Discontinued Products

An Empirical Study of Service Parts Management

Luís Miguel D. F. Ferreira

, Amílcar Arantes

and Cristóvão Silva

Department of Mechanical Engineering, University of Coimbra, Polo II Pinhal de Marrocos, 3030 Coimbra, Portugal

CERIS, CESUR, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, Lisboa,1049-001, Portugal

Keywords: Service Parts, Discontinued Products, Empirical Study.

Abstract: The procurement and inventory management of service parts for discontinued products has often been

overlooked by companies, resulting in several problems such as stock outs, rush orders or obsolete stocks.

Accordingly, the main aim of this work is to develop a methodology to deal with these issues following

product discontinuation. To this end, an empirical study – based on action research principles – was carried

out in a producer of household appliances, which is bound by law to provide service parts for its products for

a period of 15 years after they have been discontinued. The work was developed in three stages:

characterization of the company situation; definition of a procedure to eliminate obsolete stocks; and

definition of a procedure to manage active service parts. The resulting methodology and respective procedures

are presented and the results obtained with the implementation are discussed.

1 INTRODUCTION

Rapid technological innovation and recurrent

changes in consumer preferences are decreasing

product lifecycles. This places pressure on stock

management, with a requirement for suitable stock

levels for all service parts. Service parts are used to

replace old parts that are no longer operational due to

total failure or malfunction. Moreover, after a product

has been discontinued, many of the service parts

needed in the post product life cycle are often out of

production (Inderfurth and Mukherjee, 2006).

Service parts for products like household

appliances or automobiles are considered an

important element of a company´s business. In some

industry sectors the service parts business can

represent up to 25% of the revenues and 40% to 50%

of the profits of manufacturing firms (Dennis and

Kambil, 2003; Cohen et al. 1999). Thus, it is

important for industrial companies to guarantee the

availability of service parts in order to provide the

desired after-sales service level. As a result,

companies are forced to stock an enormous amount

of service parts.

It is often the case that many service parts for

discontinued products are recognized as a major

source of inventory stock-out or obsolescence. The

inventory costs associated with service parts for

discontinued products are much higher than those of

service parts for current products. Moreover, the level

of competition that exists in the market means that

any stock-out of service parts cannot be tolerated,

since this has a negative impact on the brand image

of the company (Hong et al., 2008).

Therefore, it is important to manage service parts

carefully, because production lines used for

manufacturing a particular service part are likely to

be discontinued, prior to demand falling to zero.

Moreover, in some countries, manufacturers are

required by law to provide past model service parts

for several years after production has ceased

. Yet,

procurement and inventory management of service

parts are complex subjects due to the high number of

service parts involved; the intermittent nature of their

demand patterns; the high responsiveness needed to

minimise the downtime cost for the customer; and the

high risk of stock obsolescence (Hong et al., 2008).

Hence, the main aim of this work is to develop a

methodology to help manage these service part issues

in the period following product discontinuation,

namely: eliminating obsolete stock and managing

active service parts.

Miguel D. F. Ferreira L., Arantes A. and Silva C.

Discontinued Products - An Empirical Study of Service Parts Management.

DOI: 10.5220/0006185800660074

In Proceedings of the 6th International Conference on Operations Research and Enterprise Systems (ICORES 2017), pages 66-74

ISBN: 978-989-758-218-9

2 LITERATURE REVIEW

After-sales service is a period following on from the

start of the product lifecycle and before the end of life

date. When after-sales service ends, the manufacturer

no longer guarantees the supply of replacement parts

for the product. This period may be a legal

requirement, or it may be set by the company. In some

cases, it can go beyond the legal limit, in a bid to

increase the service level and improve the company´s

image. As it is normal that replacement parts are

needed after some period of use, their demand peaks

sometime after demand for the product itself. This

can be seen in Figure 1 below. While this is normally

the case, there may be exceptions to this rule.

The end of after-sales service may also occur

unexpectedly, severely complicating the management

of service parts. Where a machine is dependent on a

part to continue functioning, a stock out of that part

will make the machine obsolete. This means that all

the other parts of that machine reach the end of the

after-sales service period earlier than expected.

