J. Service Science & Management, 2010, 3, 390-407
doi: 10.4236/jssm.2010.34046 Published Online December 2010 (http://www.SciRP.org/journal/jssm)
Off-Line and User-Oriented Approach for
Supplier Selection in Dynamic Environment: A
Case Study in the Healthcare Services*
Shalom Moalem1, Avi Herbon1,2, Haim Shnaiderman1, Joseph Templeman3
1Department of Management, Bar-Ilan University, Ramat-Gan, Israel; 2Department of Management and Industrial Engineering, Ariel
University Center of Samaria, Ariel, Israel; 3Joseph Templeman, The College of Business Administration, Rishon LiTzion, Israel.
Received August 19th, 2010; revised October 4th, 2010; accepted November 10th, 2010.
ABSTRACT
Evaluating and selecting supplies are critical activities in the process of purchasing and supplying materials. Many
manufacturing and service organiza tions operate in a con stantly changing business environmen t and occasionally have
to reconsider their steps in terms o f supplier selection. This paper offers a methodology that takes into accoun t the im-
pact of a dynamic business environment on the supplier selection process. This methodology represents an applicable
tool supported by decision for the planned selection of a single supplier that changes with time out of a supplier group
over a finite planning ho rizon. The suggested methodology has been tested in a large Israeli organization-Clalit Health
Services, which comprises large-scale logistical entity working with hundreds of suppliers on an ongoing basis. Our
analysis of application results shows that the suggested strategy of switching suppliers over a predefined planning ho-
rizon according to the business environment forecast is over 10% more efficient compared to a strategy that does not
change the leading supplier throughout the planning horizon. This average improvement is translated into expected
efficiency gains on most operative dimensions which represent selection parameters, such as cost per unit, supply
lead-time, reputation and more. Nevertheless, some of their value is lost due to some dimensions.
Keywords: Supplier Selection, Dyn amic Environment, Case Study, Healthcare Services
1. Introduction
1.1. Supplier Selection
Manufacturing and service organizations contract with
external suppliers for the products or services which they
market. These may be raw material, finished products or
service suppliers. Selecting the right supplier is a com-
plex decision made by companies and organizations,
with potentially significant impact on the organization’s
ongoing performance. Many studies [1,2] have indicated
that finding the right suppliers is essential for business
organizations and crucial to their success. It affects criti-
cal areas along the organization’s supply chain, such as
production, transportation, inventory, and quality. By
selecting the right suppliers, the organization can gain in
efficiency and enhance supply chain cost-effectiveness.
On the other hand, failing to select the right suppliers
could compromise the organization in economic/financial
terms or in terms of service quality and reputation. A key
tool in supplier selection is supplier ranking, which is a
central aspect of the quality management field, and is
attracting growing attention nowadays [3].
One of the key characteristics of the business envi-
ronment in which most organizations currently operate is
its dynamics. These dynamics are a significant and deci-
sive aspect of business environmental uncertainty [4,5].
Environmental dynamics are governed by several factors,
including consumer behavior, and predictable or unpre-
dictable events in the local or global business environ-
ment. Today’s customers are typically more sensitive to
quality, more exposed to competition and hence more
demanding in terms of requiring higher quality standards
and better value for money. The intense competition en-
ables them to demand quicker and more customized re-
sponse, which directly impacts the dynamics of the busi-
ness environment. Another important factor is the rapid
advances in information technology which allow for cost
reduction, shorter supply times, informational reliability
and accelerated manufacturing and procurement proc-
esses, along with more streamlined integration of opera-

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment: 391
A Case Study in the Healthcare Services
tional processes with marketing and customer service
approaches.
Financial parameters also affect the dynamics of the
business environments. For example, in a dynamic busi-
ness environment manufacturing or service organizations
contract with suppliers from many different countries,
contracts affected by exchange rate fluctuations. Another
important financial factor is changes in interest rates,
which may affect inventory maintenance costs. These
changes often affect supplier selection considerations,
since the cost of products or services provided by the
supplier is related to the interest rate.
1.2. Review of Supplier Selection Methods
The supplier selection technique literature is extensive,
including numerous studies beginning in the 1960’s [6]
for a comprehensive review of early works. These tech-
niques include using satisfaction level categories to rank
suppliers, relying mainly on the experience and skills of
the procurement manager charged with selecting a sup-
plier [7]. Other techniques discussed in the literature are
elimination [8], which specifies minimal standard scores
for each supplier selection criterion, as well as various
models of linear combinations of criteria weights used to
evaluate suppliers [9-11]. Common selection criteria are
price, quality and due-dates. The criteria specified by
Dickson [11] remain relevant to this day, although their
relative importance has changed. Another technique re-
lies on converting selection criteria to cost units based on
a pricing of various activities, to produce a single, total
cost [12]. Finally, the AHP (analytic hierarchy process)
technique enables decision makers to use comparative
quality assessments instead of qualitative indicators ba-
sed on precise weight calculations [13].
The literature also offers many mathematical pro-
gramming models, including linear programming [14],
MIP (mixed integer programming) [15], non-linear pro-
gramming [16], MOP (multi-objective programming)
[17], and AHP models [18].
Probabilistic and statistical models are also common.
Soukup [19] suggests a supplier selection methodology
requiring the decision maker to estimate the probability
of obtaining future supplier performance evaluation giv-
en certain scenarios. Liao and Rittscher [20] suggest a
genetic model designed to enhance supplier flexibility in
the supplier selection process given normally distributed
random demand and capacity constraints with the ac-
ceptable assumption that such enhancement would con-
tribute to greater flexibility throughout the supply chain.
In a case study of replacing the bus fleet of a large or-
ganization, Keles and Hartman [21] used an MIP model
and sensitivity analysis to obtain an optimal solution for
dynamic business environment changes. The objective of
this study was to determine optimal long-term timetables
for replacing the old bus fleet, with several manufactur-
ers submitting proposals.
To conclude, note that as already mentioned the lit-
erature lacks direct reference to the dynamic nature of the
business environment as a key characteristic informing
the supplier selection process.
1.3. The Need for a New Supplier Selection Me-
thodology: The Clalit Case
Supplier selection is directly affected by factors related
to the suppliers and to the nature of supply. According to
our approach, supplier selection should also be affected
by multiple exogenous factors such as demand and sup-
ply, which characterize a dynamic business environment.
Dynamic change in parameter values characterizing the
business environment represents a change in environ-
mental conditions which could make decision makers
consider replacing the present supplier or selecting fur-
ther suppliers. For example, a supplier deemed irrelevant
in the past could become relevant due to the falling ex-
change rate of a certain currency and the resulting re-
duced procurement cost. Despite the critical importance
of the supplier selection process, however, there are still
many organizations which do not conduct structured
supplier selection processes. Many organizations rely
exclusively on economic parameters when selecting their
suppliers.
The bulk of studies on supplier selection do not enable
corporate decision makers to practically deal with chan-
ges resulting from the dynamic nature of the business
environment, changes that often require rapid response.
At the same time, many organizations in Israel and
abroad operate under conditions of a constantly changing
business environment which requires them to reconsider
their decisions at a much higher rate than in the past.
The present study presents a supplier selection meth-
odology applied in retrospect to data provided by a large
Israeli HMO-Clalit Health Services (CHS). This organi-
zation comprises Supply and Procurement Administra-
tions which are among the largest in the Israeli economy,
serving some 60% of the Israeli population. As these
work with hundreds of suppliers on an ongoing basis,
making an efficient supplier selection process is highly
important to them. In view of the common approaches to
this problem which lacked sufficient reference to envi-
ronmental dynamics, however, CHS has clung to less-
than-optimal suppliers without any practical option of
replacing them, or with the ability to replace them only
by incurring a high financial penalty.
The CHS supply and Procurement Administrations

