Int. J.

Production Economics 112 (2008) 548569

Economical assessment of the impact of RFID technology and
EPC system on the fast-moving consumer goods supply chain
Eleonora Bottani

, Antonio Rizzi
Department of Industrial Engineering, University of Parma, Viale G.P. Usberti 181/A, 43100 Parma, Italy
Received 28 February 2006; accepted 11 December 2006
Available online 9 June 2007
Abstract
This paper describes a research whose aim is to quantitatively assess the impact of radio frequency identication (RFID)
technology and electronic product code (EPC) system on the main processes of the fast-moving consumer goods (FMCG)
supply chain.
A three-echelon supply chain is examined, composed of manufacturers, distributors and retailers of FMCG. A
questionnaire survey was performed to collect both quantitative and qualitative data related to logistics processes of each
player. Starting from these data, a feasibility study has been carried out in order to assess the economical suitability of
RFID and EPC adoption in the FMCG supply chain, both for each player and for the FMCG supply chain as a whole.
Results of the feasibility study show that RFID and EPC implementation is still not protable for all echelons examined.
Specically, both from a non-integrated and from an integrated scenario, RFID adoption with pallet-level tagging
provides positive revenues for all supply-chain players. Conversely, adopting a case-level tagging, substantial costs arise for
manufacturers, involving negative economical results. Outcomes of this study provide an economical justication to the
RFID and EPC implementation in the FMCG supply chain.
r 2007 Elsevier B.V. All rights reserved.
Keywords: RFID; EPC; Investment analysis; FMCG supply chain
1. Introduction
Radio frequency identication (RFID) is the
generic name for technologies that use radio waves
to automatically identify individual items. There are
several methods of identifying items using RFID,
but most of the systems consist of a reader and a
tag, the latter being made up of two main
components, namely an antenna and a chip. The
reader sends out electromagnetic waves that form a
magnetic eld when they join with the antenna on
the RFID tag. The tag draws power from the
magnetic eld and uses it to power the microchips
circuits. The chip then modulates the waves that the
tag sends back to the reader and the reader converts
the new waves into digital data. Data are stored into
the tag chip in the form of an electronic product
code (EPC). EPC standards have been developed by
the Auto-ID Center, a partnership founded in 1999
by ve leading research universities and nearly 100
leading retailers, consumer products makers and
software companies (Niemeyer et al., 2003).
EPC data collected are then passed to and shared
through the EPCglobal network. The latter has been
dened as a way of leveraging the internet to access
ARTICLE IN PRESS
www.elsevier.com/locate/ijpe
0925-5273/$ - see front matter r 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijpe.2007.05.007

Corresponding author. Tel.: +39 0521 905872;

fax: +39 0521 905705.
E-mail address: [email protected] (E. Bottani).
a large amount of logistics information that can be
shared among authorized partners (EPC Global,
2004). Once EPC data are collected by reading
RFID tags of cases and pallets, EPC numbers
become secure data on companies middleware.
Thus, they can be passed to other legacy applica-
tions to generate business events and can be made
available for selected and authorized trading part-
ners through the Internet network and a suite of
selected services, namely EPC information services
(EPCIS) and discovery services (DS) (Harrison,
2004). EPCIS and DS enable users to nd on the
Internet data related to specic EPCs and to gain
access and retrieve those data. Through this real-
time data-sharing mechanism, companies have
broad and plain visibility over logistic ows and
can leverage this information to optimize supply-
chain management.
Recently, the adoption of RFID technology and
EPC standards for products identication, as well as
of EPCglobal network for information manage-
ment, is experiencing an increasing diffusion in the
logistics pipeline, where they are expected to have a
major impact on labour efciency, processes auto-
mation and accuracy (Agarwal, 2001; Prater et al.,
2005).
The purpose of this paper is to describe a research
carried out by the Department of Industrial
Engineering of the University of Parma, whose aim
was to quantitatively assess the economic impact of
RFID and EPC adoption in the fast-moving con-
sumer goods (FMCG) supply chain. A quantitative
feasibility study, encompassing the costs and benets
of such technologies, has been carried out to quantify
the economical protability of the implementation
and to provide a justication for technology invest-
ments. The entire research was supported by GS1
Italy, the Italian representative of EPCglobal Inc.,
which coordinates diffusion and correct implementa-
tion of EPC standards in more than 100 countries.
The remainder of the paper is organized as
follows. In the next section, literature related to
RFID and EPC is examined in detail. Then,
research aims, methodology and steps are described.
The main results of the research are presented and
discussed in Section 4. Finally, concluding remarks
and future research directions are pointed out.
2. Literature review
Scientic literature has widely debated the rea-
sons and the current limits to RFID and EPC
adoption, as well as related benets on supply-chain
processes. Jones et al. (2004) argue that a main
reason for RFID diffusion is the capability of
tags to provide more information about products
than traditional barcodes. Manufacturing site,
production lot, expiry date and components
type are among information that can be stored into
the tag chip. Moreover, tags do not need line-
of-sight scanning to be read, since they act as
passive tracking devices, broadcasting a radio
frequency when they pass within yards of a reader
(Karkkainen, 2003). RFID tags also solve some of
the inefciencies associated with traditional bar-
codes. For example, reading barcodes requires
manual operations on packages, that is either the
packages with barcodes or the reading devices
should be manually handled to read the codes
(Boxall, 2000; Bylinsky, 2000; Jones, 1999). This
may result in time consumption and difcult data
capture if large amounts of goods have to be
handled, such as in distribution centres or retail
stores. In some cases, readability of barcodes can
also be problematic, due to dirt and bending,
bringing about reduced accuracy and low reading
rate (Ollivier, 1995; Moore, 1999).
Prater et al. (2005) have discussed the main
benets of RFID and of the EPCglobal network
adoption for supply-chain processes, for the specic
case of the grocery retailing. The availability of real-
time information is regarded as the main benet,
although additional outcomes can be found in
increased inventory visibility, stock-out reduction,
real-time access and update of current store
inventory levels, automated proof of delivery
(Fernie, 1994), availability of accurate points of
sale data, reduction of labour associated with
performing inventory counts of shelved goods,
improved theft prevention and shrinkage, and better
control of the whole supply chain (Bushnell, 2000).
Despite the achievable benets, several authors
agree that the main limit to a wide use of RFID
technology has to be found in its cost (Prater et al.,
2005; Karkkainen and Holmstrom, 2002; Burnell,
1999; Riso, 2001). Consequently, critics to RFID
argue that investments in tags and readers, as well
as in the related informatics infrastructure, are still
not protable. Nonetheless, quantitative feasibility
studies about the economic protability of RFID
and EPC implementation still lack in literature.
To this extent, Prater et al. (2005) argue that
cost/benet justication is a major barrier to the
adoption of RFID and EPC, concluding that future
ARTICLE IN PRESS
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 549
research should investigate how to best implement
these technologies in the grocery industry.
3. The research
3.1. The FMCG supply chain and related processes
The impact of RFID technology and EPCglobal
network on the FMCG industry was quantied on a
three-echelon supply chain, whose general model is
provided in Fig. 1.
The supply chain hypothesized starts with pro-
duction of palletized SKUs at the production site of
a manufacturer, where FMCG are shipped to the
manufacturers DC. Here, pallets are rst received
and checked in, in order to apply the received goods
to the correct transfer/purchase order and to
identify possible discrepancies. Once received, pal-
lets are moved to the storage area. Fork lift truck
(FLT) drivers pick them up at the receiving bays
and drop them in the correct storage locations.
Inventory management systems have to be updated
through ofce operation to correctly manage
inventory levels and locations. Eventually, stored
pallets will be either moved to shipping bays for
direct shipment or delivered to picking locations. In
the latter case, picking personnel usually retrieve
cases or items required on an order-picking basis.
For each picking operation, the current inventory
has to be updated in the companys information
system (IS), checking that the correct quantity was
removed. Pallets should then be shipped to either
retailers DCs or retail stores. Operations at the
shipping bays involve not only loading pallets on
the trailers but also checking out goods, the latter
entailing identication and correction of possible
mistakes and discrepancies. Clerical activities are
also needed to create shipping documents. The same
processes are repeated for the retailers DC, with
a stronger emphasis on receiving and picking
processes.
