Proposal of an SKU Classification Framework

A Multicriteria Approach

Sara Santos

, Luis Miguel D. F. Ferreira

and Amílcar Arantes

Economics, Management and Industrial Engineering Department, University of Aveiro, Aveiro, 3810-193, Portugal

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

Keywords: SKUs, Classification, Multicriteria ABC Analysis.

Abstract: Changes to an organization’s internal and external environment may cause an increase in the number of

Stock Keeping Units (SKU) in inventory. Therefore an SKU classification and corresponding grouping

become highly important for improving the inventory management process. In this paper we propose a

framework for SKU classification in an industrial context using a multicriteria approach considering three

criteria: value of usage; criticality and demand variability. This approach emphasizes the importance of

SKUs that despite their small value are of vital importance for the operations/production of the organization.

1 INTRODUCTION

Companies cannot ignore the reality of managing a

large number of SKUs. As such, classifying SKUs

can bring significant benefits (van Kampen et al.,

2012).

The biggest challenge for inventory management

that companies face is controllling a large number of

items. This is a very complex task to when using

individual SKUs (Soylu and Akyol, 2014). Grouping

items together makes it easier for managers since the

decisions are taken for a group of SKUs. When

classifying SKUs companies need to have a clear

understanding of the context, and of the aim of the

company inventory management policy.

Bacchetti et al. (2013) mention that the gaps

between theory and practice show that empirical

studies have not been properly validated. As a result,

inventory management solutions that are adequate

for some cases may not be suitable for others. For

that reason, some researchers have suggested that

additonal studies are needed looking at ways of

achieving more integrated solutions.

Questions of how to operationalize an SKU

classification, or the determination of the ideal

number of classes are popular topics in the literature;

moreover, the context of the company is decisive for

the choice of which method to apply (D'Alessandro

and Baveja, 2000; Soylu and Akyol, 2014; van

Kampen et al., 2012).

Another issue that increases the complexity of

inventory management is the fact that reality is

dynamic. This results not only from market changes,

but also internal changes in the organization, with a

consequent impact on stock size which increases the

cost of inventory control activities (Soylu and

Akyol, 2014). Therefore, it is very important that

organizations realize that an efficient SKU

classification could represent an important source of

competitive advantage.

This paper proposes a framework for classifying

SKUs, that can serve as a useful tool in the decision

making process of inventory management. The next

section of the paper presents the framework, while

the final section presents the conclusions and

recommendations for future research.

2 FRAMEWORK/ PROPOSED

APPROACH

In the management of an organization, not every

SKU has the same level of importance. The stock-

out of some SKUs may jeopardize the organization’s

normal production activities; other SKUs, due to

their high value of usage and high demand, require

additional attention by managers.

It is not wise to apply the same inventory

management policy to every SKU; however, it is

also true that managing SKUs individually is a very

complex task (Soylu and Akyol, 2014). As a result,

413

Santos S., Miguel D. F. Ferreira L. and Arantes A..

Proposal of an SKU Classiﬁcation Framework - A Multicriteria Approach.

DOI: 10.5220/0005287004130418

In Proceedings of the International Conference on Operations Research and Enterprise Systems (ICORES-2015), pages 413-418

ISBN: 978-989-758-075-8

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

several researchers proposed different frameworks to

help managers classify SKUs into groups and apply

an adequate inventory management policy to each

group (Bošnjaković, 2010; Cavalieri et al., 2008;

Duchessi et al., 1988; van Kampen et al., 2012).

2.1 Methodology

The main aim of this paper was to develop a

framework for classifying SKUs. The framework

was developed in the context of a company in the

car industry, with the purpose of helping managers

to improve the inventory management process in the

spare parts and consumables warehouse.

In this study we applied the concepts of action

research (Figure 1). This is a method of

collaborative research which may be used to

establish a link between companies and researchers.

Sexton and Lu (2009) define action research as a

“phenomenon-change” (or action) and critical

learning that lead to a change and produce new

knowledge (research) in a social scenario where

researchers and practitioners intervene. By

intervening in the context, the aim is to modify the

scenario by actively participating in the research.

Furthermore the authors say that action research

generates a mutual development of know-that and

know-how.

This choice of method was made as reflection

and co-working were important for assessing the

phenomenon and it was not necessary to control

environmental elements. In any case, the focus of

the research is to introduce changes in reality

(Baker, 2012). Or, we might say that action research

matches theory and practice through a change in a

problematic situation.