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment:
392
A Case Study in the Healthcare Services
S
E
purchase medicines and medical hardware and distributes
them to customers. The Supply Administration works
with some 350 suppliers, with about 2,000 supply points,
and with more than 5 million units of 5,000 items pro-
vided annually, for a total value of around 3.1 billion NIS
($840 m in current values).
The specific product selected for applying the sug-
gested methodology is latex gloves. Latex is a milky flu-
id derived from the rubber tree and processed to its fin-
ished form. This product is in high demand among CHS
customers. Its price is affected by environmental factors
such as sharp raw material cost fluctuations, and strug-
gles over controlling the processing industry. Any short-
age of latex glove supply could mean immediate sus-
pendsion of all medical activities in operating rooms and
laboratories. At the same time, unexpected environ-
mental changes could significantly affect CHS’s posi-
tioning vis-à-vis latex suppliers. For example, due to
environmental changes-mainly reduced raw material
supply - during 2004-2007, raw material costs have sky-
rocketed, and due to CHS’s long-term contractual com-
mitment to the supplier of this market-dependent item, it
had to pay premium prices. This meant that further busi-
ness relations with that supplier were no longer worth-
while, and that a new supplier needed to be found, which
could better meet the environmental changes. However,
lack of preparation to the predictable rise in the item’s
price made it difficult for CHS to immediately replace
the supplier, leading to high expenses and a lower service
level. We suggest that such environmental dynamics
should be factored into the organization’s decision mak-
ing already at the supplier selection stage.
1.4. Study Framework
The present study discusses a representative supplier
selection case as a model for large organizations con-
tracting with a wide variety of suppliers and operating in
a dynamic business environment. Our objective is to ap-
ply a methodology for selecting a single supplier at any
point in time over a given planning horizon under dy-
namic environmental conditions, while enhancing man-
agement indicators commonly used in supplier selection.
As presented below, supplier selection is the responsibil-
ity of a single decision maker or a single representative
of a team of decision makers in a manufacturing or ser-
vice organization interested in selecting a proposal by a
single supplier among a large number of concurrent pro-
posals, . The selection process will be
based on several evaluation criteria, used by the decision
maker to subjectively evaluate each proposal. The dy-
namic business environment is represented by several
quantitative (environmental) parameters,,
whose value could change with time.
1, 2, ,,i
1, 2,j
Part 2 describes CHS and its present supplier selection
procedure. Part 3 details the methodology for construct-
ing dynamic weights as a key element in the supplier
selection process. In Part 4 we present the results of our
case study application. Finally, our findings are discussed
in Part 5.
2. Supplier Selection at Clalit Health
Services
2.1. Background
Clalit Health Services (CHS) is the largest HMO in Israel,
providing medical services to over 3.7 million customers
through more than 1,200 clinics, 14 hospitals, some 400
pharmacies, and hundreds of institutes and labs. Its pro-
fessional staff includes thousands of physicians, nurses,
pharmacists, and paramedics.
The present study focuses on CHS’s Supply and Pro-
curement Administrations. The Supply Administration is
a key element in the organization’s logistical function, as
it is in charge of inventory management and supplying
medical institutes. The Procurement Administration’s
main responsibilities include initiating and managing
procurement contracts, constructing infrastructures and
developing management and control systems based on
market surveys, selecting potential sources and suppliers,
creating competition and taking advantage of inherent
economies of size. The fundamental concept of CHS’s
procurement processes is to secure a high-quality accu-
rate procurement channel at minimal cost and maximal
flexibility. The Procurement Administration follows sev-
eral guidelines, including compliance with medicine
prescription authorization requirements, meeting the re-
quirements of professional committees, meeting standard
safety requirements, reliability and first-rate financial
terms. Once potential sources and suppliers are identified,
the administration is required to specify the contractual
relationship with them. The tender technique is applied
based on CHS’s obligatory tender procedure, according
to which a dedicated tender committee evaluates the var-
ious suppliers in a tender procedure for procuring the
hardware in question. Committee members include a
comptroller representative, a member of the public, a
buyer, a financial executive, an economist and the head
of the relevant department.
The specific product referred to in this case study is
latex gloves (see Figure 1). Latex is a milky fluid de-
rived from certain plants and processed using various
methods to its finished form. The most important latex
ingredient is produced from the Hevea brasiliensis tree,
also called “gum tree” (see Figure 2). Many products

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment: 393
A Case Study in the Healthcare Services
used in medicine and science, as well as every- day life,
are made of or contain latex. Apart from gloves, they
include ventilation masks, operating room equip- ment,
dental medicine equipment, balloons, toys, balls and
condoms.
Latex gloves are in high demand among CHS custom-
ers. Their price is influenced by environmental factors
such as raw material cost fluctuations and struggles for
control of the processing industry. Glove shortage could
mean suspension of operating room and laboratory ac-
tivities. In 2002, a supply contract was signed with one
of this product’s suppliers based on contemporary de-
mand and consumption forecasts. The contract period
was five years, without an option for reevaluation based,
among other things, on changes in the business environ-
ment.
The following years saw regulatory changes in the
producing countries, such that some of them banned all
latex exports to the west. These environmental changes
meant lower raw material supply and made continuous
production of latex gloves more difficult, leading to sky-
rocketing costs. The international market price of 100g of
latex was $0.91 in early 2004, rising to $1.4 in early
2007. Following these changes, the glove supplier de-
cided to raise the unit price during the contract period,
without any negotiation with CHS representatives, taking
advantage of the rigid contract.
It was such extreme environmental fluctuations, par-
ticularly those which lead to the unilateral steps taken by
the supplier of this product, which motivated our choice
of latex gloves to demonstrate the application of our
proposed methodology. When examining the decision
making process leading to this supplier’s selection in late
Figure 1. Latex gloves.
Figure 2. Latex ra w material.
2002, we will also look into the option of changing sup-
pliers in the future based on forecasts of predictable en-
vironmental changes.
2.2. Supplier Proposals
In 2002, the CHS Procurement Administration began a
process of selecting latex glove suppliers. The product is
particularly in demand in the medical community, with
monthly consumption of some 1m units. An RFP was
sent to four suppliers (S = 4), designated hereafter as S1,
S2, S3 and S4. The price offers were requested per a 100
glove unit. The suppliers were also requested to meet
various selection parameters according to CHS require-
ments, based on Procurement Administration policy.
These commonly used criteria included cost, quality,
supply rate and flexibility [22]. In turn, the suppliers’
proposals were translated to selection parameter values.
Appendix 1 presents these selection parameters and
their values, based on the suppliers’ proposals. The Rep-
utation parameter refers to historical supplier perform-
ance data. In the absence of such data, the supplier in
question will be evaluated without any consideration of
this parameter. The Credit Rating parameter refers to the
supplier’s economic strength and financial stability, and
is measured using a 1-5 Likert scale based on financial
data required of the supplier. Geographic Location refers
to the distance, in kilometers, between the suppliers’ and
the customers’ warehouses. Finally, the Payment Dead-
line refers to the latest date following the end of a current
work month in which the supplier demands a financial
return for the supplies delivered during that month.
After specifying the customer’s requirements, the sup-
pliers’ proposals were translated into measurable data by
composing a supply criteria list, ,
(Set 1), in which each criterion is made up of several
relevant parameter indicators, ,.
The supplier criteria list was specified in a similar way
, and designated Set 2, with each
criterion made up of relevant parameters, ,
. S2 proposed a discount (for the whole
amount), based on the number of units purchased in each
order, as illustrated in Table 1 below.
1
1
1, 2,kQ
11
,kn 1, 2
p
1
1
k
q
1
11
k
nP
22
,kn
1
p
2
2
k
q