The logistics channel analysed ends at the shop
oor of the retailer, such as a supermarket or a
department store. When a store receives pro-
ducts shipped from its DCs or from a third-party
supplier, it will either check that the goods received
pair what was expected or automatically receive
the shipment without punctual verications. The
former process usually applies either to high-value
goods or to products shipped from third-party
suppliers, which are worth a punctual checking of
actual quantities and items at the point of
reception. Conversely, automatic receiving is
adopted for goods shipped from the local DC,
which have already been checked. In both cases,
further accounting activities are required to update
on hand inventories.
Received goods can be moved either to the
backroom or to the shop oor shelves. Intense shop
oor replenishment activities usually occur in
department stores, while supermarkets and small
retailers are served on a daily basis with negligible
backroom storage activities. As long as products are
going out of stock on the shop oor, they should be
replenished. This requires employees to continu-
ously check the shelves and the backroom for
product availability.
3.2. The data collection phase and the research
methodology
The data collection phase of the research involved
a panel of 11 major companies, namely 6 manu-
facturers and 5 distributors of FMCG, adhering to
GS1 Italy. For each participant, the analysis was
focused on either a DC or a warehouse, previously
indicated by the company and which could be
considered enough illustrative of related processes.
As an example, DCs with advanced IS or with
considerable ows of goods were chosen. Analo-
gous criteria have been adopted to dene a retail
store for each distributor involved in the study.
ARTICLE IN PRESS
Procurement Operations Distribution
Manufacturers DC Distributors DC Distributors retail
store
Fig. 1. The supply chain analysed.
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 550
Table 1 provides a list of the participants, as well as
the location of the structures chosen for the
research.
For each participant, the relevant processes that
could be affected by the adoption of the RFID
technology were examined. A questionnaire survey
was carried out to systematically gather information
related to supply-chain processes of manufacturers
and distributors. Two different questionnaires have
been designed and developed for comprehensive
data collection, respectively, for DCs (manufacturer
and retailer) and for retailers (either supermarket or
department store). In designing the questionnaires,
two opposite goals had to be accomplished. On the
one hand, whenever possible, the questionnaire
should be as analytic and quantitative as possible,
in order to collect quantitative data about the
current processes of the player examined. On the
other hand, the whole questionnaire would not
require more than three hours to be lled in, that is
the average time given from participants for the
visits.
Both questionnaires are composed of a rst part,
mainly aimed at collecting general information as
well as common data for all processes examined.
Examples of such information include products type
and amount, average value of pallet/case, number of
employees and related costs, and DC/retail store
area. Subsequent sections aim at collecting data
related to the processes performed by DCs and
retail stores, such as receiving, putaway,
picking and sorting and shipping for DCs,
and receiving, backroom management and
expositive area management for retail stores. As
far as retail stores are concerned, our analysis
focused on processes that could be impacted by
pallet/case-level tagging; conversely, we do not
consider processes, such as point-of-sale check-outs,
which can only be impacted by an item-level
tagging. Table 2 shows the scheme of questionnaires
adopted, as well as the main data collected.
For each process, both quantitative and qualita-
tive parameters were examined, such as the amount
of goods ow or the identication technologies
currently adopted in the process. For quantitative
parameters, the numerical values were collected
(the amount of the goods ow at the receiving docks
is 2000 pallets per day), while descriptions were
provided for qualitative information (the opera-
tions carried out by the employee at the receiving
dock are the followings). Moreover, current per-
formances of the DC/retail store were analysed.
Particular attention was paid to performance
parameters that, according to the literature, can be
improved by adopting RFID technology and
EPCglobal network, namely, process accuracy
(number of mix, quantity, documentation and
identication errors and related consequences),
labour efciency and inventory management (stock
levels, number of manual inventory counts, inven-
tory rotation, safety stock level and stock-outs).
Data were collected through a visit campaign
which took place from November 2004 to April
2005. Each visit involved academics of the Uni-
versity of Parma, a representative of GS1 Italy and
one or more members in charge for the visited
organization. As a matter of fact, respondents
needed to have a good knowledge and experience
about the processes examined. All sessions were
taped, in order to easily retrieve data that could
have been missed during the interviews. On average,
23 h were spent for each participant, equally shared
in visiting the distributive structure and answering
the questionnaire.
Data collected were adopted to edit as many case
studies which emerged as a viable tool to analyse
logistics and supply-chain management issues (see,
for instance, Byrne and Heavey, 2006). The
structure of these reports for DCs and retailers
follows the same scheme as adopted in the related
questionnaire. Case studies aim at quantitatively
describing the processes currently performed by the
DC/retail store, highlighting features suitable to be
affected by RFID and EPC implementation.
Around 15 case studies have been edited, corre-
sponding to 6 manufacturers DC, 5 distributors
DC and 4 retail stores.
ARTICLE IN PRESS
Table 1
List of participants to the research and related distributive
structures
Participant Distribution centre Retail store
Auchan Calcinate (BG)
Campari Stezzano (BG)
Carrefour Cameri (NO) Torino
Finiper Soresina (CR) Orio al Serio (BG)
Heineken Stezzano (BG)
LOre al Calvignasco (MI)
Nestle Albairate (MI)
Nordiconad Modena Bologna
Pam Spinea (VE) Bologna
Procter&Gamble Pomezia (RM)
Sony Paderno Dugnano (MI)
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 551
ARTICLE IN PRESS
Table 2
Scheme of the questionnaires adopted
DC processes Quantitative data Qualitative data Performance
Receiving Number of pallets/day received;
number of orders/day received;
number of receiving bays; number
of pallets to be labelled with bar
codes
Equipments currently adopted for
products identication and related
connection with the companys IS;
traceability information included in
products received; manual operations
required for products identication
Errors made in products receiving
(identication, documents,
product type and quantity); time
required to amend those errors;
time required to perform manual
operations on pallets
Putaway Storage capacity; number of FLT;
number of pallets/day handled;
number of inventory counts/year;
time required to identify the
storage location; time required to
update the IS; number of pallets/
year reaching the expiry date
Warehouse type; items allocation
policy; equipments currently adopted
for products identication and related
connection with the companys IS;
manual operations required for
products identication; inventory
management policy; time required for
replenishment
Average stock out; amount of
safety stocks; time required to
perform manual operations on
pallets; time required to perform
inventory counts
Picking and
sorting
Number of items in the picking
area; number of trucks; number of
pallet/day handled in picking and
sorting; orders prole; number of
orders/day fullled; time required
to identify the picking location;
time required to update the IS
Picking policy; equipments currently
adopted for products identication
and related connection with the
companys IS; replenishment policy;
manual operations required for
products identication during picking
and sorting
Errors made in picking operations
(product type and quantity); time
required to amend those errors;
time required to perform manual
operations during picking and
sorting; time required to check
picking pallets; amount of shrink
Shipping Number of pallets/day shipped;
number of orders/day shipped;
time for orders fullment; number
of shipping bays; time required to
update the IS
Equipments currently adopted for
products identication and related
connection with the companys IS;
traceability information included in
products shipped; manual operations
required for products identication
Errors made in products shipping
(identication, documents,
product type and quantity); time
required to amend those errors;
time required to perform manual
operations on pallets
Retail stores
processes
Quantitative data Qualitative data Performance
Receiving Number of pallets/day received; number of
orders/day received; number of receiving
bays; number of pallets to be labelled with
bar codes
Equipments currently
adopted for products
identication and related
connection with the
companys IS; traceability
information included in
products received; manual
operations required for
products identication
Errors made in products
receiving (identication,
documents, product type and
quantity); time required to
amend those errors; time
required to perform manual
operations on pallets
Backroom
management
Storage capacity; number of FLT; number of
pallets/day handled; number of inventory
counts/year; time required to identify the
storage location; time required to update the
IS; number of pallets/year reaching the
expiry date; number of urgent deliveries/
month
Warehouse type; items
allocation policy; equipments
currently adopted for
products identication and
related connection with the
companys IS; manual
operations required for
products identication;
inventory management
policy; time required for
replenishment
Average stock out; amount of
safety stocks; time required to
perform manual operations
on pallets; time required to
perform inventory counts
Expositive area
management
Number of gates connecting to the
backroom; number of inventory counts/year;
amount of thefts/year; amount of items/year
reaching the expiry date
Equipments currently
adopted for replenishment
management and related
connection with the
companys IS
Average stock out; time
required to perform items
replenishment; replenishment
rate; time required to perform
inventory counts
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 552
3.3. The representative supply chain and the
scenarios examined
Starting from the case studies, a representative
FMGC supply chain has been dened, based on the
similarities between processes analysed in the panel
of companies involved. The representative supply
chain is composed of three echelons, namely a
manufacturers DC, a distributors DC and a
retailer. Representative structures are characterized
by average features, which have been obtained from
data collected. Specically, quantitative parameters
of the representative supply chain have been derived
as the mathematical average of the quantitative
values. This is, for instance, the case of the amount
of pallets transhipped at the DC/retail store, of the
time required to perform manual operations, as well
as of errors currently made in processes examined.
