
called bandwidth. Some work centers and some operations have a greater ability to handle
variations in the routing requirements. This ability is typically acquired by choosing general
purpose equipment instead of special purpose equipment, and flexible layouts that
facilitate wide variations in the patterns of work flow. Bandwidth can also be considered in
terms of tolerance to volume surges. McDonalds, for example, can handle two busloads of
boy scouts that arrive unexpectedly, but a fine restaurant would not be able to manage such
a surge. McDonalds, then, has more bandwidth or higher volume flexibility than the fine
restaurant.
Surges in volume can be handled by designing the process so there is excess capacity
available to meet demand peaks, or so that extra capacity can be acquired very quickly. For
example, cross-training workers to do a variety of tasks provides increased flexibility for
accommodating peaking capacity requirement in any one work center, but not in all work
centers simultaneously.
BALANCE AND BOTTLENECKS
Maintaining the efficient flow of goods through the operating unit is a key management
concern. Balance refers to the relationship between cycle times in a process. When cycle
times of the different operations are unbalanced, some steps in the process are idle while
others are working all or most of the time. When a process is well balanced, the capacity
required for each operation is fairly similar.
For example, assume a copy center job involved three steps: copying, collating/stapling, and
packaging batches of faculty working papers. Each of these three steps takes ten minutes, on
average, for the worker at each workstation. This process is well balanced; all workers are
similarly busy and jobs flow through the process from one step to another with little, if any,
wait. You should note, however, that the tasks require 10 minutes on average, which means
that there is some variation in task times and that there will be some waiting depending on
the degree of variation in the task times.
What if, instead of collating/stapling, the process is changed, substituting 19-hole punching
and spiral binding at the second task? The spiral binding task requires 20 minutes per batch
of working papers, rather than 10 minutes for collating/stapling. What would happen?
If there is a stack of working papers to be processed in front of the copy workstation, that
worker will copy a batch of working papers every ten minutes. The spiral binder, who
requires 20 minutes to bind the batch, will find an ever-growing stack of working papers
accumulating at the binding workstation. The packaging worker will receive a batch of
8
working papers only every 20 minutes, and will, therefore, be idle for 10 minutes out of every
20. Though it still takes the packager 10 minutes to package a batch, he or she will only be
able to package a batch every 20 minutes because the pace of the process is determined by
the task with the longest task time. The spiral binding task is the bottleneck task: the task
with the longest task time. The bottleneck task time defines the cycle time for the process.
This three step process will complete batches of working papers only every 20 minutes.
If this process was allowed to continue in this unbalanced state, what would ultimately
happen to the worker at the copy work center? Eventually, because there would be no more
space to stack finished copies waiting for spiral binding, the copy worker would wait to begin
a copy job until the spiral binder took a job to bind. Ultimately, then, the bottleneck task
affects both tasks that follow it (downstream tasks) and tasks that come before it (upstream
tasks).
One way to avoid bottleneck problems is to schedule work so that capacity requirements are
more balanced. For example, if a large number of spiral binding jobs come in, a second spiral
binding machine could be purchased and another worker scheduled for that work center. If
two workers each perform a task that takes 20 minutes, the task time for that work center
will be 10 minutes, because two jobs will be completed every 20 minutes, or one job every
ten minutes on average. It might also be possible to release spiral binding jobs into the
operating unit in combination with other jobs that do not require capacity at that work
center. In the copy center example, this might mean releasing spiral binding jobs with jobs
that require neither stapling nor spiral binding, so that capacity required at the
binding/stapling work center is in balance with the copy and package work centers.
Balancing the work achieves a smooth flow through the operating unit.
TIME STANDARDS
In order to plan and control the capacity of equipment and personnel it is necessary to have
time requirement estimates for various routing steps. Work measurement is a field of study
encompassing analytical methods for studying work to find improvements, maximum
efficiency, and good time estimates for various tasks. The methods include direct
observation using stop watches and indirect synthesis of time standards for a task based on
adding times for well known parts of the task. For example, in the copy center, the time to
change paper color, the time to run a job of a particular number of sheets, and the time for
setup under normal conditions should be well known. These times can be added together to
determine time standards for more complex tasks.
A time standard is the time required to accomplish a particular task, given that the work
center is available for working on that operation. The time standard is based on some
expectation of worker and machine efficiency, and may or may not be achieved. For
example, Firm A might find that its workers operate at 120 percent efficiency, based on the
firm's time standards. Firm B may operate at 80 percent of its standards. It is not
necessarily true that firm A is more efficient or hardworking than firm B. Part of the
difference may be how the time standards were determined. The key requirement is
consistency if time standards are to be useful. It is easy enough to forecast actual time
requirements taking into account the relationship between actual and standard times as long
as the relationship is reasonably consistent.
Time standards are important for predicting the completion time of a particular job,
estimating its cost, planning future capacity needs, balancing work flows, reducing inventory
levels, and planning the use of work center capacities. All of these are concerns of the
management of an operating unit.