2
2
1, 2,k
2
22
1, 2k nP
Q
2
Table 1. S2 Unit cost based on order size.
Order Size Cost per 100 Units (NIS)
100,000 -3,000,000 18
3,000,001 -9,000,000 15
9,000,001 -15,000,000 12

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment:
394
A Case Study in the Healthcare Services
2.3. Supplier Proposals
After receiving the supplier proposals and converting
them into selection parameter values, CHS procurement
executives had to select the supplier. The selection proc-
ess included several stages. At first, each set of criteria
received a relative weighting based on the decision mak-
ers’ subjective preferences and as suggested by several
studies [3]. A weight of 0.4 was given to the Supply cri-
teria set, and a 0.6 weight was given to the Supplier cri-
teria set. Following that, a relative weight was given to
each specific criterion based on the decision makers’
perspective and their experience. Next, the decision
makers were required to weigh the parameters relevant to
supplier selection, based on their criterion weighting.
Appendix 3 presents the weights as determined by CHS.
The selection parameter weights were determined in
several phases. First, three importance levels were speci-
fied: high, medium and low. The weights were then de-
termined based on the importance levels and the level of
score variance in the various proposals per each selection
parameter. For example, high importance combined with
high variance in the supplier standard scores for the same
parameter meant a high weight, while high importance
combined with low variance meant a medium weight.
Similarly, low importance combined with high variance
meant a medium weight, while low importance combined
with low variance meant a low weight. Whenever a cer-
tain criterion was represented by a single selection pa-
rameter, this parameter received the full weight of the
criterion it represented.
The decision makers were then required to rate the
various suppliers on each selection parameter. This rating
was based on a specification of score scales for each pa-
rameter, ,
g
g
g
kn
p

,, defined by a lower and upper
boundary for each selection parameter, and
,, respectively. The various scale ratings were
determined by using two approaches based on parameter
type. One was the 1-5 Likert scale (1 representing the
lowest value), and the other was binary. A rating per
each parameter, ,
1, 2g

,
gg
g
kn
Lp


gg
g
kn
Up
g
g
g
kn
d,, in each of the criteria
group was computed using a transformation function
substituting the selection parameter raw value
1, 2g
,
g
g
g
kn
p

g
by
the rating value , such that

,
gg gg
kn kn
dTransp,
g




,,
gg gg
gg
kn kn
dLp



,
,gg
g
kn
Up



. Appendix 2 presents the
value transformation of all selection parameters ,
g
g
g
kn
p
into the rating scores ,
g
g
g
kn
d. For each supplier pro-
posal , the raw score
1, 2,3, 4i,
g
g
g
kn
d was transformed
into a standard score ,
g
g
g
xkn using a uniform 0-100 scale
for each criteria set g = 1,2, with 100 being the highest
value for each parameter. Supplier proposal standard
score sums for all selection parameters based on the
Likert scaling approach following the transformation are
presented in Appendix 1.
The binary approach was used to evaluate one re-
maining parameter: Quantity Discount. If the supplier
offers such discount, he will receive the optimal 100
score. If not, the score will be 0. Table 2 presents sup-
plier data for this parameter.
Since S2 had no historical data that could be used to
specify the Reputation parameter, he received a 59.78
score on this parameter, equal to the proportional weighted
average score of all the other selection parameters.
After specifying the weights for each parameter and
determining the standard scores for supplier proposal, all
that remained was to calculate the total weighted score
for each supplier proposal. This was done by multiplying
the relevant weight by the score, as suggested by the lit-
erature surveyed above and as prevalent in practice. The
total score ifor each i supplier proposal, X1, 2,3, 4i

,
was given by the following:


11
22
ii
kn
kn
px


1
1
1
1 1
11
2
2
2
22
22
11
,,
11
22
,,
11
, 1,2,3,4
k
k
P
Q
kn
kn
P
Q
ikn
kn
Xpxp
p i








(1)
such that


,
gg
g
kn
p  is the weight of each parameter
,
g
g
g
kn
p, and ,
g
g
g
kn
x is the standard score, 1, 2
g
g
g
k
nP
1, 2g

. Using this formula for the various suppliers
produced the following final scores: 162,S2 59.78,S

366, 469SS

. On the basis of these scores S4 was
selected for the 2003-2007 contract.
3. Constructing Dynamic Selection Weights
3.1. Proposed Methodology Principles
As already mentioned, one of the key characteristics of
modern business environments is their dynamics. Supplier
selection must take this factor into account. Below, we
propose a retrospective qualitative and quantitative de-
scription of environmental influences on supplier se lec-
tion considerations. No less important, we propose how to
present them visually to the decision makers, so as to en-
able them to actually “see” the selected supplier, as well as
predictable changes in the business environment.
First, the proposed methodology focuses on selecting a
supplier under initial environmental conditions, as de-
cribed in Part 2, on the basis of characterizing customer s

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment:
A Case Study in the Healthcare Services
395
Table 2. Standard scores for the quantity discount parameter


2
3,1
p.
Supplier S1 S2 S3 S4
Standard score 0 100 0 0
Selection parameter
quantity discount