For qualitative characteristics, such as the set of
activities required to perform identication and
transhipment of products or traceability informa-
tion included in the SKUs received, the most
frequently encountered situations have been as-
sumed as representative.
Outlining a representative supply chain pro-
vides a twofold result. First, such a supply chain can
be assumed as the current scenario of the distribu-
tive logistics for FMCG, since it was derived from
DCs/retailers previously analysed. Second, the
representative supply chain was adopted to assess
the economical impact of RFID and EPC imple-
mentation for three representative players of the
FMCG industry, as well as for the FMCG supply
chain as a whole. To this extent, the economical
assessment was made by analysing two different
scenarios, namely non-integrated and inte-
grated scenario, whose architecture is shown in
Fig. 2.
As can be seen from Fig. 2, in the non-
integrated scenario, EPC data are not shared
between supply-chain partners, being, conversely,
real-time available for internal processes. Thus, the
ARTICLE IN PRESS
Manufacturer
production site
Manufacturers
DC
Retail store Retailers DC
Non integrated scenario
Data warehousing
and Information
System
Manufacturer
Manufacturer
production site
Manufacturers
DC
Retail store Retailers DC
Data warehousing
and Information
System
Retailer
EPC data
Integrated scenario
real-time information sharing
flow of tagged pallets/cases
tagged pallet/case
Legend:
Fig. 2. Qualitative schemes of the non-integrated and integrated scenarios.
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 553
main benets of RFID technology and EPC
standards can be found in the improvement of
accuracy, automation and labour efciency of
internal processes. In the integrated scenario,
EPC data on products and pallets become available
to all players by means of the EPCglobal network.
Additional benets can thus derive from real-time
information availability, data sharing through the
whole supply chain and improved visibility. More-
over, the global data synchronization network
(GDSN) may avoid inaccurate or outdated data,
by providing an environment for streamlined data
sharing throughout the supply chain (UCCnet,
2005).
Two deployment levels of RFID and EPC have
also been hypothesized, namely pallet-level and
case-level tagging. EPC Class 1 gen 2 UHF smart
labels (RFID Journal, May 2005) are assumed to be
adopted for tagging. Conversely, the application of
RFID tags on items has not been considered,
although it can be assumed to be the ultimate
deployment level of RFID. As a matter of fact, this
scenario is still far from a practical adoption, due to
high costs of RFID technology compared with the
average value of FMCG.
Combining technological level with relationships
among partners, four different scenarios have been
examined in total, which have been reengineered as
described in the following sections.
3.4. Reengineering models
Reengineering (To be) models refer to the
application of RFID technology and EPC system in
the scenarios previously mentioned, and encompass
two main aspects. For each supply-chain player, the
models describe the technological reengineering of
activities and processes, that is, RFID equipments
required to optimally perform the supply-chain
processes (readers, xed and mobile devices, IS,
and networking, etc.). Moreover, reengineering
models involve a new denition of procedures,
activities and operations to perform those pro-
cesses, according to business process reengineering
(Hammer and Champy, 1993).
Owing to the wide number of processes and
scenarios analysed, the description of reengineered
models cannot be fully reported in the paper. Thus,
as an example, we detail the reengineered receiv-
ing process at manufacturers DC with pallet-level
tagging, whose scheme is shown in Fig. 3, from both
non-integrated and integrated perspectives.
As can be seen from Fig. 3, in the To be
scenario, pallets are identied through RFID
disposable smart labels. RF coverage is required in
the whole receiving bay of the DC, allowing real-
time connection between mobile devices and com-
panys IS. Each receiving dock should be equipped
by an RFID gate, which acts as a portal where EPC
codes of pallets can be read. Ethernet network
connects RFID gates to the IS. In the non-
integrated scenario, data related to products are
not available in advance; thus, check-in operations
do not sensibly differ from the As is scenario,
except for removing of pallet-labelling operations.
Conversely, in an integrated approach, EPC data
can be real-time retrieved from the EPCglobal
network (see Fig. 2), checking for comparisons
between expected products and those unloaded at
the dock. Availability of the GDSN makes it
possible to retrieve static data about products when
needed, avoiding manual identication processes as
well as related inaccuracies.
Reengineering models were developed for all
processes of DCs and retailers, highlighting possible
benets stemming from RFID and the EPCglobal
network. The scenarios described by Auto-ID
Center (2002), McFarlane (2002) and Agarwal
(2001) have been found particularly helpful to
dene the reengineered processes of DCs and retail
stores.
Grounding on the gap between As is and
To be models for the scenarios described in
Section 3.3, appropriate computational models
assess the economical impact of RFID and EPC
implementation. Table 3 shows costs and savings
computed by the models depending on tagging level
and process examined starting from data collected
during the survey phase (see Table 2). Costs and
savings listed in the table are computed for both the
integrated and the non-integrated scenarios;
additional savings achievable only in the inte-
grated perspective are highlighted (in italics) in the
table.
As can be seen from the table, costs and savings
are mainly related to RFID equipments, manpower
efciency and inventory optimization. Specically,
costs of RFID equipments were computed as the
product of the number of elements to be equipped,
such as docks or FLTs, and the unitary cost of
equipments, which are detailed in Section 3.5.
Similarly, costs of tags/labels were computed start-
ing from the number of pallets/cases to be tagged
or labelled, and the unitary costs of tags/labels.
ARTICLE IN PRESS
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 554
Savings in labour related to each manual operation
were derived starting from the hourly cost of
manpower, the time required to perform the
operation on each pallet/case and the number of
pallets/cases. This is, for instance, the case of costs
of manpower for labelling activities, errors amend-
ment, check-in operations on received/shipped
pallets/cases, and inventory counts, etc. The com-
putational procedure for stock-out and safety stocks
abatement is described in Section 4.2.
3.5. Feasibility study
To assess the economic competitiveness of RFID
implementation, the discounted cash ow (DCF)
that can be brought in by the reengineered models
previously described was computed.
Investments for RFID and EPC implementation
were rst quantied. As mentioned, they mainly
encompass RFID equipments, such as tags, readers,
xed and mobile devices, and related software, up to
the implementation project. Unitary costs, shown in
Table 4, have been derived from literature, market
prices, catalogues of RFID equipment producers
and vendors.
As can be seen from the table, two costs have
been hypothesized for RFID tags, which refer to
low and high quantities of tags purchased. In
particular, low quantity cost value can be applied
when thousand of tags are purchased, and has been
used when examining the pallet-level tagging
scenario. Conversely, a high quantity one is
suitable to be adopted when the amount of RFID
tags purchased reaches millions of units, as re-
quired in the case-level tagging scenario. Punctual
values were deduced from the latest press releases of
major manufacturers (RFID Journal, September 15,
2005).
ARTICLE IN PRESS
As is process
To be process - internal scenario
Arrival of pallets with
logistics labels at the
receiving dock of the DC
Manual identification of
pallets received
Pallet labeling
Pallet are shipped to the
storage area
Manual checking of
pallets received
To be process - integrated scenario
Arrival of pallets with
RFID tags at the receiving
dock of the DC
Manual identification of
pallets received
Pallet are shipped to the
storage area
Manual checking of
pallets received
Arrival of pallets with
RFID tags at the receiving
dock of the DC
Real-time identification of
pallets received
Pallet are shipped to the
storage area
Real-time checking of
pallets received
RFID tags to be attached
on pallets
RFID gate for each
receiving dock
RF coverage of the
receiving area
Ethernet connection with
the companys IS
RFID tags to be attached
on pallets
RFID gate for each
receiving dock
RF coverage of the
receiving area
Ethernet connection with
the companys IS
Fig. 3. Flowchart of the reengineered receiving process for manufacturers DC (note: italics RFID equipments/devices required).
