SUSS iSmartGuide 23/2/26, 10:54 PM

Study Unit 1

Data-Driven Operations Management

[Link] Page 1 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

Learning Outcomes
By the end of this unit, you should be able to:

1. Examine the role of operations in an organisation or a supply chain

2. Distinguish the ten decisions of operations management

3. Discuss eFciency, eGectiveness, and sustainability in operations management.

4. Appraise the value of data analytics in business decision-making

5. 5. Relate data analytics to business practice in diGerent sectors and organisations, as

well as to the operations management decisions.

6. Discuss the challenges, misconceptions, and four realms of data analytics.

[Link] Page 2 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

Overview
This study unit establishes a process view of an organisation or a supply chain and dePnes
the concepts of operations and operations management accordingly. The ten critical
operations management decisions are highlighted and linked to topics in the subsequent
study units. EFciency and eGectiveness, as two objectives of operations management, are
brieSy discussed. Sustainability, a contemporary topic that emphasises the interests of the
surrounding environment and an organisation’s stakeholders, is also introduced.

With the increasing availability of data through social media, cloud computing, and the
Internet of Things, managers are becoming more inclined to employ data analytics to make
informed and evidence-based decisions. This study unit introduces data analytics and
presents its applications in diGerent business settings. The challenges in data collection and
some common misconceptions about data analytics and data sciences are highlighted. In
particular, the four realms of data analytics, as well as some commonly used technologies in
data analytics, are introduced. Examples and applications of data analytics in the ten
operations management decisions are presented to help readers understand the power of
data analytics in data-driven operations management, as compared to conventional
operations management.

[Link] Page 3 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

Chapter 1: Introduction to Operations

Management

1.1 Process View of an Organisation

The best way to introduce the concept of operations management is to discuss the
diGerence between operations and other business Pelds such as accounting, Pnance,
marketing, strategy, organisational behaviour, human resource management. In a very
simple way, one can assume that diGerent Pelds in business and management look at the
same organisation from diGerent angles:

Accounting: accounting focuses on measuring and communicating the Pnancial and

non-Pnancial information within an organisation

Finance: corporate Pnance focuses on the Pnancial planning of an organisation, its

capital structure, and the allocation of Pnancial resources.

Marketing: marketing focuses on the interfacing with external customers and

communicating the value of an organisation to them.

Strategy: strategy focuses on formulating a plan to achieve an organisation’s vision in

the short or long term.

Organisational Behaviour: organisational behaviour focuses on studying the human

behaviour within an organisation, as well as to the interface between human behaviour
and the organisation.

Human Resource Management: human resource management focuses on sourcing and

eGective management of human capital within an organisation.

While operations management is related to the above Pelds, people working in operations
management take a diGerent perspective when looking at an organisation – they view every
organisation as a set of interconnected processes that transforms inputs into outputs.

Figure 1.1 Process view of an organisation

[Link] Page 4 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

As shown in Figure 1.1, every organisation takes in inputs (on the left side) and deliver
outputs (on the right side). The inputs normally include capital, materials, equipment,
facilities, labour, knowledge, and time. Through some processes, these inputs are
transformed into outputs that are desired by the markets. In Figure 1.1, the processes can be
those used to purchase raw materials from suppliers, manufacture physical products, sell
products to customers, and move products between diGerent parts of an organisation. The
outputs, in general, are products and services. In a nutshell, operations are the
transformation processes of inputs into outputs. This dePnition of operations is built on
the concept of process, which is a collection of connected activities or tasks that transform
inputs into outputs.

Example 1.1: SUSS as a service system

SUSS provides education services to its students. The key inputs of SUSS include
applicants, teaching manpower, administrative manpower, knowledge, teaching
facilities, IT system, etc. The key outputs of SUSS are educated individuals. The key
processes to transforming these inputs into outputs include student recruitment
process, student enrolment process, teaching faculty recruitment process, course
development process, teaching and learning process, and assessment process, etc.