2
3,1
p
Supplier proposal None V None None
requirements, specifying selection criteria, subjectively
weighting them in a way that reflects the decision mak-
ers’ preferences and finally, scoring these criteria. We
then present a supplier selection process based on selec-
tion criteria under dynamic environmental conditions.
The selection criteria specified (see Appendix 1) are
based on the supplier selection literature, mainly Dickson
[11]. Using these selection parameters, we can analyze
the supplier selection process in the organization studied,
as it is affected by its dynamic business environment. In
general, the selection criteria would be quantitative, but
they may also be qualitative as long as measurable selec-
tion parameters could be tailored for them.
3.2. Dynamic Weight Functions
As already described, following environmental changes –
specifically, rising raw material prices-CHS’s latex glove
supplier raised its price subject to the contract. Following
that move, CHS found itself tied to a less-than-optimal
supplier, and in retrospect, realized that it should have
been able to select a new supplier able to meet such price
fluctuations. Following the changes in the business envi-
ronment, the original weights and scores given to the
four latex glove suppliers were no longer appropriate.
The decision assessed in retrospect in this case study
refers to the following question: what is the future date in
which it would no longer be worthwhile to continue
working with the present supplier? In order to specify a
dynamic environment that would expand the scope of
supplier selection beyond the rise in raw material prices
to factor in additional environmental influences, we have
identified relevant dynamic environmental parameters
j
p, using research tools [23], as well as an
internal questionnaire distributed among CHS’s pro-
curement executives. Table 3 presents these parameters
and their initial values.
1, 2, 3,4, 5j
In order to represent various environmental influences
using dynamic environment parameters we suggest con-
structing dynamic weight functions. Moreover, in order
to enable decision makers to construct relatively simple
weight functions, we propose an initial qualitative pres-
entation of the directions of environmental parameter
influences on each selection parameter. Table 4 presents
the direction in which each environmental parameter is
expected to influence the selection parameter weight ac-
cording to the decision makers’ perspective. For each
selection parameter, the table presents changes which
would make it more difficult to conduct business in the
given environment. These changes could cause certain
selection parameters to “gain weight”, and others to “lose
weight”. For example, a sharp increase in the Demand
parameter, for example, would require suppliers to sup-
ply greater quantities, thus casing the selection parameter
Max per Shipment to gain weight. Similarly, an in-
creased Interest Rate would cause Inventory Mainte-
nance cost to gain weight; reduced Supply Time affects
the suppliers’ ability to prepare for the next shipment,
thus affecting some of the selection parameters, such as
Geographic Location; finally, lower Raw Materials Sup-
ply would cause the supplier to increase the Cost per
Unit, so that this parameter would gain weight.
The thin arrows in Table 4 () indicate the desirable
direction of response, according the decision makers’
subjective interpretation of changes in the selection pa-
rameter weights following the negative environmental
changes. Whenever a CHS decision maker can be ex-
pected to be indifferent to the parameter weight change,
no arrow appears. Therefore empty cells represent indif
ference by the decision makers regarding the required
response to the change in the environmental parameter in
question.
The thick arrows in Table 4 indicate worsening
in environmental parameter values: higher Demand,
higher Interest Rate, lower Supply Time, lower Raw
Material Supply and higher Exchange Rate.


To illustrate the logic filling out the table shown above,
we will refer to the Min per Shipment selection parame-
ter. When demand is rising, this parameter would lose
weight, since under the new conditions, larger shipments
are desirable. This parameter would gain weight as the
Table 3. Business environment parameters: Initial 2002.
Environmental
Parameter
Annual
Demand (d) Interest Rate (r)Supply Time (DD)
Raw Material
Supply (nn)
Nis-to-Dollar Exchange
Rate (cur)
Initial Value 10 0.1 5 100 4.8

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment:
396
A Case Study in the Healthcare Services
Table 4. Worsening environmental parameter values and their effect on parameter weights.
Environmental Influence/Selection Parameter Demand

Interest Rate

Supply Time

Raw Material Supply

Exchange Rate

Minper Shipment 1
1,1
p




Maxper Shipment 1
1,2
p

  
Unit Cost 1
2,1
p




Expected Defect Rate 1
3,1
p


Suggested Supply Time 2
1,1
p



Credit Rating 2
1,2
p



Reputation 2
1,3
p



Order Cost 2
2,1
p



Quantity Discount 2
3,1
p

  
Geographic Location 2
3,2
p


Payment Date 2
3,3
p



Minper Order 2
4,1
p



Maxper Order 2
4,2
p


interest rate (as it affects inventory maintenance costs)
increase. The reason for that is that due to the higher in-
terest rate, the organization would rather keep inventories
to a minimum, and attach a higher weight to the minimal
supplied quantities, assuming they are proportional to
order size. When environmental conditions cause supply
time to shrink, this parameter’s weight would increase
again, since given shorter supply time it is important to
make sure that at least a minimal amount of the product
would arrive in each shipment, or that working with
smaller shipments is made possible. A rising exchange
rate would also increase the weight of this parameter,
since when product cost is high the organization would
require smaller quantities in each shipment, hence the
greater importance attached to the supplier’s ability to
supply minimal quantities. As opposed to all these clear
outcomes, reduced raw material supply would have an
ambiguous effect: on the one hand, it could cause item
costs to rise, reducing the Min per Shipment parameter
weight as the organization would want to keep a minimal
inventory as product costs rise; on the other hand, re-
duced supply could cause buyers to increase their de-
mand as they fear future shortage, leading to the opposite
effect. CHS decision makers would therefore remain
indifferent to this environmental change.
After having specified the directions of environmental
influences, we now need to specify a more detailed func-
tional structure for calculating selection parameter
weighs as they are affected by these influences. In order
to construct more accurate functions and adjust them to
the directions presented above, we have studied the effect
of such environmental changes using a sensitivity analy-
sis in a decision maker trial and error approach by
changing the present environmental values in retrospect
and assessing their impact on the desirable weight of
each selection parameter. To simplify application without
loss of generality we have chosen two dynamic weight
functional structures. One has a linear product coefficient
yaxb

, such that a,b are parameters subjectively
determined to obtain the function’s specific configuration.
In functions of this type, a dynamic product coefficient
of a selection parameter weight with general indexes for
some dynamic environmental parameter describ-
ing an influence on a certain selection parameter would
dym
p
be

,
1
dym dym
dym cur
b
f
p
p

pb



 , such that b represents
the decision maker’s subjective sensitivity to the degree
of change in the selection parameter weight relative to
the change in the external parameter , while
is the present value of . The second type
are dynamic weight functions with an exponential struc-
ture,
dym
p
,dym cur
pdym p
x
y
e

 , such that ,

 are subjectively deter-
mined parameters for obtaining the function’s specific
configuration. In functions of this type, the dynamic
product coefficient