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 555
ARTICLE IN PRESS
Table 3
Costs and savings computed by the models
Pallet level tagging Case level tagging
Distribution
centre
Receiving Costs: Costs:
cost of RFID gates cost of RFID gates
Savings: Savings:
cost of pallet labels
costs of manpower for labelling operations
on received pallets
costs of manpower required to amend errors
in received pallets (documents, product type
and quantity, identication)
costs of manpower required for checking
operations on received pallets and to update
the informative system
cost of pallet labels
costs of manpower for labelling operations
on received pallets
costs of manpower required to amend errors
in received pallets (documents, product type
and quantity, identication)
costs of manpower required for checking
operations on received pallets and to update
the informative system
Putaway Costs: Costs:
cost of RFID equipments for FLT cost of RFID equipments for FLT
Savings: Savings:
costs of manpower required to perform
general and replenishment inventory counts
costs of manpower required to identify
pallets and location and to update the
informative system
cost of shrinkage
costs of manpower required to perform
general and replenishment inventory counts
costs of manpower required to identify
pallets and location and to update the
informative system
cost of shrinkage
Picking and
sorting
Costs: Costs:
cost of tags to be attached on pallets
cost of labels for picking pallets
identication
cost of printers of smart labels
cost of tags to be attached on pallets
cost of readers for EPC labels
cost of RFID equipments for trucks
Savings:
cost of pallets labels;
costs of manpower required to amend errors
in picking pallets (documents, product type
and quantity, identication)
costs of manpower required to identify
pallets and location and to update the
informative system
costs of manpower required for checking
operations on picking pallets
cost of shrinkage of picking inventory
costs of manpower required to perform
replenishment inventory counts of picking
inventory
Shipping Costs: Costs:
cost of RFID gates cost of RFID gates
Savings: Savings:
costs of manpower required to amend errors
in shipped pallets (documents, product type
and quantity, identication)
costs of manpower required for checking
operations on shipped pallets and to update
the informative system
costs of manpower required to amend errors
in shipped pallets (documents, product type
and quantity, identication)
costs of manpower required for checking
operations on shipped pallets and to update
the informative system
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 556
The value of RFID and EPC implementation
was computed by considering all the annual cash
ows related to prot and loss changes when RFID
and EPCglobal network are brought in. Specic
amortization periods were hypothesized for differ-
ent investment categories. As a nal result, net
present value (NPV), payback period (PBP) and
internal rate of return (IRR) were quantied over a
5-year period and considering a 5% interest rate.
4. Results
Results obtained for the representative supply
chain allow quantifying the economical impact
ARTICLE IN PRESS
Table 3 (continued )
Pallet level tagging Case level tagging
Savings Savings
reduction of safety stocks
reduction of stock outs at the retail store
reduction of safety stocks
reduction of stock outs at the retail store
Retail store Receiving Costs: Costs:
cost of RFID gates cost of RFID gates
Savings: Savings:
costs of manpower required to amend errors
in received pallets (documents, product type
and quantity, identication)
costs of manpower required for checking
operations on received pallets and to update
the informative system
costs of manpower required to amend errors
in received pallets (documents, product type
and quantity, identication)
costs of manpower required for checking
operations on received pallets and to update
the informative system
Backroom
management
Costs: Costs:
cost of RFID equipments for FLT cost of RFID equipments for FLT
Savings: Savings:
costs of manpower required to perform
general and replenishment inventory counts
costs of manpower required to identify
pallets and location and to update the
informative system
cost of shrinkage
costs of manpower required to perform
general and replenishment inventory counts
costs of manpower required to identify
pallets and location and to update the
informative system
cost of shrinkage
Expositive
area
management
Costs:
cost of RFID gates for cases
Savings:
cost of shrinkage in the expositive area
costs of manpower required to perform
replenishment inventory counts of the
expositive area
cost (turnover losses) for lack of products in
the expositive area
Savings Savings
reduction of safety stocks reduction of safety stocks
Note: italic additional costs/savings to be computed in the integrated scenario.
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 557
of RFID technology and EPC standard imple-
mentation for the four scenarios, that is, inte-
grated vs. non-integrated, pallet- and case-level
tagging.
4.1. Non-integrated scenario
Outcomes of the feasibility study for non-
integrated scenario and pallet-level tagging are
shown in Table 5.
As can be seen from the table, manufacturers DC
benets from about h220,000 NPV of RFID and
EPC implementation, which was computed accord-
ARTICLE IN PRESS
Table 4
Costs of RFID equipments
Hardware and software RFID equipments Costs
RFID tag UHF wide quantity (h/tag) h0.10
RFID tag UHF low quantity (h/tag) h0.15
Pallet label (h/label) h0.035
Case label (h/label) h0.005
Printer of logistics labels (h/printer) h2500.00
Printer of case labels (h/printer) h2500.00
RFID reader (h/reader) h2000.00
RFID gate (h/gate) h2500.00
Equipment of a RFID truck (h/truck) h5000.00
Software and implementation project (h) h200,000.00
RF coverage (h/m
2
) h4.00
Table 5
Results of investment evaluation for the FMCG supply chain in non integrated scenario with pallet level tagging
Manufacturers DC Years
0 1 2 3 4 5
Discount rate 1 0.95 0.91 0.86 0.82 0.78
Cost savingscost arising
Tagging and labelling 45,000 45,000 45,000 45,000 45,000
Receiving 112,665 112,665 112,665 112,665 112,665
Warehouse management 34,709 34,709 34,709 34,709 34,709
Order picking 12,500 12,500 12,500 12,500 12,500
Shipping 69,780 69,780 69,780 69,780 69,780
Total cost/saving 159,654 159,654 159,654 159,654 159,654
Amortization 59,000 59,000 59,000 59,000 59,000
Gross margin 100,654 100,654 100,654 100,654 100,654
Taxes 34,222 34,222 34,222 34,222 34,222
Net prot 66,432 66,432 66,432 66,432 66,432
Investment 324,500 0 0 0 0 0
Net cash ow 324,500 125,432 125,432 125,432 125,432 125,432
Discounted cash ow 324,500 119,459 113,770 108,353 103,193 98,279
Net present value (h) 218,554
Payback period (years) 2.842
Internal rate of return (%) 26.90%
ROI (%) 26.86%
Distributors DC Years
0 1 2 3 4 5
Discount rate 1 0.95 0.91 0.86 0.82 0.78
Cost savingscost arising
Tagging and labelling
Receiving 79,645 79,645 79,645 79,645 79,645
Warehouse management 52,323 52,323 52,323 52,323 52,323
Order picking 31,500 31,500 31,500 31,500 31,500
Shipping 42,000 42,000 42,000 42,000 42,000
Total cost/saving 142,468 142,468 142,468 142,468 142,468
Amortization 91,000 91,000 91,000 91,000 91,000
Gross margin 51,468 51,468 51,468 51, 468 51,468
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 558
ing to the following set of equations:
NCF and DCF being the net and DCF, respectively,
j 0,y,n the year considered and i the interest rate
hypothesized.
For this player, investments are mainly due to
costs for purchasing RFID tags, which were
quantied starting from the number of pallets
received at the DC and the low quantity cost of
tags proposed in Table 4. Moreover, costs of RF
coverage of the receiving bay and costs to equip
receiving docks with RFID gates, as well as for
software implementation, have been included in DC
investments. As mentioned in Section 3.4, such costs
were derived starting from the area of the receiving
bay, the number of receiving docks and the costs of
RF equipments shown in Table 4. The PBP of DCs
investments was estimated in about 3 years,
ARTICLE IN PRESS
Table 5 (continued )
Taxes 17,499 17,499 17,499 17,499 17,499
Net prot 33,969 33,969 33,969 33,969 33,969
Investment 500,500 0 0 0 0 0
Net cash ow 500,500 124,969 124,969 124,969 124,969 124,969
Discounted cash ow 500,500 119,018 113,350 107,953 102,812 97,916
Net present value (h) 40,549
Payback period (years) 4.586
Internal rate of return (%) 7.90%
ROI (%) 9%
Retailer Years
0 1 2 3 4 5
Discount rate 1 0.95 0.91 0.86 0.82 0.78
Cost savingscost arising
Tagging and labelling
Receiving 10,452 10,452 10,452 10,452 10,452
Putaway 29,880 29,880 29,880 29,880 29,880
Order picking
Shipping
Total cost/saving 40,332 40,332 40,332 40,332 40,332
Amortization 25,720 25,720 25,720 25,720 25,720
Gross margin 14,612 14,612 14,612 14,612 14,612
Taxes 4968 4968 4968 4968 4968
Net prot 9644 9644 9644 9644 9644
Investment 141,460 0 0 0 0 0
Net cash ow 141,460 35,364 35,364 35,364 35,364 35,364
Discounted cash ow 141,460 33,680 32,076 30,549 29,094 27,709
Net present value (h) 11,648
Payback period (years) 4580
Internal rate of return (%) 800%
ROI (%) 9%
NCF
j
Total cost=saving
j
amortization
j
taxes
j
investment
j
,
DCF
j
NCF
j

1
1 i
j
,
NPV
X
n
j0
DCF
j
, 1
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 559
according to the following equation:
PBP min n
X
n
j0
DCF
j
40

( )
. (2)
Finally, the IRR of the investment was computed
as the interest rate that makes DCF null at the end
of the 5-year period considered:
IRR

i
X
n
j0
DCF
j
0

( )
. (3)
In this scenario, substantial benets are achieved
when applying RFID technology to receiving,
warehouse management and shipping pro-
cesses. Specically, from the numerical results of
the computation, the following conclusions emerge.