Based on this concept of operations, the process view of an organisation is established – that
is, every organisation has one or several key processes that transform inputs into outputs.
Therefore, when we study an organisation’s operations, we are essentially examining the
organisation’s key transformation processes.

Depending on the objectives of the analysis, these transformation processes can be analysed
at diGerent levels of detail (see Figure 1.2). For example, if the objective of an analysis is to
help answer questions related to an organisation’s long-term planning problems, then we can
aggregate the organisation’s transformation processes into three broad categories: namely
the buying process, the making process and the selling process. The analysis is then carried
out on these broad categories. On the other hand, if the objective of the study is to help the
Prm make daily production scheduling decisions, then we may have to look at some sub-
progresses – such as mixing, grinding, and packing – under the making process.

Figure 1.2 DiGerent levels of transformation processes

[Link] Page 5 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

Example 1.2: Sub-processes of the course development process

Just like the Example 1.1, the course development can be seen as a key transformation
process in SUSS. However, we can break down this key process into more micro-level
processes. For example, before a new course is developed, the needs of this course
should be identiPed and justiPed. This is then followed by the process of identifying
the relevant subjects, the processes of course material procurement, the process of
assessment writing, and the review process of course materials. For a course
development process, the typical planning horizon is 1 year, while the process of
assessment writing can be a few weeks to several months.

1.2 What is Operations Management?

Given that operations are dePned as the transformation processes of inputs into outputs,
and the study of operations focuses on the analysis of key transformation processes in an
organisation, we can dePne operations management as the management of the
transformation processes that create value for customers (Jacobs & Chase, 2018). The
scope of operations management includes activities related to the design, operation and
improvement of the transformation processes.

Figure 1.3 Schematic view of operations management

However, an organisation does not exist on its own. The organisation must interact with its
internal and external stakeholders such as suppliers and customers. An organisation,
together with its suppliers, customers, suppliers’ suppliers and customers’ customers
constitute a supply chain. Similarly, the process view can be applied to a supply chain. The
main processes in various stages of a supply chain can be categorised as follows:

Planning process: these are the processes that are required for the eGective functioning
of an existing supply chain

Sourcing process: these processes select suppliers to deliver the goods and services
needed to create an organisation’s products or services

[Link] Page 6 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

Making process: this process produces the physical products or provides the services

Delivering process: this refers to the logistics service between diGerent parts of an
organisation or a supply chain.

Returning process: this process collects products that are worn-out, defective, or
excessive.

Through these processes, materials, components, and services (as inputs) are supplied to
plants or warehouses as inputs, while Pnished goods or services (outputs) are delivered to
customers. Since a supply chain consists of diGerent organisations, the processes belonging
to these organisations should also be considered as the processes within that supply chain.

To diGerentiate between the foci of processes at the organisational level and supply chain
level, we restrict the dePnitions of operations process and supply chain process in this course
as follows:

Operations processes refer to the manufacturing and service processes to transform

resources into products within an organisation. The deployment of these processes
requires analysing the capacity, labour, and material needs over time.

Supply chain processes refer to the processes that move information and materials to
and from an organisation. For example, the logistics processes that move products
physically, and the warehousing and storage processes that position products for fast
delivery to customers.

Watch

Skillsoft YouTube. (2016, December 15). Operations management [Video]. YouTube.

[Link]

This video introduces the concepts of operations management and the link between
operations management, goods and services. It also discusses some key topics in operations
management such as strategic operations management, product design, supply chain
management, and quality management.

1.3 Ten Decisions of Operations Management

According to Heizer et al. (2019), there are ten areas of strategic operations management
decision-making that are among the most considered by managers. With the availability of
massive data collected through various computer data processing systems (such as ERP
systems, process control systems, condition monitoring systems, and the Internet of Things)
and the development of data analytics technologies, managers are equipped with
tremendous power to gain a better understanding of operational issues and make more
informed decisions based on the patterns, trends, causality, and insights revealed from
operational data.