Another source of uncertainty occurs when one part

is substituted by another. An example may be when a

firm manufacturing mobile phone batteries brings a

new battery to market which is cheaper and lasts

longer; when the consumer changes battery, they will

likely choose the new model over the old model. This

source of uncertainty is difficult to predict (Hong et

al., 2005).

There is a considerable body of literature

concerning demand forecasting and inventory

management of service parts. Several documented

approaches exist, ranging from relatively simple

models like Croston’s method (Croston, 1972) and its

refinement (Syntetos and Boylan, 2001; Syntetos and

Boylan, 2005) to more complex algorithms like the

stochastic forecasting model presented by Hong et al

(2008). This last method considers four major factors

in their model to forecast service parts: product sales,

the discard rate of the product, the failure rate of the

service part and the replacement probability of the

service part.

Wang and Syntetos (2011) presented an

innovative idea to forecast demand for spare parts that

relies upon the demand generation process itself,

comparing it successfully with a traditional time-

series method. Leifker et al. (2014) presented two

approaches for dealing with the problem of extending

maintenance or supply contracts for spare parts of

discontinued products: one includes the use of a

continuous-time dynamic program and the other

makes use of a two-stage stochastic algorithm. Rego

and Mesquita (2015) presented a case study on spare

parts inventory management, comparing several

methods using different forecasting techniques and

inventory management policies by simulating with

field data (10 032 spare parts references); results of

the simulations allowed the recommendation of best

Figure 1: The lifecycle for spare parts. [Adapted from Inderfurth and Mukherjee (2006)].

Begin of lifetime

Manufacturing begins

Initial load order

End-of-life

Manufacturing ends

Final order

End-of-service (EOS)

All warranties have ceased

Possible disposal of stock

Product in Market within

Sales of products

Demand for spare parts

Discontinued Products - An Empirical Study of Service Parts Management

policies to be followed within each spare part

category. Some authors have proposed practical

approaches to manage service parts for discontinued

products in practice. Teunter and Fortuin (1998)

present a model developed to determine the size of a

final order for service parts which was implemented

in an electronic equipment company, with the

objective of covering the demand until all service

obligation has ended. Teunter and Haneveld (1998)

refer to a case study of a company that produces

appliances for which they developed a model for

service parts management in its final phase. The

model considers an ordering policy which consists of

an initial order up-to-level at time 0, the instant when

the product is discontinued, and a subsequent series

of decreasing order-up-to levels for various intervals

of the planning horizon.

However, previous research has focused primarily

on the planning and operational aspects (e.g. the

determination of optimum spare parts inventory

levels) and has neglected the strategic and

organisational problems manufacturing companies

must solve to manage their spare parts business

effectively during post product life cycle. Moreover,

a gap between research and practice has been

identified by several authors; for a comprehensive

literature review on service parts management and on

the gap between research and practice we refer to

Bacchetti and Saccani (2012). Here, for instance, the

authors make the point that “despite the wealth of

literature on the subject, no attention has in practice

been paid to proper management and control of

service-parts inventory” or that “incremental

mathematical inventory research is not likely to

enhance practice

”.

This gap between research and practice is

probably due to the mathematical complexity of the

proposed methods and their need for data which are

often not available. It is also important to mention that

the adoption of simple but formalised procedures for

the management of service parts during the post

product life cycle can help companies achieve

substantial benefits, not only by reducing costs but

also by improving the company image (Botter and

Fortuin, 2000).

3 RESEARCH METHOD

In this paper, the case of a company which produces

household appliances is presented. In the past, the

procurement and inventory management of service

parts during the post product life cycle had been

overlooked by the company. This resulted in several

problems such as stock outs, rush orders and obsolete

stock – stock of service parts for which the service

contract has expired. Moreover, sometimes service

part production also required components for which

the original supplier was no longer operating and this

would imply a time-consuming negotiation process

with a new supplier.