,1
dym
dym cur
p
dym p
fp 

 of the selec-
tion parameter weight, for a certain environmental pa-
rameter describes an influence on a given selec-
tion parameter, such that parameter
dym
p

represents the

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment: 397
A Case Study in the Healthcare Services
decision maker’s subjective sensitivity to the degree of
change in the selection parameter weight relative to the
change in the external parameter .
dym
p
The choice of a certain functional structure will be
subject to the given selection parameter weight’s sensi-
tivity to changes in the business environment over a spe-
cific horizon, based on the decision makers’ final deter-
mination. An exponential structure would enable to rep-
resent a varying sensitivity of selection parameter weight
as dependent on environmental parameter values.
In order to calculate the dynamic weight of a certain se-
lection parameter as a functional dependency on environ-
mental parameter values, we must first multiply the prod-
uct coefficients of all the various environmental parame-
ters by the selection parameter’s initial weight. The as-
sumption behind this coefficient multiplication approach is
mutual independence of the selected parameter values
representing environmental changes. This independence
guides their selection even if it is unnecessary for applica-
tion purposes. When the decision maker is indifferent, the
product coefficient value is set to 1. A schematic repre-
sentation of a linear dynamic product coefficient structure
is presented in Figure 3, with the example of Min per
Shipment as it changes relative to Demand.
In order to represent the desirable directions of chang-
ing selection parameter weights in response to changes in
parameter values (Table 4), CHS decision makers have
specified subjective parameter values to obtained de-
tailed descriptions of linear or exponential dependence of
each selection parameter on environmental parameters.
These subjective parameters are presented in Appendix 4.
This specification is based on a graphic representation of
influences as shown above, and on a calibration of sub-
jective parameter values using visual trial and error, as
represented in the example in Figure 3. Specifying a
while setting a subjective parameter means a linear b
Figure 3. Dynamic product coefficient of min per shipment
as a function of annual demand.
subjective parameter

means exponential influence,
influence. An empty cell in the table means the decision
maker is indifferent with regard to the desirable response.
Equation (2) is an example of a dynamic weight func-
tion for selection parameters as dependent on environ-
mental parameters. The dynamic weight function for Min
per Shipment, prior to normalization calculation, is given
by

1
1,,,,1(2) 13
0
0.1 1.723
1,1 00
1
0
0.04
d
diDDcur d
fi
iDD
Cur
Cur
  


 DD

 



 

 








(2)
In itself, the requirement that the total dynamic selec-
tion parameter weights for all parameter values adding
up to 1 does not guarantee that all requirements for pa-
rameter weight change trends as represented in Table 5
will in fact be met. One reason is the differences in the
specific intensities of change sensitivity as determined by
CHS decision makers for the weight of each selection
parameter relative to each environmental parameter. An-
other is the possible existence of requirements to enable
opposing trends. The weights obtained by the dynamic
weight functions for the various selection parameters
must be normalized. The normalization requirement is
such that the total weights of all selection parameters add
up to 1 per each future environmental parameter vector
value. Table 5 presents selection parameter weights as a
result of business environment changes versus the initial
selection parameter weights, in reference to a 5% wors-
ening in all environmental parameters. The environ-
mental parameter change directions are presented in Ta-
ble 5.
3.3. Dynamic Weight Functions
In order for the CHS decision makers to be able to easily
select the supplier given changes in their business envi-
ronment, we have plotted each environmental parameter
on a two-dimensional graph, with the horizontal axis
representing the environmental parameter value and the
vertical axis represented the selected supplier index, such
that all other environment values remain constant at their
current values (given in Table 4).
Looking at Figure 4, we can see that S3 becomes
dominant when the required supply time shrinks. This is
because this supplier excels in relevant selection pa-
rameters such as Reputation and Supply Time which gain
weight the more Supply Time shrinks. High scores in
these parameters, together with their increased weight,
raise S3’s general score. Similarly, Figure 5 shows that
when Supply falls, S3 becomes dominant as it excels in

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment:
A Case Study in the Healthcare Services
398
Table 5. Selection parameter weights: current vs. dynamic environment (5% worse).
Selection Parameter Current Parameter Weight Dynamic Parameter Weight
Minper Shipment 1
1,1
p
0.1 0.07
Maxper Shopment 1
1,2
p
0.07 0.03
Unit Cost 1
2,1
p
0.15 0.04
Expected Defect Rate 1
3,1
p
0.08 0.04
Suggested Supply Time 2
1,1
p
0.1 0.11
Credit Rating 2
1,2
p
0.06 0.05
Reputation 2
1,3
p
0.07 0.09
Order Cost 2
2,1
p
0.1 0.06
Quantity Siscount 2
3,1
p
0.03 0.01
Geographic Location 2
3,2
p
0.03 0.03
Payment Date 2
3,3
p
0.04 0.06
Minper Order 2
4,1
p
0.07 0.04
Maxper Order 2
4,2
p
0.1 0.04
relevant selection parameters such as Reputation and
Supply Time which gain in weight as Supply falls. High
scores in these parameters, together with their increased
weight, raise S3’s general score. Finally, we can see that
a highly significant fall in Supply may lead to selecting
S1, which excels in the Supply Time and Geographic
Location selection parameters which gain in weight as
Supply falls.
plier index. The figure shows that S1 has now become
almost irrelevant, although Figure 5 has shown it to be
potentially relevant following a significant reduction in
supply. This shows that the previous conclusion is inap-
plicable once demand changes as well. This information
would lead decision makers to assume that S1 would
probably not be relevant in the future, both due to the
high probability of demand changes and due to the trend
of reduced supply time.
Figure 6 presents a three-dimensional graph illustrat-
ing the selected supplier’s identity as a function of
changes in two environmental parameters – Demand and
Raw Material Supply-with the others held constant. The
horizontal axes represent two environmental parameters
changing concurrently, and the vertical axis is the sup-
In order to identify the best supplier in different points
in time, as a function of the previously specified environ-
mental parameters,


12
,,
E
pp p

, we can use the term


 E
p
1
arg max,
i
iXp, where
Figure 4. Selected supplier as a function of supply time change (DD).

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment: 399
A Case Study in the Healthcare Services
Figure 5. Selected supplier as a function of saw material suppl y (r m).