1.3.1 Design of Goods and Services

[Link] Page 7 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

Design of goods and services usually determines the speciPcations, cost and quality of
products or services that satisfy consumer needs. It also deals with the sustainability aspect
of the product or service and the required human resources.

The most critical factor that distinguishes a service system from a manufacturing system is
the involvement of customers in the provision of service. The physical presence of customers
in the service system is called customer contact. The degree of customer contact required
largely determines the design of the service system.

Thanks to the involvement of customers in the production of services and their idiosyncratic
needs, a service system normally faces a higher degree of demand-induced variability
compared to a manufacturing system. Such variability has great impact on system
performance and presents a unique challenge to the service operations management. A
central topic in service operations management is diGerent ways to eGectively address
customer variability and provide a service of quality. Study Unit 5 discusses the challenges in
service operations and the strategies to redesign a service operation to better address the
variability presented in the processes. Study Unit 5 also presents examples of how Prms can
use information and data to better manage their service operations.

1.3.2 Quality Management

Quality management involves decisions with respect to determining the design quality
(product or service speciPcations) to meet customer expectations; establishing policies,
procedures, and tools to monitor production or service processes; discovering and rectifying
quality problems; and sustaining the production and provision of quality output.

Study Unit 3 discusses the relevant concepts and performance metrics in total quality
management. Study Unit 3 presents data analytics tools in the management of quality as
well as examples of how Prms in Singapore and other countries use data to improve the
quality management of their services or products.

1.3.3 Process and Capacity

Process and capacity decisions deal with the way products or services are produced, and the
required technologies and capital investment. Capacity planning is an important decision in
an organisation. DiGerent time horizons are used for capacity planning: long-term,
intermediate-term, and short-term.

Study Unit 4 discusses the intermediate-term planning – i.e., aggregate planning problem in
an organisation. Data analytics techniques are introduced to help develop an aggregate plan
including capacity planning for an organisation.

1.3.4 Location

Location decision concerns the placement of an organisation’s operations relative to its

customers and suppliers. Many factors must be considered for a location – political,
economic, demographical, logistics, infrastructure, human resource, etc.

1.3.5 Layout

[Link] Page 8 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

The layout decision is undertaken after strategic-level decisions on issues such as location,
capacity, product, and service design, are conPrmed. Layout decision deals with the design
of the Sow of materials, human resource, and information to achieve eFciency in an
organisation’s operations.

1.3.6 Human Resources and Job Design

Human resources and job design deal with the recruitment, motivation, and retainment of
human resources with the required talent and skillsets. This decision is particularly important
in service industries, as human capital is the most critical and expensive resource in such a
setting. This decision is closely related to other Pelds in business studies, such as human
resource management and organisational behaviour.

1.3.7 Supply Chain Management

Supply chain management deals with an organisation’s relationship with its suppliers.
Decisions in supply chain management aim to develop a supplier and distribution network to
ensure sustainable, reliable, quality, and cost-eGective inSow of raw materials, service parts,
and delivery of Pnished products or service to customers. The decisions with respect to
supply chain management determine the selection of suppliers and distribution partners,
costs, quantity, KPIs, and conditions of liability.

1.3.8 Inventory Management

Inventory management refers to decisions on ordering and holding so that an optimal

amount of inventory is held to satisfy customer needs while taking supplier capability and
production schedule into account.

Inventory is a critical asset of an organisation. An organisation needs to decide when, where,

and how much to hold inventory to ensure its smooth operations, while at the same time
minimising the related costs. Inventory management usually includes decisions with respect
to order quantity, pricing, and warehousing. As a result, inventory management is closely
related to an organisation’s marketing and supply chain management decisions.

Study Unit 2 provides an extensive discussion on inventory management and presents some
analytical tools that can be applied to manage inventory with uncertainty.

1.3.9 Scheduling

The decision of scheduling deals with intermediate-term or short-term planning of physical

and human resources to meet customer demands in an eFcient and eGective manner. These
are the given constraints determined at a higher level of planning. The decision of scheduling
often determines who, what, where, when, and how much to dispatch resources on a weekly
or daily basis.