To solve these problems, the company decided to

start a project to define procedures to manage the

procurement and inventory management of service

parts during the post product life cycle. Moreover, it

was the intention of the company managers that the

procedures to be implemented should avoid using

complex mathematical algorithms that users would

find hard to understand as well as models that would

require data which would not be easily accessible, or

even available, from the company information

systems.

Therefore, this study applied the principles of

action research, allowing a link to be established

between the company and the researchers (Middel et

al. 2005). Nevertheless, two conditions need to be

respected for the approach taken to fall within the

field of action research. Firstly, the research objective

and project plan were driven by the researcher’s

agenda rather than by the participating company

representatives. Secondly, the project plan was

motivated primarily by the development of

procedures for the management of service parts

during the post product life cycle and not by the aim

of transforming the individual organization’s

practices. In any case, the focus of the research is to

introduce changes in reality (Baker and Jayaraman

2012).

4 THE PROPOSAL

In this section, we present the proposal for the

management of service parts during the post product

life cycle. The project to define the requested

procedures was developed in three stages, which are

described in the following subsections: the

characterisation of the company service parts; the

definition of a procedure to eliminate obsolete stocks

and the definition of a procedure to manage service

parts.

4.1 Characterisation of the Company

In the case study company, the procurement process

for service parts during the post product life cycle was

as follows: when the stock for a given service part

reaches its reorder level, a production order for this

ICORES 2017 - 6th International Conference on Operations Research and Enterprise Systems

service part is scheduled by the logistics department.

To manufacture the required service part, the

necessary components are taken from stock and these

can also eventually reach the reorder level. If this

happens, an order for the component is placed by the

company purchasing department. The size of this

order is calculated to satisfy the estimated demand for

the next six months, based on a simple average of the

demand seen for this component since the time the

product became discontinued and adjusted according

to the conditions of the contract between the company

and the supplier.

This stage of the project was designed to provide

a clear understanding of how many spare parts were

exclusively used in the post product life cycle. The

company information systems held information about

all the components used in the service parts of

discontinued products. However, accessing the

required information was difficult (for example, data

about the time remaining until the end of the service

contracts). In part, this information was incomplete

because each component could be used in several

service parts and each service part may be necessary

for several products. Moreover, the data was spread

over two different information systems (IS 1 and IS 2

in Figure 2).

Therefore, a support database was developed to

easily access all relevant data about the components

of discontinued products. This database was fed with

data from both company information systems (Figure

2), and was used to identify all the necessary

information for the required procurement process

Using data from the first company information

system (IS1), each component (Comp.) was related to

the service parts (SP) where it is used. The second

information system (IS2) shows how each service

part was related to the final products where they are

used, and for each final product (FP) the date of the

last production run was recorded. Furthermore, the

corresponding supplier of each component has also

been identified. The result of this operation is

presented in Table 1.

The information contained in the database

provided a description of the discontinued product

service parts and their respective components. Some

of these characteristics (namely the “Time remaining

to end of service”) allow the parts and their

components to be divided into three major groups: for

Figure 2: Data collection from the company information systems.

Discontinued Products - An Empirical Study of Service Parts Management

Table 1: Component information.

Compo-

nent

Time to end

service(years)

Stock

(units)

Stock

(€)

Supplier

… … … … …

C10 5 1524 480 XYZ

C11 1 149 30 ?

… …

Cn -2 46 202 ABC

… … … … …

the first group, the company is no longer required

to provide support; for the second group, the company

is required to provide parts and components within

the 15-year limit; and for the third group, no end-of-

service date exists (this represents around 50% of the

components listed). The complete listing of Table 1

showed that the company manages approximately

1300 components, exclusively used in service parts

for discontinued products.

These 1300 components are required in around

1800 service parts for discontinued products (there

are more service parts than components as each

component can be used in various parts). In turn,

these parts are used in 6530 end products

(appliances). The 1300 components represent 14% of

the total number managed by the company and 9% of

the company stock value. Three percent of these

components representing obsolete stock. Another

conclusion from analysing Table 1 was that

approximately 50% of the components had no active

supplier assigned.

