International Journal of Production Research,

Vol. 45, No. 10, 15 May 2007, 2317–2325

The Pareto managerial principle: when does it apply?

ABRAHAM GROSFELD-NIRy, BOAZ RONEN*z and NIR KOZLOVSKYx

yAcademic College of Tel-Aviv-Yaffo, Israel

zTel Aviv University, Faculty of Management, Israel
xTel Aviv University, Faculty of Management, Israel

(Revision received March 2006)

The well-known 80/20 Pareto principle and its refinement into ‘‘A’’, ‘‘B’’,
and ‘‘C’’ categories give rise to a managerial methodology consisting of
three steps: classification; differentiation; and resource allocation. This is an
easy-to-implement and extremely effective methodology. It starts with the
creation of ‘‘Pareto diagrams’’, i.e. bar charts of attributes and their relative
frequency, presented in descending order. Typically, Pareto diagrams are useful in
that they provide managers with a summary of practical information, revealing
critical attributes. However, sometimes a Pareto diagram is less informative than
it might be, because the relative frequency is almost uniform. The objective of
this article is to provide an analytical tool (an index) that employs the above-
mentioned methodology to measure the closeness of empirical Pareto diagrams to
an ‘‘ideal’’ Pareto diagram. The index developed is based upon entropy.

Keywords: Pareto principle; Entropy; Theory of constraints

1. Introduction

The overwhelming flow of information forces busy managers to seek focusing

methods that allow for efficient utilization of limited time resources. Among these
methods are the well-known theory of constraint (TOC), which advocates the
detection of critical resources (bottlenecks), and total quality management (TQM),
which advises focusing on resources and activities imperative for quality. In this
context, the Pareto methodology, described below, is one of the most important
managerial tools that can be used to identify factors worthy of attention.
Early in the 20th century, the economist Vilfredo Pareto stated that 20% of the
population possesses 80% of the wealth. This basic 80/20 principle was then adopted
in describing other realities; and, later, enhanced to the ‘‘A’’, ‘‘B’’, ‘‘C’’ classification,
in which the ‘‘A’’ group (the ‘‘vital few’’), consisting of approximately 20% of the
attributes (items), accounts for 80% of the phenomenon; the ‘‘B’’ group, i.e. the
next 30% of the items, accounts for 10% of the phenomenon, and the ‘‘C’’ group
(the ‘‘trivial many’’), which contains 50% of the items, accounts for only 10% of the
phenomenon.

*Corresponding author. Email: [email protected]

International Journal of Production Research

ISSN 0020–7543 print/ISSN 1366–588X online ß 2007 Taylor & Francis
http://www.tandf.co.uk/journals
DOI: 10.1080/00207540600818203
2318 A. Grosfeld-Nir et al.

Many researchers find it useful to model managerial scenarios according to the

Pareto principle: approximately 20% of the projects generate 80% of the revenues
(or profits); 20% of the customers generate 80% of a firm’s earnings; 20% of the part
numbers in a given bill of materials (BOM) account for 80% of the value of all parts.
The same goes for stock keeping units (SKUs) at the supermarket and so on. Koch
(1998) presents examples in many areas of management, including business,
marketing, finance, human resources, production and logistics.
Ronen et al. (2005) suggested and explored a ‘‘Pareto focusing methodology’’
that consists of three steps:
. Classification – sort out symptoms (attributes) of a phenomenon to build
a Pareto diagram; use the diagram to identify the three classes ‘‘A’’, ‘‘B’’
and ‘‘C’’.
. Differentiation – set a specialized policy for each class.
. Allocation – assign resources accordingly.
Typically, a useful or ideal Pareto chart (figure 1) is one where about 20% of the
attributes have an 80% weight in terms of relative frequency, thus revealing critical
features. However, sometimes, a Pareto chart is less informative, because the
different attributes occur with almost uniform relative frequency (figure 2).
While some empirical Pareto charts are easy to classify as useful or useless
(almost uniform), many others are in between and it is not absolutely obvious
whether the three-step Pareto focusing methodology applies. The objective of this
article is to provide an analytical tool (an index) which advocates the use of the
methodology to measure the closeness of empirical Pareto charts to an ‘‘ideal’’
Pareto chart. The index developed is based upon entropy.

