Professional Documents
Culture Documents
Production Planning & Control: The Management of Operations
Production Planning & Control: The Management of Operations
Production Planning & Control: The Management of Operations
To cite this article: S. S. Chakravorty & J. Brian Atwater (2006) Bottleneck management: theory and practice, Production
Planning & Control: The Management of Operations, 17:5, 441-447, DOI: 10.1080/09537280600682752
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
Production Planning & Control,
Vol. 17, No. 5, July 2006, 441–447
Advocates of TOC believe that bottleneck resource restricts an operation’s ability to make
money, and the best way to maximise income is to fully exploit the bottleneck resource.
Almost all TOC literature focuses on situations where 100% bottleneck utilisation is applied.
Based on the implementation experience, the finding is that the optimal level of bottleneck
utilisation should be less than 100% and any attempt to increase utilisation beyond the
optimal level brings disastrous results for a door manufacturing plant. In order to improve
and maintain the performance of the plant effective bottleneck management is critical.
The experience provides a deeper understanding of how to design such plants, which could be
beneficial for practising managers and academics working with TOC concepts.
In section 2, a review of relevant literature is investigating the optimal level of bottleneck utilisation.
presented followed, in section 3, by a discussion This case study examines the best level of bottleneck
of methodology of the study is provided. In section 4, utilisation in a door manufacturing operation.
the implementation experience in improving the perfor-
mance of a door manufacturing plant is delineated.
Lastly, conclusions from this study, including directions 3. Methodology
for future research, are presented in section 5.
3.1 Case study
A case study approach was used to document the
2. Literature review
application of bottleneck management in improving
a door manufacturing plant. According to Yin (1994)
Recognition that manufacturing operations should
there are three reasons why a case study is appropriate
allow a certain percentage of capacity slack is not new.
for our objectives. First, when ‘how’ and ‘why’ questions
Monden (1984) claimed that 10% extra capacity is
are investigated, the case study approach is preferred
necessary in the production lines to accommodate
Downloaded by [University of Chicago Library] at 10:41 20 November 2014
2. Quantitative data was collected in the form of products with facilities in Georgia, USA at Duluth,
archival records of financial data, customer Flowery Branch, and Ball Ground. The plants used in
complaint reports, ordering processing, manufac- this study were their manufacturing plants located in
turing and quality reports, and shipping reports, Duluth and Flowery Branch. These plants employed 123
purchase orders, operational data (such as managers and workers who process doors, windows,
machine and labour utilisation), routing informa- and millwork. Their door manufacturing plant consisted
tion, performance measurements (such as annual of four separate door lines, which used Norfield, and
sales and lead-time). other custom equipment. At the time of this implemen-
3. Additional qualitative data was collected through tation, the company held an estimated 5% of the
extensive interviews with senior managers, super- southern-eastern residential and light commercial build-
visors, and employees. Based on the recommen- ing product market.
dations of Fontana and Frey (2000) the interviews The redesign effort was initiated to reverse the
were conducted in an open-ended nature, which company’s deteriorating performance. At the time the
allowed respondents to provide objective opinions study was conducted these plants were experiencing
of the events as well as insights into certain several problems. For example, in the 3 months
Downloaded by [University of Chicago Library] at 10:41 20 November 2014
responsible for accepting orders into the manufacturing. the entire operation and they establish the
The production writer turns an order into a production sequence and the processing time of the opera-
order and delivers the order to the supervisor of the tion. The cycle time reduction of this operation or
door manufacturing. The supervisor manages the order parallel processing of some parts of the operation
through the manufacturing and is ultimately responsible was restricted by the type of machines.
for on-time delivery of orders. 3. The group initially found that the bottleneck
The group found serious problems with the order resource experiences about 40% down time,
processing stage of the door manufacturing operation. which consists of breakdowns, repair, routine
An elaborate information system was installed to track maintenance, and setup times.
orders (e.g. customer number, jobsite, order due date,
Further analysis revealed that breakdowns, repair,
order release time, and order status report, etc.).
and routine maintenance were typical to Norfield
No capacity check was performed before orders
machines. However, the setup times of the machines
were accepted into the door manufacturing. The
were about 20 minutes. The group applied SMED
prevalent notion was that ‘we cannot influence how (a.k.a. quick change over) techniques (Shingo 1985) to
Downloaded by [University of Chicago Library] at 10:41 20 November 2014
our customers’ order and we have to do everything to reduce the setup times to 2 minutes, which reduced
take care of our customers’. To remain competitive in the bottleneck resource downtime to about 30%.
the business environment, the company must offer short
lead-times (e.g. rush 4 hours) with a wide variety of
products. This prohibited the door manufacturing from 4.3 Determining the appropriate bottleneck
stocking large amounts of finished goods inventory. resource utilisation
As a result the group found that door manufacturing
was overloaded with orders and manufacturing was Although there are four door lines in the entire door
perpetually behind schedule and was encountering manufacturing operation, in order to protect proprie-
tary information the subsequent analysis is presented as
serious problems with their due date performance.