This represents a major difficulty for the

management and procurement of those components,

because if a stock-out of one of these components

occurred, the company would have to start a

negotiation process with a new supplier, meaning

long lead times. The lack of an active supplier could

be due to two reasons: the usual component supplier

has ceased trading or a long time has passed since the

company has placed an order with the supplier who,

due to production changes, is not able or not willing

to supply that component anymore.

4.2 A Procedure to Eliminate Obsolete

Component Stocks

The previous project stage identified some obsolete

components stocks. A procedure to eliminate these

stocks was required. The stock of these components

could simply be sold as scrap; nevertheless, they

could still be valuable as components for the

production of service parts of discontinued products,

despite the fact that the company has no obligation to

do so.

In the past, if the company received an order for a

service part which was no longer active (covered

under the post product life cycle period) and for

which there was no available stock, the client was

informed that the company was no longer required to

provide this service part and the stock out was not

considered a service-level failure by the company.

Nevertheless, it is clear that this situation could reflect

badly on the company. To avoid these kinds of

problems a procedure to eliminate obsolete stocks

was defined, as shown in Figure 3.

In this situation, clients are informed which

service parts have reached the end-of-service period.

Clients are informed that if they wish they may place

a final order for these service parts. If the stock

components for the of end of life service parts is

sufficient to cover the client’s final order, the order is

delivered. If the stock of components is not sufficient

to cope with the service part production, an order for

the required components is placed. After the final

client orders, if some service parts or components

remain in stock, they are then sold as scrap. This

procedure is now scheduled for the beginning of each

year for all service parts which have reached their

end-of-life date the previous year.

This procedure allows obsolete stock to be

removed, including both spare parts and the

components which they alone use, given the direct

link between the two. In addition, the warehouse

space allocated to these parts and components

becomes available for other stock. Finally, and most

importantly, this approach ensures customer

satisfaction (service level), given that the customer is

warned in advance of the service part’s end-of-life.

They are provided with the opportunity to make a last

order, guaranteeing a fixed quantity and with a lead

time of around one month. It was also decided that

this procedure would be carried out at the start of each

year, covering the service parts which had reached

their end-of-life during the previous year.

4.3 A Procedure to Manage Active

Service Parts

This last stage of the project defined a procedure to

improve the management process of active service

parts. The main objective of managing service parts

is to establish a component ordering procedure which

ensures an adequate inventory level. The availability

of components to produce service parts is vital to the

company. Long lead times are associated with

component orders and this is not acceptable to the end

customer who is waiting for their household

appliance to be repaired.

ICORES 2017 - 6th International Conference on Operations Research and Enterprise Systems

Figure 3: Flowchart showing the procedure to eliminate obsolete stock.

The process for acquiring the components used to

manufacture the service parts for end products which

are still current is relatively simple as this depends on

the demand (for spare parts). However, when there is

intermittent demand for service parts, forecasting

methods are used so that inventory levels can be

properly managed.

In the past, the company used the average service

part demand over the previous six months to forecast

the demand for the next six months, whenever the

component reorder point was reached. The forecast

demand for the next six months would help anticipate

the component procurement process and the service

part production, thus avoiding long lead times for the

final client. In this respect, two distinct situations are

possible: (1) service parts with components for which

there are active suppliers and (2) service parts with

components for which there are no active suppliers.

In both cases demand forecasting is important to

guarantee an effective procurement process and

efficient inventory level management. Additionally,

in the second case forecasting is also important to

help the purchasing department negotiate with

potential new suppliers.

At the beginning of the project the company

considered that the forecasting method used was not

the most effective. Thus, in order to identify which

forecasting method to use for managing service parts

it was decided to study the demand patterns (using

available data from a seven-year period).

The service parts were classified according to

their demand patterns as: slow moving, intermittent,

erratic and lumpy, as proposed by Syntetos et al.

(2005). This classification was made considering the

values calculated using: inter-demand interval (IDI)

and the squared coefficient of variation of the demand

sizes (CV

Considering that the company had asked to avoid

complex algorithms that would be difficult for the

user to understand and considering the available data,

three demand forecasting models were tested: the

Croston Model (Croston, 1972), the SBA model

(Syntetos and Boylan, 2005) and the simple average

in use by the company.