30
25
20
15
10
5
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Figure 1. A differentiating Pareto diagram.

10
8
6
4
2
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Figure 2. An almost uniform Pareto diagram.

 Pareto managerial principle 2319

Section 2 of the paper presents the literature and research carried out. Section 3
describes the motivation for developing the tool. Section 4 develops the entropy-
based index to measure the closeness of a given Pareto chart to an ideal chart.
Section 5 demonstrates the application of the model. Section 5 summarizes and
concludes the paper.

2. Literature

Many managerial tools are useful in that they reveal attributes worthy of attention.
Such as, for example, the TOC, advocating a focus on critical resources
(bottlenecks); TQM, advocating a focus on resources and activities imperative for
quality; and the ‘‘complete kit’’, advocating starting a job only after all the required
parts are available. For literature describing the TOC, see Goldratt (1990); Ronen
and Spector (1992); and Mabin and Balderstone (2000). The literature relating to
TQM is vast: e.g. Deming (1985) and Juran (1988). The ‘‘complete kit principle’’ is
described in Ronen (1992) and Grosfeld-Nir and Ronen (1998). In this context the
Pareto principle, coupled with Pareto charts, is an extremely useful and easy-to-
implement tool that reveals attributes worthy of attention.
Despite its high potential value and many applications, the Pareto principle has
been the subject of few research and academic studies. Herbert (1995) suggested the
use of the principle in managing storage houses: the use of Pareto charts allowed
focusing on stored items and placing them in the warehouse in the most optimal way.
Gleason (1995) applied the Pareto principle to the paper mill industry, using Pareto
charts to identify frequent failures in order to increase machine utility. Indeed, the
America-Israel Paper Mills used Pareto charts successfully for several years as a tool
to increase productivity of high-cost resources. The approach was labeled ‘‘lost time
analysis’’ (LTA) (Gleason 1995). A banking application of the Pareto principle was
employed by Hales (1995) to locate and develop profitable sectors for targeting
marketing efforts. An information systems research study using the Pareto principle
was conducted and reported by Ronen and Spiegler (1991), where the method-
ological analysis performed on an insurance database demonstrated that 20% of the
data are used to generate over 80% of the information accessed. In an age of
exponentially growing databases, this finding has a significant effect on the design of
databases and storage technology.
Temme (1994) established prioritization rules taking Pareto as the first and
foremost among them. He suggested that the Pareto principle be employed to set
priorities for focusing on the topics requiring most management attention.
An interesting case study describing a successful and insightful application of
Pareto in a large industrial firm is reported by Tatikonda et al. (1999). In this study,
the firm decided to build a pricing system based on the ABC activities. At the start
of this long and expensive project, five senior workers in the accounting department
were sent to a seminar on the subject. Then, a steering committee of 15 workers
from the various departments was established. After 2 years, in which thousands
of man-hours and direct software, with a cost of $200 000, were invested, the first
conclusions were formulated, the main one indicating that low-volume products
generate higher indirect costs than high volume products. But, as the ABC process
matured, the firm was acquired by a larger enterprise, and the new management
2320 A. Grosfeld-Nir et al.

immediately stopped the project. Instead, it decided on a series of steps all stemming
from the Pareto principle: focusing on the 20% of the items that contribute 80% of
the yield, cutting the number of suppliers and reducing the number of purchased raw
materials. These steps, in addition to a ‘‘just in time’’ (JIT) implementation and
quality-management methods led to significant achievements in many areas. Firm
overheads diminished and various coordination and control departments were
eliminated. The purchasing department, for example, which had previously
employed 28 workers ended up with only three workers.
A model that analyses when the Pareto rule applies was not found in the
literature.