Susman and Evered (1978) propose five steps for

leading an action research project, which in the

present case study we define as: 1) Diagnosis, 2)

Criteria definition, 3) SKU classification, 4)

Framework validation, 5) Defining inventory

management policies.

2.2 Diagnosis

The diagnosis is an assessment phase, where the

main goal is to identify the context in which

researchers will intervene, which problems are

relevant and how they could affect the rest of the

organization. This step was carried out over several

meetings with the personnel of the warehouse,

purchasing, maintenance and other groups of

interest.

The researchers concluded that the main problem

for the warehouse is related to the lack of physical

space. Besides that, (unplanned) corrective

maintenance activities raise serious problems for the

supply of spare parts and several stock-outs have

been reported as a consequence.

Figure 1: Framework - action research proposal.

The clients of the warehouse are mostly from the

maintenance area. For this reason, maintenance

personnel are relevant stakeholders in this project.

After several meetings an SKU classification was

suggested. It was observed that a large number of

items are spare parts and it is in the best interest of

the company to define which SKUs should be more

strictly controlled.

2.3 Criteria Definition

The major departments of the company should

participate in the criteria definition step (Sexton and

Lu, 2009) and the classification aim should be

clearly defined (van Kampen et al., 2012). This is a

critical step, and it is not possible to move forward

without collecting all the relevant data from the

company’s activity.

The framework to be presented is built and

shaped for a specific industrial context. However, it

is possible that, with the right modifications (mainly

at the criteria level), this framework could be applied

to another context.

Choosing which criteria to apply is something

that must be discussed and adapted to the

classification context and objectives. Several studies

show that multicriteria approaches are the more

efficient way to assess spare parts and consumables

problems. Bacchetti et al. (2013) proposed a

classification method with six dimensions (life

cycle, lead time, number of orders, demand

frequency, criticality and value). Bošnjaković (2010)

ICORES2015-InternationalConferenceonOperationsResearchandEnterpriseSystems

414

presented a multicriteria framework with value,

demand frequency and criticality. Childerhouse et al.

(2002) built a classification based on life cycle, lead

time window, volume and variability, which is

named DWW

. Flores and Whybark (1986) and

Flores et al. (1992) presented frameworks which

included criticality, lead time and value in the

multicriteria ABC Analysis and proved that such

analysis is a very important tool for improving the

efficiency of inventory management. Ramanathan

(2006) remarks that multicriteria ABC Analysis is a

very effective approach for classifying SKUs,

presenting a framework of linear optimization for

solving the multicriteria problem.

This paper proposes a multicriteria approach

with three criteria, following the recommendations

of Flores et al. (1992). The process ends with the

presentation of a multicriteria ABC Analysis. This is

a very widespread technique which is easily

understood and implemented in organizations.

Managers use this analysis to help understand

which SKUs are more frequently used in the

warehouse and thus need closer monitoring.

However, just looking at the number of spare

parts used makes this analysis very narrow, as many

such items are only used in very specific time

periods. It is thus very important to include other

criteria. As several studies have shown, one of the

most important criteria to consider when analyzing

spares is criticality (Cavalieri et al., 2008;

Huiskonen, 2001; Jouni et al., 2011; Molenaers et

al., 2012), but, because these items are characterized

by an erratic consumption, it is important to verify

demand variability (Heinecke et al., 2013).

Also, while the value of usage and demand

variability are quantitative, the criticality analysis

requires data (such as managers’ tacit knowledge)

which implies a qualitative analysis.

The Framework should then incorporate three

criteria:

 value of usage, with the corresponding ABC

Analysis;

 criticality, with the corresponding VED

Classification;

 demand variability, which will be associated

with the classification HLW (High, Low,

Without Variability).

2.4 SKU Classification

2.4.1 Ranking Demand Value – ABC

Analysis

ABC Analysis shows – with a very high level of

accuracy – which SKUs have most impact on the

company in terms of value.

Cavalieri et al. (2008) says that this analysis is

very important from different perspectives.

Financially it provides data on which investments

should be taken into account depending on whether

they relate to durables or consumables. Logistically

it provides information about whether stock should

be kept for an item or not, or even if the

consumption should be linked to the demand. From

a maintenance perspective it gives the basis for a

balance between the availability of spares and

consumables and the company maintenance policies,

coordinating with purchasing the decisions of

maintenance policies to minimize the effects of

failures.