As proposed by the authors, demand for slow

moving components should be forecast by the

Croston method and the others by the SBA model.

Using historical data, the accuracy of these models

was compared with the result obtained using the

simple average procedure in place in the company.

The comparison of the models was made comparing

the forecasting errors, considering the mean absolute

percentage error (MAPE). Table 2 presents the inter-

demand interval (IDI) and the squared coefficient of

Discontinued Products - An Empirical Study of Service Parts Management

variation of the demand sizes (CV

) for three

randomly selected service parts.

The forecasting errors obtained for each

component are presented in Table 3. It is clear that

there is no forecasting method tested which

outperforms the others for all the cases. Thus, it was

decided to keep the simple average procedure,

already in place in the company. This decision is in

accordance with the findings of Sytentos et al. (2015).

Table 2: Demand patterns of three randomly selected

service parts.

SP. 1 SP. 2 SP. 3

IDI 2.35 1.02 1.94

1.25 0.37 2.7

Having identified the forecasting model to be

used, the support database developed during this

project (referred to in the previous section) was used

to elaborate a list of all the components with no

available supplier. For these components, a single

average of the past demand (since time 0) was used to

estimate their future demand till the end of their

service life. This list was sent to the purchasing

department who was responsible for initiating

contacts to find potential suppliers for those

components.

The demand estimate based on past demand

helped the purchasing managers negotiate with the

selected suppliers. To avoid the same problem in the

future (where components have no supplier

available), a simple procedure represented in Figure

4 was implemented in the company.

Table 3: Performance of the forecasting models.

Service Parts Forecasting model MAPE

SBA (α = 0.10) 43%

Average 42%

Croston (α = 0.17) 76%

Average 96%

SBA (α = 0.15) 202%

Average 203%

The database developed during this research

project allowed an historical record to be created,

comprising demand data for the service parts. At the

beginning of each year, the list of components is

checked to identify all components where demand

existed last year. For these components, all suppliers

are contacted to verify if they are still able to deliver

the required component. If not, the purchasing

department will be required to initiate contacts to find

a new supplier.

For the components with available suppliers the

procurement process in place in the company will be

as follows. Whenever the reorder point is attained, an

order to satisfy the demand for the next six months

will be made, considering the simple average of the

past six months’ demand.

5 RESULTS

This project allowed the company to make the

management of spare parts during the post product

life cycle more efficient and reliable, anticipating

some of the problems usually felt. To achieve this

objective, several procedures were developed to

support the procurement decisions for these spare

parts, according to the time period remaining that the

company has to assure the availability of spare parts.

The comparison of forecasting methods showed

that a simple six-month average could produce

adequate results to define the order quantity for

service parts. The procurement and inventory

management procedures for service parts described

were implemented in the company and have been

used on an ongoing basis.

The implemented procedures led to a 4%

reduction in stock outs, raising the service from a

level of 95% to 99%. The improvement essentially

resulted from a set of formal procedures that

guarantee the existence of suppliers for any service

parts components in the post product life cycle. This

avoided extremely long lead times from a negotiation

process which does not meet the expectations of final

clients.

Another result of this project was a 10% reduction

of the inventory value of service part components in.

This was mainly due to eliminating obsolete stock,

which also helped increase warehouse efficiency.

6 CONCLUSION

This paper demonstrates how an action research

project can provide performance improvements in the

management of service parts used in the post product

life cycle. Simple, but formal procedures are

ICORES 2017 - 6th International Conference on Operations Research and Enterprise Systems

Figure 4: Procedure to guarantee the availability of suppliers.

employed to eliminate obsolete stocks and to manage

active service parts. The simplicity of the developed

procedures was important for their acceptance in the

company, since it is well known that most managers

feel uncomfortable if they are unable to understand

the models that support the results. Moreover,

available methods described in the literature often

require data which is not available in most companies.

The procedures implemented during the project

make use of a forecasting method. To choose the

forecasting method three well known models were

tested: the Croston model, the SBA model and a

simple six-month average of past demand. It was

found that no method outperforms the others,

therefore it was decided to maintain the six-month

average already in use in the company.