3. Pareto charts

Statistically, a Pareto chart is simply a frequency block diagram displaying the

relative frequency of different attributes in descending order. This classification is an
essential step that must precede taking the corrective measures of differentiation and
allocation.
For example, suppose that while studying a certain factor, such as the number of
defectives, classification reveals 15 attributes. Thus, according to the Pareto
principle, approximately 20%, i.e. three attributes, are the principle source of 80%
of the defectives, and focusing on improving these three attributes may eliminate
most of the defectives.
It should be noted that the creation of a Pareto chart, i.e. classification, is usually
a quick and inexpensive information gathering action. It is the implementation of the
next two steps that requires a relative lengthy organizational effort. However, as the
selection of the attributes which are candidates for corrective measures depends upon
the outcomes presented in the Pareto chart, the information gathering effort needs
special care, as mistakes in the classification may lead to irrelevant correction steps.
Although many researchers have reported applications where Pareto charts
reveal that the 80/20 principle holds true, sometimes, in practice, certain
classifications produce Pareto charts which are less managerially helpful in
that the relative frequency corresponding to 20% of the attributes is much less
than 80%, leading to a situation where focusing on a small number of attributes is
impossible. Also, after obtaining a Pareto chart, managers sometimes find it hard to
reach agreement about how useful the chart is in identifying a small number of
critical attributes, that is, how closely the diagram fits the 80/20 principle. For this
reason we develop a quantitative tool, followed by a simple index, to measure the
closeness of Pareto charts to the 80/20 principle.
Taking a slightly broader point of view, a manager observing a Pareto chart may
wonder how useful this particular chart is in guiding him to focus on a small number
of attributes which are a major source of problems. For example, table 1 and figure 1
describe a situation where, intuitively, it would be desirable to focus on the first three
attributes (20%). In contrast, figure 2 displays a reality where the Pareto chart is
almost uniform and, therefore, the Pareto principle should not be used.
Our goal is to provide a numerical index which will indicate that, for example, the
attributes of figure 1 fit the three-step Pareto methodology while the attributes of
figure 2 do not. Our approach is similar to the one used in statistics, where 95% is
 Pareto managerial principle 2321

Table 1. Pareto table.

Attribute % Attribute % Attribute %

1 25.2 6 3.9 11 2.1

2 21.1 7 3.1 12 2.1
3 18.2 8 2.9 13 1.6
4 7.1 9 2.7 14 1.5
5 5.2 10 2.1 15 1.2

35
30
25
20
15
10
5
0
1 2 3 4 5 6 7 8 9 10

Figure 3. A Pareto control limit chart with 10 attributes.

used as a control limit to imply significance. Note that the 95% control limit is a
convention based upon expert opinion. Similarly, based upon expert opinion we
develop a Pareto index, with control limits, to indicate that a certain Pareto
classification is ‘‘significant’’, i.e. fits the Pareto methodology.

4. Pareto charts fitting the three-step methodology: using entropy

Consider a Pareto chart with 10 attributes; then, clearly, if the relative frequency
associated with the first two attributes is 80%, or so, the Pareto differentiating
methodology applies. However, if the relative frequency associated with the first two
attributes is much less than 80%, the Pareto differentiating methodology should not
be used. After deliberation and consulting with experts we selected as a ‘‘control
limit’’ a Pareto chart where 20% of the attributes represent a total of 60% of the
relative frequency and the remaining attributes are uniform. For example, when
there are 10 attributes, figure 3 represents the control limit chart.
To avoid confusion the relative frequencies of control limit charts for 5, 10, and
15 attributes are:
5 attributes : 60%,10%,10%,10%,10%:
10 attributes : 30%, 30%, and eight times 5%:
15 attributes : 20%, 20%, 20%, and 12 times ð40=12 ¼Þ3:33%:

The entropy corresponding to 5 attributes is:

H5 ¼
0:6  ln 0:6
4  0:1  ln 0:1 ¼ 1:77
H10 ¼
2  0:3  ln 0:3
8  0:05  ln 0:05 ¼ 2:77
H15 ¼
3  0:2  ln 0:2
12  0:0333  ln 0:0333 ¼ 3:35
2322 A. Grosfeld-Nir et al.

Table 2. Control limits based upon entropy.

No. of activities 5 10 15 20 25 30 35 40 45 50
Control limit 1.77 2.77 3.55 3.77 4.09 4.36 4.58 4.77 4.94 5.09

6
5
4
3
2
1
0
0 5 10 15 20 25 30 35 40 45 50

Figure 4. Control limits based upon entropy.