1
1
1
11
2,
1,
111
,
, 1,2,3,4
k
gg gg
P
Qgdyn g
iE i
kn kn
gkn
Xp ppxp
i







,

E
(3)
and

1,... i
X
pp

1,... iE
is a dynamic score for proposal i. The
X
pp
 1
,
gg
g
kn
score is composed of the standard scores
in proposal i for the selection parameter
i
xp
,
g
g
kn
g
p and dynamic weights

,
,
gg

g
dyn
kn
p  , normalized as
dynamically dependent on the environmental parameters,
according to the following:




1
1
1
11
,
,
,2
,
111
,,,,
,,, ,
, 1,2, 1,2,1,2...
gg
gg
k
gg
g
g
kn
gdyn
kn P
Qg
kn
gkn
gg
gk
fdrDDrmcur
p
g
f
drDDrmcur
gkQn P



 


(3)
weight coefficient functions and are presented in Appen-
dix 4. This formula dictates the supplier scores according
to which the most desirable supplier is identified – the
one with the highest score at each potential environ-
mental point.
4. Strategy for Replacing Suppliers over the
Planning Horizon
As already described, in 2002 the CHS Procurement
Administration began evaluating potential latex glove
suppliers. CHS chose one out of four candidates for a
long-term contract. Hereafter, this strategy will be called
the Basic Strategy. According to the Basic Strategy, the
fourth supplier (S4) has been selected on the basis of its
scores at the beginning of the planning horizon and is not
replaced throughout it. Conversely, in our proposed me-
thodology the procurement executives would have relied
on CHS’s periodic future planning information system to
apply a strategy that takes into account business envi-
ronment forecasts over the planning period to predict
future points in time in which the most desirable supplier
would change. The retrospective forecast presented in
Figures 7-11 enables CHS to formulate a strategy for
replacing suppliers following environmental changes
over the 5-year planning horizon. Regrettably, in reality
CHS continued to adopt the Basic Strategy of working
with a single supplier throughout the planning period,
regardless of environmental changes.
Figure 6. Selected supplier as a function of changing de-
mand and raw material supply.

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment:
400
A Case Study in the Healthcare Services
Figure 7. 2003-2007 demand forecast vs. realization.
Figure 8. 2003-2007 interest rate forecast vs. realization.
Figure 9. 2003-2007 supply time forecast vs. realization.

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment: 401
A Case Study in the Healthcare Services
Figure 10. 2003-2007 raw material supply forecast vs. realization.
Figure 11. 2003-2007 exchange rate forecast vs. realization.
Figures 7-11 present the environmental parameter
forecast for the entire planning period together with the
parameter values as observed in reality, for the same
planning period (2003-2007), with Year 1 referring to
2003, and so on.
The nature of competing strategies is evaluated based
on actual environmental changes throughout the plan-
ning horizon. The strategy suggested here can provide
decision makers with a series of future suppliers already
at the beginning of the planning horizon. Figure 12 pre-
sents the predictable environmental forecasting strategy
for the winning supplier over the planning horizon. The
winning supplier’s scores were evaluated every quarter
for the five-year period. One can see how this strategy
requires suppliers to be changed in midstream. Note that
S2 would become preferable near the end of the plan-
ning horizon. This can be explained by the fact that the
forecast predicts rising demand. A change of this type
could lead to significantly higher weighting of the Unit
Cost parameter, in which S2 is much preferred over the
others.
In order to fully appreciate the appropriateness of the
decision to remain with a single supplier, compared to
the retrospective choice of adopting a strategy that allows
for supplier replacement in response to predictable envi-
ronmental changes, we have specified a utility value per
time unit based on the selected supplier score at that
point in time. This utility indicator is based on actual
environmental data for the period in question. The com-
parison is between the Predictable Environment Strategy
and the Basic Strategy, in reference to actual environ-
mental parameter values. As before, we calculate the
winning supplier scores given the two strategies on a
per-quarter basis. The area under each curve in Figure
13 is defined as the utility value for the respective strat-
eg over the planning horizons. y

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment:
402
A Case Study in the Healthcare Services
Figure 12. Predictable environment strategy.
Figure 13. Utility values obtained by the predictable environment vs. basic strategy.
A comparison of the respective strategy utility val-
ues as presented in Figure 13 clearly shows that the
Predictable Environment Strategy is preferable, and
that in retrospect, CHS should have adopted this sup-
plier replacement strategy. The difference between the
utility values is around 20% in favor of the Predictable
Environment Strategy. This aggregate improvement
has been converted in Table 6 into several common
operational indicators. The table shows that retrospec-
tive adoption of the proposed strategy would have en-
sured superior values of most indicators for the case
under study.
The indicators presented in Table 6 have been calcu-
lated based on a raw average of each indicator over the
planning horizon. Analysis of the utility data in this table
demonstrates a significant economic efficiency gain, in-
cluding savings of about 1,300,000 NIS (about $350m in
current values) in Direct Purchasing Cost. This im-
provement is due to the fact that in the Basic Strategy,
the selected supplier is S4, which charges the highest
Cost per Unit, while the Predictable Environment Strat-
egy prioritizes S4 only for about one fifth of the planning
horizon. The environmental forecast points to a trend of
continual gain in the Cost per Unit parameter, dictating a
strategy of preferring suppliers excelling in this parame-
ter as time goes by. In retrospect, the actual environ-
mental parameter values moved closer to the predictable
environment parameter values, mainly for the Demand
parameter, which significantly affects the Cost per Unit
parameter weight.
Further aspects of efficiency gain can be identified in
other important parameters. This improvement is ex-
pressed in quantitative indicators, such as 14% shorter
upplier range (Distance from Supplier), and 8% shorter s

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment: 403
A Case Study in the Healthcare Services
Table 6. Operational indicator changes following retrospective adoption of the predictable environment strategy.
Indicator Basic Strategy
Predictable Envionment
Strategy Improvement
Direct Purchasing Cost 10,800,000 NIS 9,459,000 NIS 12%
Supply Time 3 days 2.76 days 8%
Distance from Supplier 120 km 103km 14%
Quantity Discount 0 0.28 28%
Payment Rating Cutrent+75 days Current+103 days 37%
Credit Rating 2 2.52 26%
Requtation 40 77 92%
Defective Rate PER Shipment 0.4% 0.53% -25%
Supply Time, as well as in qualitative indicators such as
Credit Rating and Reputation. We can also see that there
may be some indicators in which the Basic Strategy is
retrospectively advantageous (e.g., Defective Rate per
Shipment). S4’s proposal in this latter parameter is the
best among the four, so that not selecting it retrospec-
tively had a negative effect. Although the predictable
dynamics increased the weight of this parameter in rele-
vant environmental scenarios, but this effect was rela-
tively weak as it was felt in only two out of five envi-
ronmental parameters (shorter Supply Time and smaller
Raw Material Supply). In retrospect, the environmental
forecast proved wrong, that is, it failed to approach the
actual environment for the Raw Material Supply pa-
rameter, which had the main negative effect in terms of a
0.13% absolute increase in the Defective Rate per Ship-
ment parameter.
5. Discussion
In this article, we have examined a new executive tool
for supporting supplier selection, with particular refer-
ence to the dynamic nature of the organization’s business
environment. The methodology suggested here is based
on a qualitative and quantitative description of environ-
mental influences affecting present and future supplier
selection consideration, and includes coherent visual re-
presentation of such effects to organizational decision
makers. The methodology suggested was studied using a
case study in a leading Israeli HMO, Clalit Health Ser-
vices, which comprises large-scale logistical entities. In
evaluating the methodology in this organization, we have
assessed business environmental changes in retrospect,
and particularly changes in such parameters which would
have made it worthwhile for the organization to seek
additional or alternative suppliers, or change the market
shares of existing ones.
The proposed decision making process in Clalit Health
Services has been accepted by its experts, based on a
presentation of quantitative data in the application stages.
Our case analysis shows that had the organization ap-
plied this methodology in retrospect, this would have
improved most operational indicators by an average of
20%. Nevertheless, the process does not guarantee that
all indicators would improve, and some might even lose
value.
An important key to the proposed methodology is de-
termining a future series of suppliers, right at the begin-
ning of the planning horizon. Admittedly, in some cases
a policy of not replacing the selected supplier would be
preferable. This is particularly true when the actual envi-
ronment, as it develops over time, is significantly differ-
ent from the one predicted in advance. Nevertheless, in
such a case it is always possible to change the supplier
replacement plan suggested by the proposed policy and
adjust it to actual developments by updating the forecast
on an ongoing basis. The resulting loss to the organiza-
tion is expected to be smaller, on average, than the loss
expected using the proposed strategy in this case.
One of the proposed methodology’s assumptions is
that existing supplier proposals, or at least the relative
differences between them on all selection parameters,
would be maintained throughout the planning horizon.
Should that not be the case, we suggest updating the
proposal scores from time to time, and changing prefer-
ences accordingly.
Our findings lead to the key conclusion that our me-
thodology – the Predicted Environment Strategy – is
preferable to the existing one applied by the organization
over a broad range of operational indicators, and in terms
of average total improvement. Despite the potential for