1.3.10 Maintenance

Maintenance is related to decisions about facility capacity, production requirements, as well

[Link] Page 9 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

as manpower to ensure a stable and dependable process.

1.4 EFciency Versus EGectiveness in Operations

Management

Given that operations are the transformation processes within an organisation, and
operations management is the design, operation and improvement of these transformation
processes, then a natural question is: “What do people working in operations management
want to achieve?”

Well, the simple answer is that through operations management, the transformation
processes can be made more eFcient and more eGective. Therefore, eFciency and
eGectiveness are the two classic objectives of operations management. In this aspect,
operations management is very closely tied to the Peld of industrial engineering, economics,
and optimisation.

EFciency. “EFciency means doing something at the lowest possible cost. The goal of an
eFcient process is to produce a good or provide a service by using the smallest amount
of inputs” (Jacob & Chase, 2018, p. 14). This is a quantitative measure of an
organisation’s operations.

EGectiveness. “EGectiveness means doing the right thing to create the most value for
the organisation. Value is the attractiveness of a product relative to its price. Abstractly,
value can be dePned as quality divided by price” (Jacob & Chase 2018, p. 14). This is the
qualitative assessment of an organisation’s processes or a supply chain process.

One critical way to measure eFciency is to use the productivity concept. Productivity is
dePned as “the ratio of outputs (good and services) divided by one or more inputs (such
as labour, capital, or management)” (Heizer et al., 2019, p. 51). Higher productivity can be
achieved through two ways: one is to reduce input while producing the same amount of
output; the other is to produce more output while keeping the same amount of input.

There are two measurements of productivity:

Single Factor Productivity = Output / One Input

Multifactor Productivity = Output / Multiple Inputs

[Link] Page 10 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

Example 1.3 Calculation of Productivity

A restaurant serves an average of 920 customers a week. Each customer brings in an

average of $95 in revenue. The restaurant hires 4 full-time employees. Each employee
works 8 hours a day, 5 days a week, and is paid $20 an hour. On average, the material
cost is about $3000 a week, and the rental cost is $2000 a week.

The output and inputs are calculated as follows:

Output = Revenue = 920 customers * $95 = $87,400/week;

Labour Cost = 4 * $20/hour * 8 hours/day * 5 days = $3,200/week;

Material Cost = $3,000/week;

Rental Cost = $2,000/week;

Single Factor Productivity:

Productivity (with respect to labour input) = Output / Labour Cost = 27.31;

Productivity (with respect to material input) = Output / Material Cost = 29.13;

Productivity (with respect to rental input) = Output / Rental Cost = 43.70;

Multifactor Productivity:

Productivity (with respect to all inputs)

= Output / (Labour Cost + Material Cost + Rental Cost)

= 87,400 / (3,200 + 3,000 + 2,000) = 10.66.

An organisation can focus on either eFciency or eGectiveness to build its competitive

advantages. However, there is a trade-oG between eFciency and eGectiveness – i.e., an
organisation may not be able to achieve both objectives at the same time.

Study Unit 6 provides more extensive discussions on eFciency and eGectiveness in

operations management, the eFciency measurement, and how processes can be designed in
diGerent ways to achieve eFciency or eGectiveness.

1.5 Sustainable Operations Management

Conventionally, when a Prm designs its corporate strategy, the strategy is centred around
the interests of shareholders – that is, how to best maximise shareholders’ return on
investment. In this case, a Prm’s ultimate goal is to create and sustain value for its current
shareholders by increasing revenue and reducing cost. Such a corporate strategy translates
into the two classic objectives of eFciency and eGectiveness in operations management.

However, as people are increasingly aware and concerned with the impact of an
organisation’s operations on its surrounding environment and community, it is argued that
an organisation’s strategy needs to consider not only the interests of shareholders, but also

[Link] Page 11 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

the interests of stakeholders.