Savings in receiving activities result for about
90% from shortening of time required to perform
checking operations on pallets received, and for
about 10% both from improved accuracy of the
process and from avoiding labelling operations.
About 75% of savings in warehouse management
can be attributed to reduction of shrinks, which,
however, are not completely removed. Cutting
down the time required for replenishment cycles
and to perform inventory counts involves the
remaining 18% of costs savings. Finally, shipping
operations benet from a dramatic shortening of
time required for pallets identication, which yields
most of the savings (54%) of this process.
Investment analysis of distributors DC leads to
similar remarks. Distributors DC benets from
h40,000 NPV, and initial investments are paid back
in 4.5 years. If compared with the manufacturers
DC, no costs arise in tagging and labelling
activities, since we assumed that such costs are wholly
charged to the manufacturers. Consequently, cost of
tags required for picking pallets and costs for
software and implementation represent the main
investments for distributors DC. In shipping
processes, savings result from shortening of time
required to perform checking operations on pallets
shipped (80%), from better accuracy of the process
and from removing labelling operations (20%), the
latter being also the main benet achievable in
receiving. Warehouse management benets from
a dramatic shortening of time required to perform
replenishment cycles and inventory counts. Conver-
sely, no substantial benets are achieved in shrink
abatement, since product losses due to expiration
date have emerged as trivial for the retailers DC.
Finally, retail stores benets from a NPV of
about h11,000, and initial investments are paid back
in 4.5 years. As shown in Table 5, only receiving
and putaway operations can be affected by the
implementation of RFID and EPC standards in a
pallet-level tagging scenario. For receiving opera-
tions, costs arising are primarily due to investments
in software required for processes management,
installation of RFID gates and RF coverage of the
receiving area. Savings mainly result from reducing
the time required to perform counts and check-in
operations on pallets received (75%). Similarly, RF
coverage and FLT equipments count for 30% and
70% of the arising costs in putaway activities.
Nonetheless, cost savings can be achieved in terms
of reduction of time required for inventory counts,
which allows about 90% of total savings.
Table 6 shows the results of investment analysis in
the non-integrated scenario with case-level tagging.
ARTICLE IN PRESS
Table 6
Results of investment evaluation for the FMCG supply chain in non integrated scenario with case level tagging
Manufacturers DC Years
0 1 2 3 4 5
Discount rate 1 0.95 0.91 0.86 0.82 0.78
Cost savingscost arising
Tagging and labelling 1,489,500 1,489,500 1,489,500 1,489,500 1,489,500
Receiving 112,665 112,665 112,665 112,665 112,665
Warehouse management 34,709 34,709 34,709 34,709 34,709
Order picking 69,681 69,681 69,681 69,681 69,681
Shipping 69,780 69,780 69,780 69,780 69,780
Total cost/saving 1,202,665 1,202,665 1,202,665 1,202,665 1,202,665
Amortization 69,000 69,000 69,000 69,000 69,000
Gross margin 1,271,665 1,271,665 1,271,665 1,271,665 1,271,665
Taxes 0 0 0 0 0
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 560
ARTICLE IN PRESS
Table 6 (continued )
Net prot 1,271,665 1,271,665 1,271,665 1,271,665 1,271,665
Investment 379,500 0 0 0 0 0
Net cash ow 379,500 1,202,665 1,202,665 1,202,665 1,202,665 1,202,665
Discounted cash ow 379,500 1,145,396 1,090,853 1,038,908 989,436 942,320
Net present value (h) 5,586,412
Payback period (years) 45
Internal rate of return (%)
ROI (%)
Distributors DC Years
0 1 2 3 4 5
Discount rate 1 0.95 0.91 0.86 0.82 0.78
Cost savingscost arising
Tagging and labelling
Receiving 79,645 79,645 79,645 79,645 79,645
Warehouse management 52,323 52,323 52,323 52,323 52,323
Order picking 148,259 148,259 148,259 148,259 148,259
Shipping 42,000 42,000 42,000 42,000 42,000
Total cost/saving 322,227 322,227 322,227 322,227 322,227
Amortization 129,000 129,000 129,000 129,000 129,000
Gross margin 193,227 193,227 193,227 193,227 193,227
Taxes 65,697 65,697 65,697 65,697 65,697
Net prot 127,530 127,530 127,530 127,530 127,530
Investment 709,500 0 0 0 0 0
Net cash ow 709,500 256,530 256,530 256,530 256,530 256,530
Discounted cash ow 709,500 244,314 232,680 221,600 211,048 200,998
Net present value (h) 401,139
Payback period (years) 3.052
Internal rate of return (%) 23.70%
ROI (%) 24%
Retailer Years
0 1 2 3 4 5
Discount rate 1 0.95 0.91 0.86 0.82 0.78
Cost savingscost arising
Tagging and labelling
Receiving 10,452 10,452 10,452 10,452 10,452
Putaway 29,880 29,880 29,880 29,880 29,880
Order picking 96,222 96,222 96,222 96,222 96,222
Shipping 10,452 10,452 10,452 10,452 10,452
Total cost/saving 136,554 136,554 136,554 136,554 136,554
Amortization 27,720 27,720 27,720 27,720 27,720
Gross margin 108,834 108,834 108,834 108,834 108,834
Taxes 37,004 37,004 37,004 37,004 37,004
Net prot 71,830 71,830 71,830 71,830 71,830
Investment 152,460 0 0 0 0 0
Net cash ow 152,460 99,550 99,550 99,550 99,550 99,550
Discounted cash ow 152,460 94,810 90,295 85,995 81,900 78,000
Net present value (h) 278,541
Payback period (years) 1.638
Internal rate of return (%) 58.80%
ROI (%) 62%
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 561
For manufacturers DC, high costs arise due to
investments for purchasing wider quantity of RFID
tags, required to tag each case entering the supply
chain. As can be seen from the high negative NPV,
resulting benets do not compensate such costs.
Nonetheless, picking operations now benet from
substantial savings, which could not be achieved
under a pallet-level tagging scenario. Specically,
reduction of time required to perform picking
activities and related checks, and shrink abatement,
represent the main components of total savings,
accounting for 17%, 58% and 19%, respectively.
Additional savings, whose importance is substan-
tially smaller, derive from the removal of logistics
labels, which are now avoided and substituted by
RFID tags.
An opposite situation occurs for the distributors
DC, which benets from an NPV of about
h400,000. This result can be justied considering
that no expenses for RFID tags should be sustained
by the distributor; thus, cost savings widely over-
come costs arising for this player. In case-level
tagging scenario, distributors DC can also achieve
higher savings than in the previous one. Specically,
additional savings emerge in picking operations,
mainly due to time shortening for picking cycles and
checking activities performed on mixed pallets.
These elements account for about 70% and 25%
of the total cost savings, while lower benets (1%)
can be achieved thanks to the improved accuracy of
picking operations, due to reduction of errors of
product quantity/type in picking pallets. Such errors
could not be identied nor rectied in a pallet-level
tagging scenario.
Finally, for the retail store, the resulting NPV is
about h280,000. By comparing Tables 5 and 6, no
substantial changes are observed in outcomes
related to receiving and putaway processes.