Bacchetti et al. (2012) state that an SKU has an

important role in the total amount of inventory held.

Bošnjaković (2010) remarks that any SKU has an

associated value and when it is taken from the

warehouse it becomes a cost. So value-usage of an

item is defined as the product of the cost of an SKU

with the annual demand.

An ABC Analysis reveals that only a small

number of items is responsible for the most of the

value. Likewise SKUs are usually classified into 3

groups – group A includes 5% of the items that

represent 75% of value-usage, group C includes

75% of items representing only 5% of value-usage,

the rest of items will be placed in group B, with 20%

of items representing 20% of value-usage.

Nevertheless, this analysis is proven to be

unsuitable when inventory is not homogenous –,

mainly when the major differences are not related to

the value of the SKUs. In this case it is important to

introduce other criteria, but these criteria must

represent factors which are significant to the

company (Flores and Whybark, 1986; Molenaers et

al., 2012; Ramanathan, 2006).

2.4.2 Ranking Criticality – VED Analysis

Criticality is the most important attribute when

classifying spare parts and components (Huiskonen,

2001). This analysis is very subjective. Considering

the industry and the organization context, the criteria

could potentially be very well defined. It is

important that all parties concerned (purchasing,

maintenance) should share ideas and reach

agreements. However, maintenance has more

influence in this case (because maintenance

managers know better than anyone else which SKUs

could compromise the normal running of the

company).

ProposalofanSKUClassificationFramework-AMulticriteriaApproach

415

Several authors (Cavalieri et al., 2008; Molenaers et

al., 2012) have conducted a criticality analysis, and

used a VED Classification which divides SKUs into

3 groups: Vital (Group V), Essential (Group E) and

Desirable (Group D). Although other techniques

could be applied, most studies consulted use this

technique. A VED Analysis allows SKUs to be

easily understood and ranked according to their

criticality, allowing the most critical items to be

quickly identified.

Defining criticality is not an easy task, although

this concept could be linked to the type of activity

for which the SKU is used (Bošnjaković, 2010).

This author assesses criticality through four

attributes: criticality for plant production, criticality

for safety, criticality for supply and criticality for

inventory. Duchessi et al. (1988) claims that

criticality is a function of the level of criticality of

the equipment where the SKUs is installed.

After close observation in the case study, it was

found that the company did already distinguish

between SKUs based on criticality. The company

uses two attributes to measure criticality; one

assessing the consequences for production and the

other related to the safety of operators. This is an

idea shared in the literature, where criticality is

measured as a function of the failure of a piece of

equipment (Duchessi et al, 1988; Huiskonen, 2001;

Molenaers et al., 2012).

Therefore, following the recommendations of

maintenance managers to build the framework and

proceeding as Flores and Whybark (1987)

reccommend, the main concern of management

should not be the cost of keeping of an item, but the

consequences of not keeping it.

Assessing criticality is very hard because it is

mainly based on subjective judgments and opinions

of managers (Botter and Fortuin, 2000). To achieve

a more systematic measure of criticality we decided

to use an Analytic Hierarchy Process (AHP).

The AHP sets pairwise comparisons for the

different criteria using a predefined scale (Saaty,

1980). This is a procedure used in Cavalieri et al.

(2008), Flores et al. (1992) and Moleanaers (2012)

with the purpose of establishing a ranking of

criticality.

2.4.3 Ranking Demand Variability

Bošnjaković (2010) claims that the frequency of

demand is a very important criteria when selecting

the inventory model. But as frequency of demand

may differ widely among SKUs, management

should be adjusted to reflect the pattern seen in the

frequency of demand. SKUs with the same pattern

of demand should then be grouped together.

Nevertheless the frequency of demand does not

account for erratic demand, as annual average

consumption does not reveal peaks of consumption

over time (Heinecke et al., 2013; Syntetos and

Boylan, 2005).

Calculating of the coefficient of variation (CV –

a measure that establishes a ratio between standard

deviation and average demand) reveals the

variability that exists between SKUs, showing how

they differ in volume and distribution of

consumption. Although the CV does not have an

intrinsic meaning, D’Alessandro and Baveja (2000)

present an example which illustrates what we should

see when this measure is used for analyzing demand.

If an SKU has a CV of 0.25 it varies little, so its

demand is more predictable than an SKU with a CV

of 0.75.

The boundary between high and low variability

SKUs is determined by using the procedure of

D’Alessandro and Baveja (2000). Here, a Pareto

Principle is applied, using an 80/20 rule to assess the

cut-off between SKUs. This same principle is

applied in ABC Analysis.