Shareholders are individuals or organisations who own shares of a Prm or use Pnancial
instruments to provide capital to the Prm.

Stakeholders are individuals, organisations, or the environment that are inSuenced,

either directly or indirectly, by a Prm’s activities. These individuals and organisations
include those who are aGected by the surrounding environment that the Prm operates
in, the Prm’s employees and/or the customers that the Prm serves.

The corporate strategy that takes care of stakeholders’ interests then cascades down to
operations strategy. In this case, we have conventional operations strategy versus
sustainable operations strategy.

Conventional operations strategy predominantly focuses on the economic dimension.

However, many operational processes also give rise to environmental and social impact.
The environmental and social criteria ought to be accounted for.

A Prm’s sustainable operations strategy expands the view of conventional operations

strategy. It requires the Prm not only to take care of its shareholders’ economic
interests, but also address the environmental and social impacts on its key stakeholders.
See Figure 1.4 for a graphical comparison of the conventional and sustainable
operations strategies.

Figure 1.4 Conventional versus sustainable operations strategies

Study Unit 6 provides an extensive discussion on the concept of sustainable operations

management and the commonly used tools and techniques to achieve sustainability in
operations.

1.6 Emerging Topics in Operations Management

With the advent of new business models and technologies, conventional operations
management is facing new challenges and opportunities. Study Unit 6 presents some of
these challenges and opportunities: the operational challenges introduced by the sharing
economy, and how organisations leverage on new technologies such as blockchain, artiPcial
intelligence, and cloud computing to improve their operational eFciencies and customer
services.

[Link] Page 12 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

Lesson Recording

Introduction to Operations Management

[Link] Page 13 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

Chapter 2: Data-Driven Operations

2.1 Data Analytics and Decision Making

Today, it is less common to see managers make business decisions solely based on their
personal judgement, intuition, and experiences. Instead, the ability to collect, manipulate,
analyse, and interpret factual evidence and data becomes a key successful factor in their
decision-making process. On the other hand, the amount of available data and information
to support decision-making – e.g., data collected by companies or collected through mobile
network, Internet of Things, social media – is growing at an exponential rate and the
unstructured nature makes them increasingly diFcult to understand and analyse.

According to Evans (2020), data analytics is “the use of data, information technology,
statistical analysis, quantitative methods, and mathematical or computer-based models to
help managers gain improved insight about their business operations and make better, fact-
based decisions” (p. 4).

As an illustration, data analytics can be used to improve the decision-making process of the
following business decisions:

Pricing: data analytics can be used to help set prices for physical products, services, or
contracts.

Customer segmentation: data analytics can be used to identify the targeted groups of
customers in diGerent industries.

Merchandising: data analytics can be used to determine the brands to promote, as well
as the quantities and allocations.

Location: data analytics can be used to identify the best locations for manufacturing
sites, service facilities and oFce buildings.

Supply chain design: data analytics can be used to identify the best suppliers, transport
modes and routes.

StaFng: data analytics can be used to manage human resources within a Prm, e.g.,
staFng level, deployment, hiring and training (sometimes referred to as “people
analytics”).

Health care: data analytics can be used to schedule and allocate available resources
(e.g., operating rooms, nurses), optimise patient admission, transfers and their care path,
sourcing of medical supplies, and forecast health risk factors.

Appendix 1.1 presents some examples of how data analytics are used in some of the ten
operations management decisions listed in Chapter 1.3.