Conversely, order picking and shipping are
now included in the analysis, since these pro-
cesses are affected by both investments and savings
when case-level tagging is adopted. Specically,
investments are related to RFID gates to be
installed for real-time monitoring the number of
cases moved from the backroom to the shop oor,
while savings mainly derive from reducing the
time required for inventory replenishment in the
shop oor as well as in the backroom. Additional
benets can be achieved from shrink abatement,
thanks to the real-time visibility of inventory and
possible promotion on products close to expira-
tion dates.
Results of the non-integrated scenario allow
concluding that the case-level tagging is not cost
competitive for all supply-chain players; in parti-
cular, manufacturers should sustain the higher costs
of implementation. Since costs are mainly due to
purchasing RFID tags, it can be concluded that
10 ch tags cost is still too high to enable the
adoption of case-level tagging. A viable break
even cost of tags, that is the unitary cost of tags
allowing the manufacturers investment to be paid
back in 5 years, has thus been estimated. As
mentioned, the initial investment is a function of
the unitary cost of tags, as well as the whole NPV,
according to Eq. (1). The break even cost of tag
has thus been computed by imposing the NPV of
the manufacturers DC to score zero at the end of
the 5-year period hypothesized, as shown in the
following equation:
break even cost tag cost
X
n
j0
DCF
j

0
( )
. (4)
Such cost has been estimated in about 1.45 ch for
the non-integrated scenario.
4.2. Integrated scenario
In the integrated scenario, data sharing and
real-time availability through EPCglobal network
make additional benets possible, namely:
Reduction of safety stocks for each supply-chain
player, due to the improved visibility of inven-
tories brought in by the EPCglobal network in
the whole supply chain, and thus to the reduction
of the bullwhip effect. As a consequence, lower
costs of holding stocks result.
Reduction of stock-outs at the retail store. Both
manufacturers and distributors can benet from
such a reduction, mainly in terms of increasing
turnover and improved service level to custo-
mers, respectively.
To compute savings in safety stocks, we followed
the analytical approach proposed by Chen et al.
(2000), which have quantied the bullwhip effect
reduction when comparing two opposite situations,
namely sharing of demand data between supply-
chain partners and absence of information sharing.
Applying the mathematical expressions provided by
the authors to compare the non-integrated and
ARTICLE IN PRESS
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 562
the integrated scenarios, results for manufacturers
DC show a reduction of demand standard deviation
of 46.54% and 55.42% for pallet- and case-level
tagging, respectively. These results have been
obtained assuming an average procurement lead
time of 10 days and a mobile mean interval of 5
days, which have been obtained from the data
collection phase. For distributors DC, decrease in
demand standard deviation is less appreciable: no
reduction results in the pallet-level tagging scenario,
while about 13.45% of reduction in demand
standard deviation has been computed in the case-
level tagging scenario. As a matter of fact, RFID
and EPC implementation cannot improve inventory
visibility in distributors DCs when adopting a
pallet-level tagging: since information related to
retailers orders is currently available, no additional
data are provided exploiting pallet-level tagging.
Numerical results have been obtained starting from
characteristics of the representative FMCG sup-
ply chain, namely average procurement lead times
of 1 day for the retailer and 4.5 days for the
distributors DC, and mobile mean interval of 5
days both players.
To quantify the economical impact of safety
stocks reduction, cost of holding stocks has been
also computed, starting from data collected during
the survey phase. For manufacturers DC these
costs are 143.47 h/year and pallet, while for dis-
tributors DC results show a cost of stocks of
136.51 h/year and pallet. The main reason for this
difference has to be found in the higher average
economical value of pallets for the manufacturer,
and in the higher inventory turn rate for the
distributors DC.
As a result, for manufacturers DC, benets
achievable in terms of safety stocks reduction
account for 267,095 and 318,043 h/year in the pallet-
and case-level tagging scenario, respectively. Nu-
merical results have been obtained starting from
average safety stocks of about 4000 pallets, which
represents 25% of the average stock, as emerged
from the representative FMCG supply chain.
Similarly, for distributors DC, benets achievable
in terms of safety stocks reduction have been
computed starting from average safety stocks of
about 3500 pallets, which represent 12.5% of the
average stock. Results, for the case-level tagging
scenario, show that the reduction of safety stocks
can provide 64,272 h/year of savings. Those benets
have been added as savings when computing
the annual NCF of each player. Tables 7 and 8
summarize the results obtained in pallet- and case-
level tagging, respectively.
As can be seen from Table 7, in a pallet-level
tagging, improvements in NPV are observed for all
supply-chain players. More specically, manufac-
turers DC benet from savings, resulting from
decrease in safety stocks, which allow reaching
h980,000 NPV. Conversely, distributors DC and
retailer benet from additional savings in receiving
operations, due to the real-time availability of data
on pallets received, which allow reducing the time
required for checking and acceptance operations.
In the case-level tagging scenario (see Table 8),
the additional benets of safety stock reduction do
not compensate the high costs of tags for manu-
facturers DC, providing a high negative NPV.
Applying Eq. (4) to the numerical results obtained,
the break even cost of tags has been estimated as
about 3.5 ch for this scenario. Conversely, both
distributors DC and retailer can benet from
positive NPV of about h1,220,000 and h370,000,
respectively. Again, both players benet from
savings in receiving operations, while distribu-
tors DC can gain additional savings from safety
stocks reduction and from decrease in time required
to perform picking operations.
Moreover, starting from data computed for the
representative FMCG supply chain, effects re-
sulting from the reduction of stock-out at the retail
store have been quantied for both manufacturers
and distributors DCs. Such benets, however, have
not been taken into account in the integrated
scenario, due to a twofold reason. First, they have
been quantied in terms of additional annual
turnover, but no information is available on their
prot margin on the annual NCF. Moreover, the
annual turnover refers to the manufacturing com-
pany as a whole, while the costs/benets analysis is
focused on a specic DC/retail store; thus, those
costs are not directly comparable.