The order manufacturing stage involves the manu- if there is only one door line. The processing times and
the subsequent analysis are presented in single door
facture of door units and basically consists of four steps.
units. Other door units (e.g. single units with transoms
1. An automated operation performed by Norfield or double units) are converted to equivalent single door
machines that involve loading a stack of door units. Using 8 hours (480 minutes) of working time per
units, preparing the door units, and applying day and 300 units of customer demand rate per day, the
hinges to them. required takt time is 1.6 (480/300) minutes. Considering
2. A manual operation that consists of pulling the 30% down time, the group was aware that 1.95
door unit and placing head and side jambs on it, (1.5 1.3) minutes cycle time of bottleneck resource
then pushing it to a conveyor line. was not sufficient to meet the takt (the rate at which
3. A manual operation that nails a casing on the each product needs to be completed to meet customer
door unit, flips it over, and pushes the unit to demand) time. The group had to establish initial level
the next stage on the conveyor. of bottleneck utilisation in order to continue to do
4. This is also a manual operation that nails casings business.
and skid blocks and then places the unit on a cart. The group studied customer order and production
reports to establish the initial level of bottleneck
The group carefully studied the operational details
utilisation (see table 1). Without considering variation
of the door manufacturing and made three major
at the bottleneck resource, average daily door produc-
observations.
tion is 224 {(1 0.3) (480 minutes/1.5)} door units.
1. The group found that the manufacturing typically Two observations can be made from the table. First, as
operates about 8 hours a day. Considering long- the bottleneck utilisation increased from 71.4% to
term demand, the group was unable to provide 80.3%, 20 additional door units were accepted, yielding
cost justification to add a second shift to the a $1280 increase in throughput generation. However,
manufacturing. this increase in throughput was offset by a decrease
2. There are four steps in the door manufacturing in on-time delivery performance from 98% to 89%.
process. The average processing time is as follows: Second, as the bottleneck utilisation increased from
Station 1, 1.5 minutes; station 2, 0.5 minutes; 80.3% to 84.8%, the performance of the door manu-
station 3, 0.5 minutes, station 4, 0.5 minutes. The facturing worsened. In this instance, only 10 additional
bottleneck resource is the first station, which is an door units were accepted, yielding only $640 increase
automated operation. Norfield machines perform in throughput. However, this increase in throughput
Bottleneck management: theory and practice 445
Table 1. Bottleneck resource utilisation and manufacturing performance.
Note: The data is extracted from the months of June, July, and August.
was offset by a huge decrease in on-time delivery As a result the group introduced a new position
performance from 89% to 70%. ‘Order Manager for Door Orders’ (note this individual
In the highly competitive environment of the also handles windows and other millwork as well).
company, the door manufacturing aims for about 90% The primary responsibility of this new position was to
on-time delivery performance, yielding an optimal level evaluate orders and make a decision regarding
Downloaded by [University of Chicago Library] at 10:41 20 November 2014
of bottleneck utilisation of 80.3%. Prior to this analysis, whether to process the order at the door manufactur-
no one would have believed that by accepting an order ing plant or subcontract the order to another
of 10 doors the manufacturing’s on-time delivery manufacturer. The number of door orders accepted
performance would deteriorate so significantly. This for processing at the plant is based on bottleneck
finding implies that the temptation to accept additional resource utilisation, which is determined by perform-
orders beyond the optimal amount should be avoided, ing daily capacity check on the bottleneck resource.
as this is not beneficial for the company’s delivery Under no circumstances is the optimal utilisation of
reputation, which will hurt future business. the bottleneck resource exceeded. In order to manage
orders subcontracted to other manufacturers, two
actions were taken.
4.4 Managing the bottleneck resource utilisation
1. A group of small manufacturers with excellent
Given that there was a large variation (e.g. all door
lead-time and quality performance was identified.
orders 150 to 450 door units) in door orders from one
2. Changes to the information system were made
day to another, the group knew that managing the
optimal level of bottleneck resource utilisation was to accommodate the release, follow-up, and if
going to be challenging. The group decided to approach necessary expediting of orders to meet the
these issues in two ways. company’s lead-time and quality expectations.