Potentially, many SKUs may not vary at all over

time. This may occur if the SKUs are not being

consumed or if the pattern of consumption is so

regular that the variability is almost zero.

2.4.4 Associating Criteria

More criteria could be considered, but Flores et al.

(1992) remark that the main purpose of SKU

classification is to simplify the stock and inventory

management. These authors argue that although it is

possible to include more than three criteria, the

analysis would be very complex (Flores et al., 1992).

The framework should only include those criteria

that are really important to management, and each

group of SKUs should have a matching inventory

management policy.

After selecting the classification criteria, SKUs

will be brought together to create groups of

homogeneous items. The outcome should result in a

scheme of classification that associates all three

criteria. This results in 27 different possible

combinations. Visually the outcome is a

tridimensional scheme as displayed in Figure 2.

Nevertheless, one of the purposes of this work is

to present a multicriteria classification, so in this

step each SKU will be ranked using a multicriteria

ABC Analysis. The main purpose of this step is to

decide which SKUs deserve closer attention by

management, also allowing inventory managers to

easily identify and quickly implement the

framework.

ICORES2015-InternationalConferenceonOperationsResearchandEnterpriseSystems

416

Figure 2: Criteria association scheme.

Therefore, weights were given to the three

criteria presented – value, criticality and demand

variability. The extra importance placed on

criticality meant this criteria had the highest rank in

terms of classification. Criticality is more important

for spare parts or components, but value is more

significant for consumables. In view of this, a 40%

weight is assigned to each of these criteria. The

demand variability, which is an indicator of

consumption patterns, received the remaining 20%

weighting. This multicriteria approach and ranking

of SKUs was also conducted by Flores et al. (1992)

and Ramanathan (2006).

It’s important to note that the outcome of criteria

weights reflects the concerns of all parties involved

in the project. These weights were arrived at using

the AHP technique that was explained earlier.

As mentioned before, the criteria were chosen

taking into account the industry, organization and

sector specificity.

2.5 Framework Validation

Once the SKUs have been classified it is necessary

to perform the validation processes. This can be

considered to be one of the most important steps in

the framework. A meeting is held with researchers,

purchasing, warehouse, financial and maintenance

personnel where the results are presented. This

meeting assesses if the framework is an important

tool for the decision-making process and if

adjustments are necessary.

2.6 Defining Inventory Management

Policies

The purpose of the framework was to identify SKUs

that share the same inventory management policies

and group them together. These policies may

determine that there is either “no need to stock”, or

there is a need to maintain a “safety stock”, or that

the traditional models of periodic/continuous review

policies should be applied.

SKUs with a zero stock policy should only be

purchased when the demand arises. This policy

means that a minimum amount of financial resources

are invested. The decision of “no need to stock” can

only apply to those SKUs that, if unavailable, do not

affect the normal operation of the company (Braglia

et al, 2004), or to SKUs with low criticality where

the inventory holding cost is higher than the “stock

out” cost (Bošnjaković, 2010). Nevertheless, these

policies should match with very reliable suppliers.

Following a “safety stock” policy means that

orders are made when a reorder point is reached.

This should be applied to SKUs with medium value

and medium demand variability; or SKUs with low

value, low demand variability and low criticality; or

for SKUs with high/medium criticality and with low

value and low demand variability; or lastly for SKUs

with medium criticality and medium demand

variability.

All the other cases should be managed using a

policy that defines a fixed reorder point, with

continuous reviews; however, it should be

remembered that high demand variability is a

criterion of high uncertainty. In these cases,

management should try to manage quantity

discounts for items of high and medium value and

diversify suppliers, while maintaining a tight control

on continuous reviews, especially in high and

medium criticality SKUs.

3 CONCLUSIONS

This paper proposes a framework for classifying

SKUs, which is designed to be an effective tool to

support decision making process for inventory

management.

The multicriteria classification has proven to be a

good solution for a warehouse storing a high number

of SKUs. Furthermore, the framework benefited

from co-working with personnel of different

departments of the company.

The multicriteria ABC analysis has proven that a

classification of SKUs using only one criterion is not

adequate as it ignores other criteria of vital

importance to the organization. This is particularly

relevant when the criteria of criticality and demand

variability are considered together. In this case some

SKUs, which previously were placed in groups of

low importance, emerge as critical SKUs. Where

ProposalofanSKUClassificationFramework-AMulticriteriaApproach

417

SKUs are vital for companies operations there

should be a separate inventory management policy.