For manufacturers DC, stock-out costs per
single product sold at the retail store were estimated
as the current turnover loss, turnover loss
c
(h/item),
resulting from products lack at the retail store,
according to the following equation:
turnover loss
c
case value current stock out
number of retail stores number of references:
5
As a result of the questionnaire survey, only the
average value of case and the number of references
ARTICLE IN PRESS
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 563
handled by the DC were available. Thus, the
turnover loss has been expressed as a function of
the number of retail stores in the distribution
channel of the manufacturer and the current level
of stock-out at the retail store. Benets resulting
from stock-out reductions have been computed
ARTICLE IN PRESS
Table 7
Results of investment evaluation for the FMCG supply chain in integrated scenario with pallet level tagging
Manufacturers DC Years
0 1 2 3 4 5
Discount rate 1 0.95 0.91 0.86 0.82 0.78
Cost savings cost arising
Tagging and labeling 45,000 45,000 45,000, 45,000, 45,000,
Receiving 112,665 112,665 112,665, 112,665, 112,665,
Warehouse management 34,709 34,709 34,709, 34,709, 34,709,
Order picking 12,500 12,500 12,500, 12,500, 12,500,
Shipping 69,780 69,780 69,780, 69,780, 69,780,
Safety stock reduction 267,095 267,095 267,095, 267,095, 267,095,
Total cost/saving 426,749 426,749 426,749 426,749 426,749
Amortization 59,000 59,000 59,000 59,000 59,000
Gross margin 367,749 367,749 367,749 367,749 367,749
Taxes 125,035 125,035 125,035 125,035 125,035
Net prot 242,715 242,715 242,715 242,715 242,715
Investment 324,500 0 0 0 0 0
Net cash ow 324,500 301,715 301,715 301,715 301,715 301,715
Discounted cash ow 324,500 287,347 273,664 260,632 248,221 236,401
Net present value (h) 981,766
Payback period (years) 1.136
Internal rate of return (%) 89.20%
ROI (%) 98%
Distributors DC Years
0 1 2 3 4 5
Discount rate 1 0.95 0.91 0.86 0.82 0.78
Cost savings cost arising
Tagging and labelling
Receiving 304,645 304,645 304,645 304,645 304,645
Warehouse management 52,323 52,323 52,323 52,323 52,323
Order picking 31,500 31,500 31,500 31,500 31,500
Shipping 42,000 42,000 42,000 42,000 42,000
Safety stock reduction
Total cost/saving 367,468 367,468 367,468 367,468 367,468
Amortization 91,000 91,000 91,000 91,000 91,000
Gross margin 276,468 276,468 276,468 276,468 276,468
Taxes 93,999 93,999 93,999 93,999 93,999
Net prot 182,469 182,469 182,469 182,469 182,469
Investment 500,500 0 0 0 0 0
Net cash ow 500,500 273,469 273,469 273,469 273,469 273,469
Discounted cash ow 500,500 260,446 248,044 236,233 224,983 214,270
Net present value (h) 683,476
Payback period (years) 1.968
Internal rate of return (%) 46.60%
ROI (%) 48%
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 564
ARTICLE IN PRESS
Retailer Years
0 1 2 3 4 5
Discount rate 1 0.95 0.91 0.86 0.82 0.78
Cost savings cost arising
Tagging and labeling
Receiving 44,673 44,673 44,673 44,673 44,673
Warehouse management 29,880 29,880 29,880 29,880 29,880
Order picking
Shipping
Safety stock reduction
Total cost/saving 74,553 74,553 74,553 74,553 74,553
Amortization 25,720 25,720 25,720 25,720 25,720
Gross margin 48,833 48,833 48,833 48,833 48,833
Taxes 16,603 16,603 16,603 16,603 16,603
Net prot 32,230 32,230 32,230 32,230 32,230
Investment 141,460 0 0 0 0 0
Net cash ow 141,460 57,950 57,950 57,950 57,950 57,950
Discounted cash ow 141,460 55,190 52,562 50,059 47,676 45,405
Net present value (h) 109,433
Payback period (years) 2.673
Internal rate of return (%) 29.90%
ROI (%) 30%
Table 8
Results of investment evaluation for the FMCG supply chain in integrated scenario with case level tagging
Manufacturers DC Years
0 1 2 3 4 5
Discount rate 1 0.95 0.91 0.86 0.82 0.78
Cost savingscost arising
Tagging and labelling 1,489,500 1,489,500 1,489,500 1,489,500 1,489,500
Receiving 112,665 112,665 112,665 112,665 112,665
Warehouse management 34,709 34,709 34,709 34,709 34,709
Order picking 69,681 69,681 69,681 69,681 69,681
Shipping 69,780 69,780 69,780 69,780 69,780
Safety stock reduction 318,043 318,043 318,043 318,043 318,043
Total cost/saving 884,622 884,622 884,622 884,622 884,622
Amortization 69,000 69,000 69,000 69,000 69,000
Gross margin 953,622 953,622 953,622 953,622 953,622
Taxes 0 0 0 0 0
Net prot 953,622 953,622 953,622 953,622 953,622
Investment 379,500 0 0 0 0 0
Net cash ow 379,500 884,622 884,622 884,622 884,622 884,622
Discounted cash ow 379,500 842,497 802,379 764,170 727,781 693,125
Net present value (h) 4,209,452
Payback period (years) 45
Internal rate of return (%)
ROI (%)
Table 7 (continued )
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 565
assuming stock- out at the retail store to be
improved by 16%, according to the ndings of a
study carried out by the University of Arkansas on
RFID applications on Wal Mart distributive
channel (RFID Journal, October 14, 2005). The
following set of equations was applied for the
ARTICLE IN PRESS
Distributors DC Years
0 1 2 3 4 5
Discount rate 1 0.95 0.91 0.86 0.82 0.78
Cost savingscost arising
Tagging and labelling
Receiving 304,645 304,645 304,645 304,645 304,645
Warehouse management 52,323 52,323 52,323 52,323 52,323
Order picking 148,259 148,259 148,259 148,259 148,259
Shipping 42,000 42,000 42,000 42,000 42,000
Safety stock reduction 64,272 64,272 64,272 64,272 64,272
Total cost/saving 611,499 611,499 611,499 611,499 611,499
Amortization 129,000 129,000 129,000 129,000 129,000
Gross margin 482,499 482,499 482,499 482,499 482,499
Taxes 164,050 164,050 164,050 164,050 164,050
Net prot 318,449 318,449 318,449 318,449 318,449
Investment 709,500 0 0 0 0 0
Net cash ow 709,500 447,449 447,449 447,449 447,449 447,449
Discounted cash ow 709,500 426,142 405,849 386,523 368,117 350,588
Net present value (h) 1,227,720
Payback period (years) 1.698
Internal rate of return (%) 56.30%
ROI (%) 59%
Retailer Years
0 1 2 3 4 5
Discount rate 1 0.95 0.91 0.86 0.82 0.78
Cost savingscost arising
Tagging and labelling
Receiving 44,673 44,673 44,673 44,673 44,673
Warehouse management 29,880 29,880 29,880 29,880 29,880
Order picking 96,222 96,222 96,222 96,222 96,222
Shipping
Safety stock reduction
Total cost/saving 170,775 170,775 170,775 170,775 170,775
Amortization 27,720 27,720 27,720 27,720 27,720
Gross margin 143,055 143,055 143,055 143,055 143,055
Taxes 48,639 48,639 48,639 48,639 48,639
Net prot 94,416 94,416 94,416 94,416 94,416
Investment 152,460 0 0 0 0 0
Net cash ow 152,460 122,136 122,136 122,136 122,136 122,136
Discounted cash ow 152,460 116,320 110,781 105,506 100,482 95,697
Net present value (h) 376,326
Payback period (years) 1.326
Internal rate of return (%) 75.30%
ROI (%) 81%
Table 8 (continued )
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 566
computation:
benefit of stock out reduction turnover loss
c
turnover loss
e
turnover loss
e
case value expected stock out
number of retail stores number of references
expected stock out current stock out 1 16%. 6
being turnover loss
e
the expected turnover loss when
RFID is adopted. The results obtained for manu-
facturers DC are provided in Fig. 4.
As can be seen from the gure, benets of stock-
out reduction have the potential to signicantly
impact the overall protability of the implementa-
tion. Thus, availability of real-time data becomes a
fundamental issue for manufacturers, to ensure
products to be on the shop oor when required.
EPCglobal network can be regarded as a viable tool
to address this issue, since data available on the
EPCglobal network allow manufacturers to know
precisely what will be needed, when and in which
quantity, preventing products shortage at the retail
store.
A similar computation was performed to quantify
the benets of stock-out reduction for distributors
DC. In this case, turnover loss
c
(h/retail store) at the
retail store has been computed starting from the
service level provided by the retail store and its
annual turnover. For distributors DC, although
stock-out of a product at the shop oor does not
necessarily imply a sell loss, since the customer may
buy an alternative product, we quantify it as a sell
loss because of the poor service level provided. The
following equation has thus been adopted for the
computation:
turnover loss
c
turnover current stock out:
(7)
From data collected, both turnover and the
current stock-out level at the retail store were
directly available. Two different computations were
performed to assess the benets of stock-out
reduction for supermarkets and department stores,
since data can highly vary depending on the type of
retail store examined; for instance, stock-out level
resulted in about 2.75% and 1.5%, respectively, for
supermarkets and department stores. By applying
Eq. (6), benets resulted in about h338,400 and
h38,247 for each department store and supermarket
the distributors DC serves. The overall saving
should be computed taking into account the
distributive network of the distributor, that is, the
number and type of retail stores served.
ARTICLE IN PRESS
0
1
0
0
2
0
0
3
0
0
4
0
0
5
0
0
6
0
0
7
0
0
8
0
0
9
0
0
1
0
0
0
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
6.00%
7.00%
8.00%
9.00%
Number of retail stores served
C
u
r
r
e
n
t

s
t
o
c
k
-
o
u
t
950.00 1000.00
900.00 950.00
850.00 900.00
800.00 850.00
750.00 800.00
700.00 750.00
650.00 700.00
600.00 650.00
550.00 600.00
500.00 550.00
450.00 500.00
400.00 450.00
350.00 400.00
300.00 350.00
250.00 300.00
200.00 250.00
150.00 200.00
100.00 150.00
50.00 100.00
- 50.00
Fig. 4. Economical impact of stock-out reduction (16%) for manufacturers DC as a function of current stock-out level at the retail store
and number of retail stores served, for a single item sold at the retail store.
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 567
On the ground of these results, it can be stated
that substantial benets in terms of stock-out
reductions can be reached in an integrated
scenario for both manufacturers and distribu-
tors DCs. Consequently, the implementation of
RFID, EPC standards and EPCglobal network
emerges as a viable tool for improving integration
of supply-chain partners, allowing real-time data
availability and sharing through the supply chain.