THE JOB SHOP VERSUS THE FLOW SHOP
Two extremes in operating unit design are the job shop and the flow shop. A job shop uses
general purpose equipment and personnel to deal with small batch sizes. It has great
bandwidth for routing and product diversity, long throughput times, low machine capacity
utilizations, many storage steps, a high ratio of setup time to run time, and seemingly random
bottlenecking at different work centers.
The flow shop is at the other extreme. The pure flow shop is an operating unit such as a
refinery or chemical plant where the process is turned on at some point and runs
continuously; neither individual units nor batches of units are produced. The equipment in a
flow shop is highly specialized for a relatively small group of products, work centers are
dominated by machines rather than people, throughput times are very short, work-in-process
inventory levels are low, and there are virtually no setups done.
It is useful to compare and contrast job shops and flow shops, although the vast majority of
operating units are at neither of these production process extremes. For example, an
emergency room is similar to a job shop, while the check processing operation of a bank,
where all the checks are processed on the day they enter the system, is similar to a flow
shop. In a retail environment a small store specializing in high fashion is more like a job shop
than a high volume store that carries ready-to-wear clothing.
10
The decision to design an operating unit as a job shop or as a flow shop has a long-lasting
impact on the effectiveness of the unit. Clearly the decision has to be made in accordance
with the market or customer the unit will serve. For a market where demands are highly
diverse, a job shop is more desirable. In contrast, if a firm decides to satisfy a limited scope of
demand (probably with high volume), a flow shop can meet the requirements more
efficiently. Mismatch between market requirements and process design can cause the
operating unit to be either too inefficient or too rigid.
MEASURING PERFORMANCE: COST, QUALITY, FLEXIBILITY, and
DELIVERY
Managers judge the performance of operating units on multiple dimensions. Measures of
cost, quality, flexibility, and delivery receive varying degrees of attention depending on the
firm's competitive priorities and the focus management chooses for performance
improvement.
Cost is an important measure of the performance of an operation. Costs that should be
considered include the cost of materials, the cost of labor (both direct and indirect), and
overhead costs. Labor and material costs are affected by the overall design of the process
and by how much waste and rework occur. Process analysis can provide insight about how
to more efficiently use resources to reduce costs without sacrificing performance along other
important dimensions.
Another important measure for operating units relates to the quality of the outputs
produced. Yield can be defined as the percent of good items in a batch (output
quantity/input quantity). An alternative measure is the scrap rate, which measures the
losses per period (1 – yield). Other quality measures are used in many firms, including
customer complaint rates, field service problems, conformance to design specifications, and
conformance to statistical quality control measurements. Defining quality is in itself a
difficult process for many firms, but a clear definition of quality is needed before quality
measurement can be considered.
The bandwidth of a process, as discussed earlier, is a measure of how flexible an operation is
in terms of either product diversity or volume variability. The design of a process will affect
how flexible it can be, but careful process analysis may provide insight into how to maximize
the flexibility of a given process design.
Dependable delivery is important to customers. The ability to meet delivery deadlines is
determined to a great extent by the process itself, but it is also determined by how well the
process is understood, and therefore how work is scheduled and expedited. If the manager
of a process does not have a solid understanding of processing and wait times for a given
order, it can be very difficult to accurately set delivery dates that can consistently be met.
12
SAMPLE PROCESS ANALYSIS QUESTIONS AND ANSWERS
1. Suppose you are operating a machine that produces only one type of part
and a part is produced every 30 seconds (its cycle time). What is the
capacity of your machine in units per hour?
time available/cycle time = capacity in units
time available = 60 minutes/hour * 60 seconds/minute = 3600 seconds/hour
cycle time = 30 seconds/unit
3600 seconds/hour/30 seconds/unit = 120 units/hour
Now suppose you were operating a machine that could produce three
different parts. Part A has a cycle time of 20 seconds, Part B has a cycle time
of 50 seconds and Part C has a cycle time of 80 seconds. Which expression
of capacity (units per unit of time or time) is most useful here? Why?
Because more than one type of unit can be produced, it is difficult to express capacity in units, so
expressing capacity in time is more useful.
If, however, the proportion of each type of unit was fairly consistent, you could calculate a
weighted average cycle time for the product mix. For example, assume the mix is 50% Part A, 30%
Part B, and 20% Part C. The weighted average cycle time would be:
20 seconds * .5 = 10 seconds
50 seconds * .3 = 15 seconds
80 seconds * .2 = 16 seconds
Weighted average = 41 seconds
2. Suppose you are a physician and the average length of time required to
meet and examine a new patient is one hour and the average time required
to meet and examine a return patient is 15 minutes. On any given day, the
mix of new and return patients is unpredictable. How would you express
capacity? Why?
As above, because you are uncertain of the mix, it would be more useful to express capacity in
time, rather than in units.
3. You are at Disney World and wait 45 minutes to ride the roller coaster at
Space Mountain. Guests board the roller coaster in batches of 20. The ride
takes 3 minutes and all riders get off at the same time. What is the cycle
time of this process? What is your throughput time?
Because Disney guests leave the ride every 3 minutes, the cycle time for the process is three
minutes per batch. The throughput time is the sum of the run time (3 minutes) and the wait time
(45 minutes), or 48 minutes.