2.2 Data Analytics in the Public Sector Operations

Public sector should leverage information technology to enhance the services and capital

[Link] Page 14 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

consumption. According to SAS (n.d.), the City Council of Town of Cary in the state of North
Carolina, USA, used SAS® Analytics to examine sensor data, predict future demand, discover
hidden problems, and engage customers visually. In 2010, the City Council installed wireless
meters for 60,000 customers. The Aquastar wireless system was able to read meters on an
hourly basis, which generates 8,760 data points per customer per annum instead of 12
monthly readings. Using data analytics, the city was able to analyse half a billion points of
data on water use. The data made available through the wireless water meters, together with
the customer accessible portal driven by data analytics, allowed water loss problems to be
detected at a much earlier stage, which made it much easier to Px. Through the data
analytics process, the City Council was able to acquire a large-scale picture of water usage
that is critical to the planning of future expansions of water plants and to the promotion of
targeted conservation eGorts. The ability to visually view household or commercial customer
data on an hourly basis has resulted in the following pragmatic applications:

The City Council could notify customers of any potential water leaking problems within
days

An alert of a surge in water usage can be sent to customers on a timely basis (within
hours)

Water usage was displayed in the online portal, making it easier for customers to
monitor their water usage and increasing their awareness of the need for water
conservation.

Here is an example of how data analytics supported the proactive detection of water usage
problem. Through the alert triggered from the online portal, one company experienced a
spike in water usage even though its employees were away for the weekend. This seemed
strange and the extraordinary reading helped the company trace the surge to its business
dishwasher, which was not functioning properly and was running continually over the
weekends. Without the data from the wireless water meter and the customer-accessible
portal, this problem could have gone undetected, resulting in a waste of money and water.

Watch

SAS Software. (2015, December 14). Town of Cary relies on SAS analytics
[Video].YouTube. [Link]

This video looks at how the Town of Cary applies data analytics to the management of water
usage.

2.3 Data Analytics in the Private Sector Operations

Omata (2019) shows how banks in Singapore use data analytics. With big data, banks can
help improve their service in the following areas:

Client Process: Manual processes like customer due diligence check can be automated
by machine learning and big data, thus shortening the lead time for corporate and
institutional client onboarding. Standard Chartered Bank partnered with tech start-up
Instabase to simplify due diligence procedures needed by market requirements. Based

[Link] Page 15 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

on various watchlists, these procedures collected information and details with respect to
sanctions and unfavourable publicity from public and private sources. Citibank
developed algorithms that automatically submit reports to regulators.

Human Resource Management: Since as early as 2011, DBS Bank has been utilising big
data analytics to raise productivity in recruiting and to improve employee retention. A
Human Capital Analytics team was set up to investigate employee turnover and
assessing employee attrition risk. There has been a substantial decline in the bank's
attrition rate from 27 percent to 18 percent.

Cyber Attacks: Since 2017, 55% of cyber threats were pre-empted and 62% of the
problems were addressed right from when they arose in OCBC Bank, thanks to the
employment of predictive analytics and big data. In addition, OCBC used voice
assistants to automate activities like extracting information and triggering tasks
throughout emergencies. These measures helped released manpower for crucial
decision-making during emergencies.

Auditing: Big data can be used to detect any possible issues at a bank branch that
requires auditing. Data such as customer Sows, ATM load, and call centre volumes can
be used to identify consumer needs. This helps the bank to develop new services and
reduce customer complaints. One of the early initiatives by DBS was to develop the
sequence for replenishing ATMs at diGerent locations. Data analytics allows the bank to
discover the patterns of ATM cash depletion and forecast the timing when ATM cash
shortage (also known as a cash out) arises. The implementation of predictive analytics
allowed the bank to eliminate 98% of cash out incidents and 92% of consumer
complaints. The result shows that DBS’ customers saved almost 800,000 hours in queue
for ATM service.

Anti-Money Laundering: The conventional anti-money laundering mechanisms focus on

a rule-based strategy. The mechanism points out every transaction that breaks a set of
rules. Bank staG check transactions to identify suspicious activities manually. With big
data analytics, OCBC bank reported a 35% reduction in the number of cases that
required manual investigation in 2017. DBS uses big data to detect anomalies in trade
Pnance.