RFID tags can be regarded as a means to enhance
supply-chain integration, since costs of tags can be
balanced by benets achievable, thanks to real-time
data availability, which cannot be provided using
traditional barcodes. Specically, outcomes of the
integrated scenario show that benets of reduced
stock-out and safety stocks could overcome costs
for purchasing RFID tags for manufacturers DC,
once tag costs should fall below the estimated
break-even points. This outcome may provide an
economical justication to the adoption of the
RFID technology for FMCG supply-chain partners
in the case-level tagging.
5. Conclusions
The nal results of the research described in this
paper are provided in Table 9.
As a result of the investment analysis, the
following conclusions can be drawn. In a non-
integrated scenario, the introduction of RFID in a
pallet-level tagging scenario provides positive NPV
for all echelons examined. Manufacturers DCs, in
particular, are affected by the highest benets.
Conversely, the case-level tagging scenario generates
opposite results. Specically, substantial costs arise
for manufacturers DC, mainly due to tags costs; as
a result, negative NPV emerges. Conversely, high
benets can be achieved by retail stores and
distributors DC, which do not uphold costs for
purchasing RFID tags. In this scenario, the break
even cost of RFID tags has been estimated as
about 1.45 ch/tag.
When an integrated perspective is examined,
visibility due to EPCglobal network and data
sharing bring in additional benets for both
manufacturer and distributor. Benets are related
to safety stock reduction at the DC as a conse-
quence of the attening of the bullwhip effect, and
to turnover improvement because of stock-out
reduction.
In the case-level tagging scenario, benets differ
depending on the supply-chain player examined.
For manufacturers DC, benets achievable in
terms of safety stocks reduction do not compensate
the high cost of tagging; as a result, high negative
NPV emerges. Conversely, both distributors DC
and retailer experience increased benets, which
provide improvements in the nal NPV.
Although prots and losses are not balanced for
manufacturers DC, numerical results show that
benets obtained in terms of safety stocks reduction
have potential to provide an economical justication
ARTICLE IN PRESS
Table 9
Final results of the research
Manufacturers DC Distributors DC Retailer
Non integrated scenario Pallet level tagging NPV (h) h218,554 h40,549 h11,648
IRR (%) 26.90% 7.90% 8.00%
PBP (years) 2.84 4.59 4.58
ROI (%) 26.86% 8.90% 8.94%
Case level tagging NPV (h) h5,586,412 h401,139 h278,541
IRR (%) 0.00% 23.70% 58.80%
PBP (years) 45 3.05 1.64
ROI (%) 23.58% 61.81%
Integrated scenario Pallet level tagging NPV (h) h981,766 h683,476 h109,433
IRR (%) 89.20% 46.60% 29.90%
PBP (years) 1.14 1.97 2.67
ROI (%) 98.13% 47.83% 29.89%
Case level tagging NPV (h) h4,209,452 h1,227,720 h376,326
IRR (%) 0.00% 56.30% 75.30%
PBP (years) 45 1.70 1.33
ROI (%) 58.89% 81.25%
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 568
to the introduction of RFID and EPC even in the
case-level tagging scenario. More specically, a tag
cost of about 3.5 ch could balance prots and losses
for the manufacturers DC.
While all benets related to RFID and EPCglobal
network have been computed on the ground of
existing literature, future researches will be ad-
dressed to the development of an RFID Lab where
all supply-chain processes mapped and reengineered
in this research will be full-scale replicated. Experi-
mental campaigns will determine how and to what
extent potential saving for each process and for each
supply-chain partner could be achieved. When this
paper was submitted, the project of a 150 m
2
facility
was going to be carried out and potential stake-
holders were recruited.
Acknowledgements
Our thanks goes to GS1 Italy, which has
supported this research, and to Auchan, Campari,
Carrefour, Finiper, Heineken, LOre al, Nestle ,
Nordiconad, Pam, Procter&Gamble, Sony for the
paramount collaboration and commitment to this
research.
References
Agarwal, V., 2001. Assessing the Benets of Auto-ID Technology
in the Consumer Goods Industry. Cambridge University,
Auto-ID Center. Available at /http://www.autoidlabs.org.
uk/camwhitepapers.htmlS, accessed July 2005.
Auto-ID Center, 2002. Technology Guide. Cambridge Uni-
versity, Auto-ID Center. Available at /http://interval.hu-berlin.
de/downloads/rd/technologische%20grundlagen/Technology_
Guide.pdfS, accessed July 2005.
Boxall, G., 2000. The use of RFID for retail supply chain
logistics. Paper presented at the Tag 2000, Baltic Conven-
tions. The Commonwealth Conference & Events Centre,
London, 24 May.
Burnell, J., 1999. Users will overcome many obstacles and imple-
ment RFID, study predicts. Automatic ID News 15 (5), 26.
Bushnell, R., 2000. RFIDs wide range of possibilities. Modern
Materials Handling 55 (1), 37.
Bylinsky, G., 2000. Hot new technologies for American factories.
Fortune 142 (1), 288A288K.
Byrne, P.J., Heavey, C., 2006. The impact of information sharing
and forecasting in capacitated industrial supply chains: A case
study. International Journal of Production Economics 103,
420437.
Chen, F., Drezner, Z., Ryan, J.K., Simchi-Levi, D., 2000.
Quantifying the bullwhip effect in a simple supply chain:
The impact of forecasting, lead time, and information.
Management Science 46 (3), 436443.
EPC Global, 2004. The EPCglobal Network
TM
: overview of
design, benets and security. Available at /http://www.
epcglobalinc.orgS, accessed July 2006.
Fernie, J., 1994. Quick response: An international perspective.
International Journal of Physical Distribution & Logistics
Management 24 (6), 3846.
Hammer, M., Champy, C., 1993. Reengineering the Corporation.
Harper Business, New York.
Harrison, M., 2004. EPC Information Service. In: Proceeding of
the Auto-ID Labs Research Workshop, September 2324.
University of St. Gallen, Switzerland. Available at /http://
www.m-lab.ch/auto-idS, accessed July 2006.
Jones, L., 1999. Working without wires. Industrial Distribution
88 (8), M6M9.
Jones, P., Clarke-Hill, C., Shears, P., Comfort, D., Hillier, D.,
2004. Radio frequency identication in the UK: Opportu-
nities and challenges. International Journal of Retail and
Distribution Management 32 (3), 164171.
Karkkainen, M., 2003. Increasing efciency in the supply chain
for short shelf life goods using RFID tagging. International
Journal of Retail and Distribution Management 31 (3),
529536.
Karkkainen, M., Holmstrom, J., 2002. Wireless product identi-
cation: Enabler for handling efciency, customization and
information sharing. Supply Chain Management: An Inter-
national Journal 7 (4), 242252.
McFarlane, D., 2002. Auto-ID based control. White Paper,
Cambridge University, Auto-ID Center. Available at /http://
www.autoidlabs.org.uk/camwhitepapers.htmlS, accessed July
2005.
Moore, B., 1999. Barcode or RFID: Which will win the high
speed sortation race? Automatic ID News 15 (7), 2936.
Niemeyer, A., Pak, M.H., Ramaswamy, S.E., 2003. Smart tag for
your supply chain. The McKinsey Quarterly 4, 68.
Ollivier, M., 1995. RFID enhances materials handling. Sensor
Review 15 (1), 3639.
Prater, E., Frazier, G.V., Reyes, P.M., 2005. Future impacts of
RFID on e-supply chains in grocery retailing. Supply Chain
Management: An International Journal 10 (2), 134142.
RFID Journal, May 2005. Understanding the EPC Gen 2
Protocol, Special Report.
RFID Journal, October 14, 2005. EPC reduces out-of-stocks at
Wal Mart.
RFID Journal, September 15, 2005. Alien Drops Tag Price to
12.9 Cents.
Riso, F., 2001. Intelligent chip technologyan RFID update.
Paper presented at the Intelligent Supply Chain, CIES, The
Food Business Forum, Hotel Okura, Amsterdam, 45
October.
UCCnet, 2005. The Global Data Synchronization Network
(GDSN). Available at /http://www.uccnet.org/GDSN/
GDSN_2-9-05.pdfS, accessed January 2006.
ARTICLE IN PRESS
E. Bottani, A. Rizzi / Int. J. Production Economics 112 (2008) 548569 569