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment:
404
A Case Study in the Healthcare Services
non-negligible environmental prediction errors, in retro-
spect the proposed strategy’s effectiveness has been
shown to be maintained over a five-year planning period.
The proposed methodology may be applied to many
other manufacturing and service organization, leading to
economic and organizational efficiency gains. In all such
organizations, decision makers can evaluate environ-
mental parameters that are liable to change, leading to
possible changes in the relative importance attached to
the various supplier selection parameters. Moreover, the
decision supporting executive tool suggested here takes
the subjective element of decision making into account
both by specifying environmental parameters affecting
selection parameter weights and by specifying the weight
functional structure (linear, exponential) affecting the
degree of change. Consequently, this tool may quantita-
tively represent qualitative decision maker perspectives.
Thanks to its simplicity, the methodology’s applicability
is maintained (apart from size proportions) also in sys-
tems that are characterized by a broad range of selection
parameters liable to be affected by multiple business en-
vironmental parameters.
The proposed methodology is limited, however, when
it comes to the application stage, where decision makers
are required to subjectively determine parameters affect-
ing the dynamic weight functions’ configuration. It is
possible to deal with this difficulty by using techniques
allowing the identification of relevant parameters relying
on partial data.
Another application limitation is the difficulty of se-
lecting suppliers for a short term. For administrative
reasons, or simply due to the bureaucratic difficulty of
replacing suppliers over the short term, the proposed
strategy might be applied only approximately and
probably for long periods of time (years), as is custom-
ary. Moreover, the suggested policy could imply sup-
plier replacement times in the order of days, or even
hours. Although this constitutes no theoretical difficulty,
in practice, supplier replacement dates would have to be
adjusted to more manageable units, such as months or
years.
One promising avenue of future research is to ex-
pand the validity basis of the proposed methodology
by applying it to other organizations or using statisti-
cal simulation experiments. Another direction would
be to enhance the subjective element in the abovemen-
tioned weight functional parameter selection and
making it user-oriented on the methodological level.
Developing an effective built-in trial-and-error me-
chanism could allow decision makers to better under-
stand the significance of selecting certain parameters,
and arrive at a more subjectively balanced parameter
combination.
REFERENCES
[1] K. S. Bhutta and F. Huq, “Supplier Selection Problem: A
Comparison of the Total Cost of Ownership and Analytic
Hierarchy Process,” Journal of Supply Chain Manage-
ment, Vol. 7, No. 3, 2002, pp. 126-135.
[2] V. Jayaraman, R. Srivastava and W.C. Benton, “Supplier
Selection and Order Quantity Allocation: A Comprehen-
sive Model,” Journal of Supply Chain Management, Vol.
35, No. 2, 1999, pp. 50-58.
[3] N. D. Singpurwalla, “A Probabilistic Hierarchical Classi-
fication Model for Rating Suppliers,” Journal of Quality
Technology, Vol. 31, No. 4, 1999, pp. 444-454.
[4] L. J. Bourgeois, “Strategy and Environment: A Conce-
ptual Integration,” The Academy of Management Review,
Vol. 5, No. 1, 1980, pp. 25-39.
[5] R. B. Duncan, “Characteristics of Organizational Envir-
onments and Perceived Environmental Uncertainty,”
Administrative Science Quarterly, Vol. 17, No. 3, 1972,
pp. 313-327.
[6] C. A. Weber, J. R. Current and W. C. Benton, “Vendor
Selection Criteria and Methods,” European Journal of
Operations Research, Vol. 50, 1991, pp. 2-18.
[7] E. Timmerman, “An Approach to Supplier Performance
Evaluation,” Journal of Purchasing and Materials Man-
agement, Vol. 22, No. 4, 1986, pp. 2-8.
[8] L. E. Crow, R. W. Olshavsky and J. O. Summers, “Indus-
trial Buyers’ Choice Strategies: A Protocol Analysis,”
Journal of Marketing Research, Vol. 17, No. 1, 1980, pp.
34-44.
[9] C. C. Li, Y. P. Fun and J. S. Hung, “A New Measure for
Supplier Performance Evaluation,” Institute of Industrial
Engineering, Vol. 29, No. 9, 1997, pp. 753-758.
[10] K. N. Thompson, “Scaling Evaluative Criteria and Sup-
plier Performance Estimates in Weighted Point Pre-Pur-
chase Decision Models,” International Journal of Pur-
chasing and Materials Management, Vol. 27, No. 1, 1991,
pp. 27-36.
[11] G. W. Dickson, “An Analysis of Supplier Selection: Sys-
tems and Decisions,” Journal of Purchasing, Vol. 2, No.
1, 1966, pp. 5-17.
[12] R. M. Monczka and S. H. Trecha, “Cost-Based Supplier
Performance Evaluation,” Journal of Purchasing and
Materials Management, Vol. 24, No. 1, 1988, pp. 2-8.
[13] C. Muralidharan, N. Anatharaman and S. G. Deshmukh,
“A Multi-Criteria Group Decision Making Model for
Supplier Rating,” Journal of Supply Chain Management,
Vol. 38, No. 4, 2002, pp. 22-33.
[14] A. C. Pan, “Allocation of Order Quantity among Suppli-
ers,” Journal of Purchasing and Material Management,
Vol. 25, No. 3, 1989, pp. 36-39.
[15] R. G. Kasilingam, “Selection of Vendors – a Mixed Inte-