These numbers serve as control limits, namely, if, for example, a Pareto chart
with 10 attributes has entropy exceeding the control limit 2.77, it should not be
treated according to the Pareto methodology.
Table 2 and figure 4 display the control limits for Pareto charts with up to
50 attributes. If the number of activities is not a multiple of five, interpolations
can be used.
For example, the entropy corresponding to figure 1 is h ¼ 3.1653.35 ¼ H15,
which implies that the Pareto methodology is adequate. On the other hand,
the entropy corresponding to figure 2 is h ¼ 3.8743.35 ¼ H15, which implies that the
Pareto methodology should not be used.

4.1 Rule of thumb

Sometimes, observing the Pareto frequency chart, it is obvious that it is adequate
for the use of the Pareto methodology, and the formal use of the entropy index
becomes a burden. Mathematically, the following ‘‘rule of thumb’’ is a sufficient
condition: if the total frequency corresponding to 20% of the attributes is above 60%
the entropy index is below the control limit, i.e. the use of the Pareto methodology is
in place.

5. Applying the model

We will now demonstrate the application of the model, using the three-step
methodology.
 Pareto managerial principle 2323

5.1 Case I: a company in the aviation industry

In 2002 the company described was manufacturing about 800 different high-tech
products aimed at the aviation industry. However, in spite of an impressive
monetary figure for total sales net profit was low. It has been suggested that the
company needs to improve its focus in terms of product variety and marketing.
. Classification. First two methods of classification were used: (a) an overall
classification of all products according to their monetary contribution;
and (b) products were sorted into 14 group technologies; then groups were
classified according to their monetary contribution.
The overall classification revealed that class ‘‘A’’, consisting of about 160
products (20%), accounts for 78% of the contribution; class ‘‘B’’ consisting of about
240 products (30%) accounts for about 14% of the contribution, and class ‘‘C’’
(the ‘‘trivial many’’; 50%) consisting of about 400 products, accounts for only 8% of
the contribution. Moreover, almost all the products of class ‘‘C’’ belonged to six
group technologies. Note that, using the ‘‘rule of thumb’’ (end of section 4) for the
overall classification, the entropy index recommends the use of the Pareto
methodology. Also, omitting details, calculating the entropy index for the group
technologies, gave the same results.
. Differentiation: the company applied a different policy for each class.
Management concluded that there were too many parts requiring marketing
and logistic efforts with relatively low added monetary value. Further
investigation revealed that low profitability for some products was expected
as in the market there were competing low-price products. The company
presented three different policies to the three part classes.
. Class ‘‘C’’: about 250 products were discontinued. This eliminated two group
technologies, the residuals of which were merged with others.
Discontinuation was carried out in two ways: for some parts, a notification
stating the discontinuation of the parts was sent to the customers, allowing
time to carry stocks; for other parts, a significant increase in price was
announced, explaining the extra costs needed to continue holding the parts
active. This policy intended many ‘‘C’’ parts to discontinue, while others
move to the ‘‘B’’ class.
. Class B: use some of the available sales force to increase sales of selected
items, trying to turn them into class ‘‘A’’ items. Other than that, basically, no
change in activities.
. Class A: customers will get special attention; their needs will be learned and
their satisfaction will be monitored in great detail. Marketing will make an
effort to identify potential customers (currently buyers of competing
products) and offer specials and quality service for selected products.
. Allocation: most of the sales resources were allocated to the ‘‘A’’ items, and
the rest to the ‘‘C’’ items.
Within one year the Pareto chart moved to a more even shape, where about 30%
of the parts had 60% of the contribution. The ‘‘fear’’ that many customers would
2324 A. Grosfeld-Nir et al.

quit unless they were offered a ‘‘full suite’’ proved to be ungrounded. Although total
sales dropped, net profit increased by an impressive percentage.