Future research should validate the framework and

include any necessary adjustments. It is also

important to establish the periodicity for revising the

framework. It could also be of interest to apply the

same framework to other types of organizations and

contexts.

REFERENCES

Bacchetti, A., Plebani, F., Saccani, N., & Syntetos, A.

(2013). Empirically-driven hierarchical classification

of stock keeping units. Int. J.Production Economics,

143, 263–274.

Baker, T., & Jayaraman, V. (2012). Managing information

and supplies inventory operations in a manufacturing

environment. Part 1: An action research study.

International Journal of Production Research, 50(6),

1666-1681.

Bošnjaković, M. (2010). Multicriteria inventory model for

spare parts. Tehnički vjesnik, 17 (4), 499-504.

Botter, R., & Fortuin, L. (2000). Stocking strategy for

service parts–a case study. International Journal of

Operations & Production Management, 20(6), 656-

674.

Braglia, M., Grassi, A., & Montanari, R. (2004). Multi-

attribute classification method for spare parts

inventory management. Journal of Quality in

Maintenance Engineering, 10(1), 55-65.

Cavalieri, S., Garetti, M., Macchi, M., & Pinto, R. (2008).

A decision-making framework for managing

maintenance spare parts. Production Planning &

Control, 19(4), 379-396.

Childerhouse, P., Aitken, J., & Towill, D. R. (2002).

Analysis and design of focused demand chains.

Journal of Operations Management, 20(6), 675-689.

D'Alessandro, A. J., & Baveja, A. (2000). Divide and

conquer: Rohm and Haas' response to a changing

specialty chemicals market. Interfaces, 30(6), 1-16.

Duchessi, P., Tayi, G. K., & Levy, J. B. (1988). A

conceptual approach for managing of spare parts.

International Journal of Physical Distribution &

Logistics Management, 18(5), 8-15.

Flores, B. E., & Whybark, D. C. (1986). Multiple criteria

ABC analysis. International Journal of Operations &

Production Management, 6(3), 38-46.

Flores, B. E., & Whybark, D. C. (1987). Implementing

multiple criteria ABC analysis. Journal of Operations

Management, 7(1), 79-85.

Flores, B. E., Olson, D. L., & Dorai, V. K. (1992).

Management of multicriteria inventory classification.

Mathematical and Computer Modelling, 16(12), 71-

82.

Heinecke, G., Syntetos, A. A., & Wang, W. (2013).

Forecasting-based SKU classification. International

Journal of Production Economics, 143(2), 455-462.

Huiskonen, J. (2001). Maintenance spare parts logistics:

special characteristics and strategic choices.

International Journal of Production Economics, 71

(1–3), 125–133.

Jouni, P., Huiskonen, J., & Pirttilä, T. (2011). Improving

global spare parts distribution chain performance

through part categorization: A case study.

International Journal of Production Economics,

133(1), 164-171.

Molenaers, A., Baets, H., Pintelon, L., & Waeyenbergh,

G. (2012). Criticality classification of spare parts: A

case study. Int. J. Production Economics, 140, 570–

578.

Ramanathan, R. (2006). ABC inventory classification with

multiple-criteria using weighted linear optimization.

Computers & Operations Research, 33 (3), 695-700.

Saaty, T. L. (1980). The analytic hierarchy process:

planning, priority setting, resources allocation. New

York: McGraw.

Sexton, M., & Lu, S. L. (2009). The challenges of creating

actionable knowledge: an action research perspective.

Construction Management and Economics, 27(7), 683-

694.

Soylu, B., & Akyol, B. (2014). Multi-criteria inventory

classification with reference items. Computers &

Industrial Engineering, 69, 12–20.

Susman, G. I., & Evered, R. D. (1978). An assessment of

the scientific merits of action research. Administrative

Science Quarterly, 23(4), 582-603.

Syntetos, A. A., & Boylan, J. E. (2005). The accuracy of

intermittent demand estimates. International Journal

of Forecasting, 21(2), 303-314.

van Kampen, T. J., Akkerman, R., & van Donk, D. P.

(2012). SKU classification: A literature review and

conceptual framework. International Journal of

Operations & Production Management, 32(7), 850-

876.

ICORES2015-InternationalConferenceonOperationsResearchandEnterpriseSystems

418