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment:
A Case Study in the Healthcare Services
405
ger Programming Approach,” Computers and Industrial
Engineering, Vol. 22, No. 1-2, 1996, pp. 347-350.
[16] J. D. Hong and C. Hayya, “Just-in-Time Purchasing: Sin-
gle or Multiple Sourcing?” International Journal of Pro-
duction Economics, Vol. 27, No. 2, 1992, pp. 171-181.
[17] C. A. Weber and L. M. Ellarm, “Supplier Selection Using
Multi-Objective Programming: A Decision Support Sys-
tem Approach,” International Journal of Physical Dis-
tribution and Logistics Management, Vol. 23, No. 2, 1993,
pp. 3-14.
[18] G. Wang, S. H. Huang and J. P. Dismukes, “Product Dri-
ven Supply Chain Selection Using Integrated Multi-Cri-
teria Decision Making Methodology,” International Jour-
nal of Production Economics, Vol. 91, No. 2, 2004, pp.
1-15.
[19] W. Soukup, “Supplier Selection Strategies,” Journal of
Purchasing and Materials Management, Vol. 23, No. 2,
1987, pp. 7-12.
[20] Z. Liao and J. Rittscher, “A Multi-Objective Supplier
Selection Model under Stochastic Demand Conditions,”
International Journal of Production Economics, Vol. 105,
No. 1, 2008, pp. 150-159.
[21] P. Keles and J. C. Hartman, “Case Study: Bus Fleet Re-
placement,” The Engineering Economist, Vol. 49, No. 3,
2004, pp. 253-278.
[22] R. Verma and M. Pullman, “An Analysis of the Supplier
Selection Process,” International Journal of Management
Science, Vol. 26, No. 6, 1998, pp. 739-750.
[23] R. A. D’Aveni and D. J. Ravenscraft, “Economies of
Integration versus Bureaucracy Costs: Does Vertical In-
tegration Improve Performance?” The Academy of Man-
agement Journal, Vol. 37, No. 5, 1994, pp. 1167-1206.
Appendix 1. Supplier proposal and their total standard scores.
Supplier 1S 2S 3S 4S
Selection Parameter
Proposal 10,000 10,000 5,000 1,500
Min per Shipment 1
1,1
p
St. Score 40 40 80 100
500,000 1,000,0001,500,000 2, 000, 000
Proposal
1
1,2
p
St. Score 20 40 60 80 Max per Shipment
0.17 0.18 0.15 0.12 Proposal
Unit Cost (NIS) 1
2,1
p
St. Score 80 100 60 40
Defective Rate per
Shipment
0.8 0.7 0.5 0.4 Proposal
1
3,1
p
St. Score 40 60 80 80
Supply Time (days) Proposal 1 4 2 3
2
1,1
p
St. Score 100 40 80 60
Proposal 3 2 3 2
2
1,2
p
Credit Rating St.
Score 60 40 60 40
Very high perform-
ance Low performance
Proposal High performance N/A
2
1,3
p
Reputation
St. Score 80 59.78 100 40
Cost per Order (NIS) Proposal 700 780 850 615
2
2,1
p
St.
Score 80 60 40 80
Distance from Sup-
plier
Proposal 20 60 120 120
2
3,2
p
St. Score 100 80 60 60
Payment Date (Cur-
rent+)
75 120 60 90 Proposal
2
3,3
p
St. Score 40 60 60 80
50,000 100,000 30,000 80,000
Proposal
2
4,1
p
Min per Order St.
Score 80 40 80 60
10,000,000 5,000,000 25,000,000
Proposal 15,000,000
2
4,2
p
Max per Order St.
Score 40 60 60 100

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment:
406
A Case Study in the Healthcare Services
Appendix 2. Suggested likert-sca le rank score transformation.
Rating/Parameter 1 2 3 4 5
33
10 107.5 10m
1
1,1
p
Min per Shipment 3
10 10m 44
0.75 100.510m33
510 2.510m 3
2.5 10m


Max per Shipment
1
1,2
p
5
510  65 105 10 66 1.5 1010 66 210 1.510

  6 210
Unit Cost 1
2,1
p
0.2c 0.180.2c 0.160.18c

 0.140.16c 0.14c

Expected Defect Rate
1
3,1
p
1P 0.8 1P 0.60.8P

 0.40.6P

 0.4P

Suggested Supply
Time 2
1,1
p
5LT 45LT 34LT

 23LT

 2LT

Credit Rating 2
1,2
p
1Cre

2Cre

3Cre

45Cre  Cre

Reputation 2
1,3
p
Very low
performance Low performance Medium performance High performance Very high performance
Order Cost 2
2,1
p
900oc  900800oc 800700oc 700600oc 600oc

Quantity Discount2
3,1
p
200D 150200D 100150D

 50100D

 50D

Geographic Location
2
3,2
p
30CR 3060CR 6090CR

 90120CR

 120CR
Payment Date 2
3,3
p
5
10o 55 100.75 10o 55 0.75 100.5 10o

 55 0.510 0.2510o

 5 0.25 10o


Min per Order 2
4,1
p
6
510O 77 0.5 1010O 77 101.5 10O

 77 10210O

 7 210O
The parameter – Quantity Discount – has been specified with a transformation which expresses a binary rating,
as shown in the table below.
2
3,1
p
Selection Parameter/ Rating 0 1
2
3,1
p
Quantity Discount None √
Appendix 3. Selection parameter weights.
Selection Parameter Weight
Min per Shipment1
1,1
p
0.1
Max per Shipment1
1,2
p
0.07
1
2,1
p
Unit Cost 0.15
1
3,1
p
Expected Defect Rate0.08
2
1,1
p
Suggested Supply Time0.1
2
1,2
p
Credit Rating0.06
2
1,3
p
Reputation 0.07
2
2,1
p
Order Cost0.1
2
3,1
p
Quantity Discount0.03
2
3,2
p
Geographic Location0.03
2
3,3
p
Payment Date0.03
2
4,1
p
Min per Order 0.04
2
4,2
p
Max per Order0.07

Off-Line and User-Oriented Approach for Supplier Selection in Dynamic Environment:
A Case Study in the Healthcare Services
407
Appendix 4. Subjective parameters in dynamic selection weights.
As a function of
annual demand
Selection Parameter As a function of
interest rate
As a function of
supply time
As a function of
raw material cost
As a function of
exchange rate
Min per Shipment1
1,1
p
1.7

 2b

 3b

0.04


Max per Shipment1
1,2
p
0.08

 3b

1b

 2b

Unit Cost1
2,1
p
0.0001

 0.0001


8b 0.0001


Expected Defect Rate1
3,1
p
6b

0.05


Suggested Supply Time2
1,1
p
0.0001
  10b

4  
Credit Rating2
1,2
p
0.0001
  2b 0.05  
Reputation 2
1,3
p
0.0001

 10b

5b
Order Cost2
2,1
p
2b 0.001


3b

Quantity Discount2
3,1
p
1.5b 4b

1.5b

 4b

Geographic Location2
3,2
p
10b

10b
Payment Date2
3,3
p
0.0001   0.0003   0.0003  
Min per Order2
4,1
p
2
  2b

 0.002  
Max per Order2
4,2
p
0.05

 2b

5.1

b