5.2 Case II: a supermarket chain

In 2001, the supermarket chain described here owned 220 stores. Again, in spite of an
impressive monetary figure for total sales, net profits were low. When a new
management team came into office, they started to analyse the profitability of the
chain via the stores. First, they analysed the contribution of each store and drew a
Pareto chart. Then, they applied the Pareto methodology as follows:
. Classification: they classified the stores by their contribution into ‘‘A’’, ‘‘B’’
and ‘‘C’’. As the ‘‘A’’ stores (20%) contributed over 60% of the revenue it
became clear that the entropy index was below the control limit, showing that
there was a potential to carry on the analysis using the contribution per store
as a leading variable.
. Differentiation: the ‘‘A’’ stores went through a retention program to keep the
customers in the chain, and increase their contribution. The ‘‘B’’ stores went
through a growth program that increased the contribution per store. The ‘‘C’’
stores were analysed: some of them were closed down (especially in residential
areas that turned commercial).
. Resource allocation: most of the managerial attention was given to the ‘‘A’’
and ‘‘B’’ stores. Also, managers of ‘‘A’’ stores were asked to spend time at
‘‘B’’ and ‘‘C’’ stores to instigate changes in terms of managerial methods and
product variety; and managers of ‘‘B’’ and ‘‘C’’ stores were asked to spend
time at ‘‘A’’ stores to see if they could adopt methods and products.
At the same time management analysed the stores’ achievements by analysing the
contribution per dollar of rent of each store. This parameter enabled management to
compare small and large stores with the same metric.
. Classification: the stores were classified into ‘‘A’’, ‘‘B’’, and ‘‘C’’ types.
An index analysis showed a significant potential for using the Pareto
methodology.
. Differentiation: the ‘‘A’’ stores were focusing on retention, while the ‘‘B’’
stores focused on growth. The ‘‘C’’ stores were candidates for management
replacement, or closure.
. Resource allocation: most of the resources were allocated to the ‘‘A’’ and
‘‘B’’ stores.
. After 3 years of continuous improvement, an index analysis was carried out
again. The figures revealed that these two parameters (contribution per store
and contribution per dollar rent) were no longer significant.

6. Conclusions

Pareto charts, describing the relative frequency of a main factor as a function of

attributes, make for an extremely useful managerial focusing tool. However, the
charts must be differentiating, i.e. the relative frequency associated with a small
 Pareto managerial principle 2325

number of attributes must be high enough. In this context, the presence of an

analytical tool helping to identify useless charts is important.
Our index, based on entropy, combines common sense, i.e. expert opinion with a
formal analytically calculated index. The resulting index provides a control limit that
depends on the number of attributes. As we observed many classifications
corresponding to this index, we concluded that it is useful and practical.
Future empirical research could consider the comparison of our index to
judgmental expert opinion and report how well it works.

References

Deming, E.W., Transformation of western style management. Interfaces, 1985, 15, 6–11.
Gleason, M., High level of operating equipment efficiency requires overall effort. Pulp Pap.,
1995, 69, 79–84.
Goldratt, E.M., The Haystack Syndrome, 1990 (North River Press: Croton on Hudson, NY).
Grosfeld-Nir, A. and Ronen, B., The complete kit concept: modeling the managerial
approach. Comput. Ind. Eng., 1998, 34, 695–701.
Hales, M.G., Focusing on 15% of the pie. Bank Marketing, 1995, 27, 29–34.
Herbert, P.R., Separate the vital few from the trivial many. Transport. Dist., 1995, 36, 84–86.
Juran, J.M., Planning for Quality, 1988 (The Free Press/Macmillan: New York).
Koch, R., The 80/20 Principle, 1998 (Doubleday: New York).
Mabin, V.J. and Balderstone, S.J., The World of the Theory of Constraints, 2000
(The St. Lucie/APICS Series on Constraints Management: Boca Raton, FL).
Ronen, B., The ‘‘complete kit concept’’. Int. J. Prod. Res., 1992, 30, 2457–2466.
Ronen, B. and Spector, Y., Managing systems constraints: a cost utilization approach. Int. J.
Prod. Res., 1992, 30, 2045–2061.
Ronen, B. and Spiegler, I., Information as inventory: a new conceptual view. Inform. Manage.,
1991, 21, 239–247.
Ronen, B., Spiegler, I. and Kozlovsky, N., The use, misuse and abuse of the Pareto principle.
Working paper, Tel Aviv University, Faculty of Management, 2005.
Tatikonda, L.U., O’Brien, D. and Tatikonda, R.J., Succeeding with 80/20. Manage. Account.,
1999, 80, 40–44.
Temme, J., Learn to manage competing priorities by focusing on what is important. Plant
Eng., 1994, 48, 80.