2.4 Data Analytics in Small and Medium Enterprises

(SMEs)

During a 2018 survey of over 1,000 local SME business owners and IT decision makers in
Singapore (Singapore News Center, 2018), Association of Small and Medium Enterprises
(ASME) and Microsoft Singapore discovered the following interesting trends:

57% of SMEs are aware of digital transformation, which is adopted by 56% of SMEs.
About 25% of these SMEs felt that their digital transformation eGorts have been
successful.

Local SMEs estimate that digital transformation will allow them to achieve 22% projected
cost savings on average and a 26% increase in projected revenue gain.

SMEs give priority to four pillars of digital transformation – customer engagement,

operations optimisation, production transformation, and employee empowerment.

[Link] Page 16 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

SMEs perceive that the major benePts of adopting digital transformation include
improvement in business operations eFciency, attracting new customers, workforce
eFciency improvement, and cost saving.

SMEs consider high cost, diFculty of integration, no urgency, and lack of digitally skilled
workforce to be the major barriers to digital transformation.

The top technological solutions SMEs plan to adopt include big data and advanced
analytics, AI and machine learning, e-commerce platform, and business process
applications.

Despite the barriers to adopting data analytics, the SMEs conceded that its potential
benePts are greater than challenges. Below are some examples of how Singapore-based
SMEs have adopted data analytics in their operations.

Following are some examples of Singapore-based SMEs’ experiences in data analytics.

Kearea

In the past, local e-commerce platform Kearea purchased stocks based on a feeling about
what would be popular among clients. This resulted in excess inventory and losses when
merchants were discounted for sale. To address this, Kearea began analysing sales data and
inventory data. They employed basket analysis to better understand client purchasing
behaviour, detecting linkages between diGerent products (such as vacuum cleaners and
Plter packs or monitors and headsets), allowing them to make more accurate stock
purchases. As a result, their business increased by 50% within six to nine months (Infocomm
Media Development Authority, n.d.).

Plover Trip

Plover Trip, founded in 2017, provides corporate travel management and technology
solutions in Singapore. Most of its clients are companies with frequent business trips.
Through the analysis of customer booking trends and past sales information, Plover Trip is
able to produce business intelligence and build strategic partnerships with its travel and
hotel companies to increase income. For example, with insights from data analytics, the
company is able to improve travel policies or create preferred agreements with airlines or
hotels that are frequently used to assist their consumers to make greater savings. Through
these improvements in service, Plover Trip managed to achieve higher customer satisfaction
and a 98 percent customer retention rate (Infocomm Media Development Authority, n.d.).

ShopBack

ShopBack is a website that rewards customers with cashback when they shop at over 1,000
partner brands on online platforms like Amazon and [Link]. ShopBack has
implemented self-serve analytics, allowing employees to access and analyse data directly
without relying on the data team. This approach speeds up decision-making and the rollout
of new products and features. Using the product analytics platform Amplitude, ShopBack's
teams can gather insights on consumer behaviour and test hypotheses in near real-time. This
capability enabled them to launch a product comparison tool and a voucher feature quickly.
The comparison tool, available in Singapore, Taiwan, and Indonesia, increased daily
searches by over 25%. The voucher feature addressed the demand for value-for-money
purchases, as a result, there was a 10x increase in gross merchandise value and a 50%
average monthly growth. Additionally, insights, such as popular search terms during sales
events help tailor marketing campaigns eGectively (Yeo, 2021).

[Link] Page 17 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

2.5 Data Analytics Landscape in Singapore

KPMG Singapore (2017) dePned Pve stages of maturity in relation to a Prm’s data collection
and use of analytics. The Pve stages are:

Awareness – this is the stage where an organisation starts to develop an awareness of

data analytics. The characteristics of the awareness stage include data collection and
analysis without structured approach, data being stored in silos, and lack of data quality
and integrity.

Experimental – this is the stage where an organisation starts to explore with data
analytics projects. Instead of a planned approach, the experimental stage is ad hoc,
with data collection and analytics undertaken only when a need arises.

Cohesive – this is the stage where the organisation has established proper data
governance procedures and data analytics programmes. During this stage, the
organisation has established a structured approach to data analytics but may not be
actively driving its decision-making by incorporating the insights derived from analytics.