First, the group was interested in studying how
customers placed their door orders. The group
thoroughly analysed the customer order reports and 5. Conclusions
found that seven customers contributed to about 70%
of the door orders. The group interviewed these There are two major conclusions from this study. First,
customers to understand what factors significantly consistent with theoretical findings, our implementation
impacted their decision to place a door order at a experience finds that optimal bottleneck resource
particular time. The group found out that frequently utilisation is less than 100% (specifically 80.3%) and
customers batched their orders because the company any attempt to increase the utilisation brought disas-
gave volume discounts on the building material such trous results for the door manufacturing plant. Based on
as door units. The group developed a new discount this implementation experience, the advocates of TOC
structure based on how evenly customers placed their should abandon promoting 100% bottleneck utilisation,
orders. Five out of seven customers adopted this new as this activity is wasteful. Of course, the optimal level
discount structure and the remaining two continued to of bottleneck utilisation will dependent on the type
use the old discount structure. Within a few months, of manufacturing operations. Manufacturing plants
the group observed that the variation in door orders experiencing higher (or lower) level fluctuations than
significantly reduced (e.g. all door orders 250 to 350 the plant will require lower (or higher) than 80.3%
door units). The group realised that since this was bottleneck utilisation. In addition, sensitivity around the
a dynamic environment in which the customer base optimal level of bottleneck utilisation was observed, and
changes and so do their needs, the new discount as the bottleneck utilisation was increased beyond the
structure alone would not be sufficient to effectively optimal level, the door manufacturing on-time delivery
manage the incoming door orders. performance deteriorated quickly. Therefore, it is
446 S. S. Chakravorty and J. B. Atwater
policy of offering discounts based on volume. A new pp. 567–606, 2000 (Sage Publications: Thousand Oaks, CA).
discount structure based on how evenly the customers Goldratt, E.M., Theory of Constraints, 1990 (North River
placed their orders was developed. Over time, this Press Inc.: Croton-on-Hudson, New York, NY).
structure dampened the door order variation consider- Goldratt, E.M. and Cox, R., The Goal, 2nd revised ed., 1992
(North River Press Inc.: Croton-on-Hudson, New York,
ably. This is a significant finding as many companies NY).
experience large variation in their customer orders and Hopp, W.J. and Spearman, M.L., Factory Physics, 1996 (Irwin
instead of trying to dampen the variations they develop Publishers: Boston, MA).
better operations to deal with this problem. More Krajewski, L.J., King, B.E., Ritzman, L.P. and Wong, D.S.,
research in this area would be very beneficial. Kanban, MRP, and shaping the manufacturing environ-
ment. Manage. Sci., 1987, 33(1), 39–57.
Obviously, business is a dynamic environment and Meredith, J., Building operations management theory through
relying only on new discount structure was not case and field research. J. Op. Manag., 1998, 16, 441–454.
sufficient. In order to maintain the optimal level of Monden, Y., A simulation analysis of the Japanese just-in-time
resource utilisation, procedures to sub-contract the technique (with kanbans) for a multiline, multistage
production system: a comment. Dec. Sci., 1984, 15, 445–447.
orders were developed.
Nakajima, S., Maximising equipment effectiveness. In
Finally, based on our implementation experience, Continuous Improvement in Operations, edited by
bottleneck management is an important area of A. Robinson, pp. 307–331, 1991 (Productivity Press Inc.:
research. This study shows an optimal level of bottle- Cambridge, MA).
neck resource utilisation is necessary to improve a door Rosenthal, R. and Rosnow, R., Essentials of Behavioral
Research, 1991 (McGraw-Hill Publishers: Boston, MA).
manufacturing plant. This is particularly relevant as Shingo, S., The Revolution in Manufacturing: The SMED
today’s competitive environment demands that compa- System, 1985 (Productivity Press: Portland, OR).
nies respond to the customers’ needs and wants at a Spearman, M.L., On the theory of constraints and the goal
faster rate than ever before. There is increasing pressure system. Prod. Op. Manage., 1997, 6(1), 28–33.
to pursue new ways of thinking as a source of Stake, R.E., Case studies. In Handbook of Qualitative Research,
2nd ed., edited by N.K. Denzin and Y.S. Lincoln,
competitive advantage. More research in this area pp. 435–454, 2000 (Sage Publications: Thousand Oaks, CA).
should contribute to the science and practice of bottle- Yin, R.K., Case Study Research: Design and Methods, 1994
neck management. (Sage Publications: Thousand Oaks, CA).
Bottleneck management: theory and practice 447