The implementation of the procedures has

allowed the company to improve the service levels of

service parts for past product life cycle.

The action research model used in the

development of the procedures to manage spare parts

for the post product life cycle contributed both to

achieving practical results in the field and to

developing new knowledge. This was achieved

through the active participation of both researchers

and company employees, which characterises the

action research model.

REFERENCES

Bacchetti, A. and Saccani, N., 2012. Spare parts

classification and demand forecasting for stock control:

Investigating the gap between research and practice.

Omega, vol. 40, no. 6, pp. 722-737.

Baker, T. and Jayaraman, V., 2012. Managing information

and supplies inventory operations in a manufacturing

environment. Part 1: An action research study.

International Journal of Production Research, vol. 50,

no. 6, pp. 1666-1681.

Botter, R and Fortuin, L., 2000. Stocking strategy for

service parts: a case study. International Journal of

Operation and Production Management; vol. 20, pp.

656-74.

Discontinued Products - An Empirical Study of Service Parts Management

Croston, J.D., 1972. Forecasting and stock control for

intermittent demands. Operational Research Quarterly,

vol. 23, pp. 289-303.

Cohen, M.A., Zheng, Y.S. and Wang, Y., 1999. Identifying

opportunities for improving Teradyne’s service-parts

logistics system. Interfaces, vol. 29, no. 4, pp. 1-18.

Dennis, M.J. and Kambil, A., 2003. Service management:

building profits after the sales. Supply Chain

Management Review, vol. 7, no. 1, pp. 42-48.

Hong, J. S., Koo, H., Lee, C. and Ahn, J., 2008. Forecasting

service parts demand for a discontinued product. IIE

Transactions, vol. 40, No. 7, pp. 640-649.

Inderfurth, K. and Mukherjee, K. 2006. Analysis of spare

part acquisition in post product life cycle. Otto von

Guericke University. FEMM Working Paper 2006/006.

Leifker, N.W., Jones, P.C. and Lowe, T.J., 2014.

Determining Optimal Order Amount for End-of-Life

Parts Acquisition with Possibility of Contract

Extension. The Engineering Economist, vol. 59, no. 4,

pp.259–281.

Middel, R., Coghlan, D., Brennan, L. and McNichols, T.,

2006. Action research in collaborative improvement.

International Journal of Technology Management, vol.

33, no. 1, pp. 67-91.

Nagarur, N., Hu, T. and Baid, K., 1994. A computer based

inventory management system for spare parts.

Industrial Management & Data Systems, vol. 94, pp.

22-28.

Rego, J.R. and Mesquita, M.A., 2015. Demand forecasting

and inventory control: A simulation study on

automotive spare parts. International Journal of

Production Economics, vol. 161, pp.1-16.

Syntetos, A. A. and Boylan, J. E., 2001. On the bias of

intermittent demand estimates. International Journal of

Production Economics, vol. 71, no. 1/3, pp. 457-466.

Syntetos, A. A. and Boylan, J. E., 2005. The Accuracy of

Intermittent Demand Estimates. International Journal

of Forecasting, vol. 21, pp. 303-314.

Syntetos, A. A. and Boylan, J. E. and Croston, J. D., 2005.

On the Categorization of Demand Patterns. Journal of

the Operational Research Society, vol. 56, pp. 495-503.

Teunter, R. H. and Fortuin, L., 1998. End-of-life service: A

case study. European Journal of Operational Research,

vol. 107, no. 1, pp. 19-34.

Teunter, R. H. and Fortuin, L., 1999. End-of-life service.

International Journal of Production Economics, vol.

59, pp. 487-497.

Teunter, R.H. and Haneveld, W.K. Klein, 1998. The ‘final

order’ problem. European Journal of Operational

Research, vol. 107, no. 1, pp. 35-44.

Wang, W. and Syntetos, A.A., 2011. Spare parts demand:

Linking forecasting to equipment maintenance.

Transportation Research Part E: Logistics and

Transportation Review, vol. 47, no. 6, pp. 1194-1209.

ICORES 2017 - 6th International Conference on Operations Research and Enterprise Systems