4. The actual waiting time is important, but the customer's perception of the
wait is also important. When are waits most tolerable? Think of waiting
line experiences that you have not minded and identify some of the ways
the line was managed to make your perception of the wait positive.
Waits are generally considered to be more tolerable when they are fair (no "cutting"), when the
product or service is worth the wait (concert tickets), when the wait is occupied rather than idle
(watching previews at the movies), when the waiter is not anxious or upset, when the waiting area
is comfortable (reasonable temperature, a place to sit or lean), when the waiting lines are moving
and the waiter can feel the progress, when the wait can be used to incorporate part of the process
and shorten throughput time (such as filling out your check at the license bureau), when servers do
not seem to be idle, when a wait is expected (at your physician's office), or when a wait is part of
the experience (waiting for concert tickets overnight in the rain).
You may pleasantly recall the way lines are managed at Disney World (many turns, constant
movement, previews of what's to come, many distractions). You may also — not so pleasantly —
recall waiting outside your boss's office or in the emergency room.
5. You are the manager of a Surgi-Center that performs minor surgical cases.
The Center has 5 operating rooms and, on average, each case takes 60
minutes to perform. After each case clean-up requires 20 minutes. If the
Center schedules its first case for 8:00 a.m., no cases are performed
between 12:00 and 1:00 p.m., and the last case must be completed by 5:00,
what is the daily capacity of the Center in cases? What are two ways you
can increase the capacity of the Center?
Here, the total time for each case is the run time of 60 minutes plus the clean-up time of 20
minutes, or 80 minutes. Each operating room has 4 hours in the morning and four hours in the
afternoon that can be scheduled for cases.
4 hours * 60 minutes/hour = 240 minutes/half-day operating session
(240 minutes/session) / (80 minutes/case) = 3 cases/session
3 cases/session * 2 sessions/day/room * 5 rooms = 30 cases/day
14
The two basic ways to increase capacity here are to increase the time available each day by either
being open for longer hours or adding operating rooms, or to shorten the length of time it takes to
do the surgery and cleanup for each case. Here, because the day is broken into two four-hour
sessions, the total time required for each case (operating time plus clean-up time) cannot exceed
60 minutes to increase the capacity by one case each session. If more than 60 minutes is required,
the one-hour lunch is violated.
6. A process for making a product has five tasks, as shown in the table below.
All machines are used for each task. A shift consists of eight actual work
hours. Calculate the yield for the process overall, assuming defective units
are identified through inspection and scrapped at each task. Calculate the
shift capacity, in units, for each task.
Task Time for Number of Shift Capacity
Task Yield Each Machine Machines (units)
1 0.95 6 minutes 3
2 0.82 4 minutes 1
3 0.97 2 minutes 2
4 0.92 3 minutes 4
5 0.89 6 minutes 3
Total XXXXX XXXXX
What is the bottleneck of this process?
You are the production manager and you want to double production output.
What steps would you take? What issues would you need to consider?
Task Time for Number of Shift Capacity Shift Capacity
Task Yield Each Machine Machines (units) (units * yield)
1 .95 6 minutes 3 240 228
2 .82 4 minutes 1 120 98
3 .97 2 minutes 2 480 465
4 .92 3 minutes 4 640 588
5 .89 6 minutes 3 240 213
Total .619 XXXXXXX XXXXX 120 XXXX
The overall yield for the process, when defective units are scrapped, is calculated by multiplying the
yields for each step (.95 * .82 * .97 * .92 * .89 = .619). The logic here may be easier to see if you
think in terms of units. If you start out at Task 1 with the materials required to produce 1,000 units,
you will only produce 950 good units to go on to Task 2; of those 950 good units, Task 2 will only
yield 779 good units (950 * .82) to go on to Task 3, and so on.
To complete the rest of the table above, you need to consider the number of machines at each task
to determine the theoretical shift capacity in units for each task. It is also appropriate to consider
the yield at each step and calculate the actual capacity at current yields for each task. [Note that
there are hidden dangers in either calculation. If you calculate only the theoretical capacity, which
assumes perfect quality, you are overstating what you can actually do with current yields. If,
however, you calculate capacity factoring in current yields, you may be, in effect, assuming that the
yields are fixed and that your capacity is necessarily constrained to those levels.] Incorporating
yields into the calculation of total capacity is a bit complicated, because yield rates before the true
bottleneck task don’t affect the overall capacity, but yield rates at each task after the bottleneck
task do.
Identifying the bottleneck here is a little tricky. The longest task times are at Tasks 1 and 5, so they
might be expected to define the bottleneck task. Because there is more than one machine at
several tasks, however, the capacity at each task determines the capacity for the process, and
therefore the cycle time for the process. Task 2 is the bottleneck task.
You need to look to the bottleneck, because capacity at that task determines capacity for the
system. The capacity at Task 2 is 120 units per shift so you could either add a machine or think
about adding a shift for that task. If you identified the need to consider yield as affecting capacity
here, which is correct, you would need to demonstrate what you would do to change it. Other
issues to consider are equipment cost, labor cost, space, and market demand.
16