Business-driven – this is the stage where insights derived from data analytics are
incorporated into an organisation’s decision-making processes. The characteristics of
the business-driven stage include the integrated inter-departmental data collection and
analytics, and the utilisation of the enriched data sets by users from a few business
functions like IT or marketing.

Embedded – this is the stage where access to data analytics is made available to all
departments and employees so that data analytics can be actively engaged in decision-
making processes at all levels.

According to KMPG’s research, six business sectors in Singapore (namely consumer retail,
Pnance, transport, logistics, healthcare, digital media, and general) are positioned in
diGerent stages of maturity in the road map as shown in Figure 1.5. While most sectors
remain in the Prst two stages of maturity, the digital media and the private ride booking
sectors have advanced to the most mature stage.

Figure 1.5 Business sectors in diGerent stages of data analytics

(Source: Adapted from KPMG, 2017, p. 9)

For example, in the least mature end (i.e., the awareness stage), only a limited amount of

[Link] Page 18 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

customer data is collected by brick-and-and-mortar retailers. Information is often obtained

at the transaction level and is not tied to individual customers. In this sector, targeted
marketing does exist, but only to a small extent constrained by the availability of useful data.

At the most mature end (i.e., the embedded stage), there are wide-ranging, continuous
consumer data collection methods utilised in the digital media sector. Transaction data,
search information, product or webpage visit details can be used to construct models to
reveal customer preference and predict their behaviours and trends. Drawing insights from
customer data, technologies developed by the advertisement technology companies can
help improve the eGectiveness of companies’ marketing eGorts. Following are some
examples how customer data can be used:

Customer segmentation (e.g., by their demographics, residential status) so that and

advertisement can be more targeted at a speciPc group (e.g., Prms who just intend to
market their services to consumers at a certain age range).

The most relevant search engines to consumers are identiPed by complex algorithm and
an advertisement is displayed on these platforms to increase the eFcacy of
advertisement campaign.

Consumers who show interests in certain products (e.g., by searching and viewing
product details) but do not complete the transactions can be identiPed by algorithms so
that the advertisement of the same or similar products can be directed (“re-target”) to
these consumers (e.g., by displaying advertisement on their most visited websites or
their social medial accounts).

2.6 Challenges in Data Collection

What is the most important ingredient in data analytics? Of course, it is data. Data analytics
is not just about the analytics skills or techniques; it is also about how to collect and use data
wisely. One might assume that in data analytics, the most important step is to analyse data.
In fact, during the whole analytics process, one might spend more than 50% of his or her
time to collect, verify, recollect or supplement, and clean the data before s/he can apply
analytics skills and techniques to analyse them. Why does one have to spend so much time in
data collection? Well, there are a couple of reasons why it is so strenuous.

An analyst needs to spend time understanding the business model and operation
processes of an organisation before s/he can specify the kinds/types of data to be
collected. The analyst cannot request for data purely based on his assumptions.

The exact data one desires may not be available, so there might be a need to use proxy
data. For example, one may request for the dates of patients’ admission to and
discharge from hospitals to calculate the time spent in the hospitals. However, due to the
concerns of privacy or conPdentiality, hospitals may not be able to provide such data. As
a workaround, other data may be used to approximate the desired information.

The analyst does not have direct access to the desired data, so s/he needs to resort to
others’ assistance to access the data. The data owner might not have incentives to share
the data with any third party. It could be a challenging job to overcome the resistance of
the data owner.

[Link] Page 19 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

The one who extracts the data from a database does not understand the speciPcations
of the data request. Even the dePnitions of data points can be subject to debate.

Data integrity can be a problem – some data are missing, inaccurate, or have errors. The
analyst needs to Pnd ways to either recollect the data or supplement the incomplete
data with other sources.

Sample size may be a problem.

The following is an example how data collection can be challenging:

Figure 1.6 Patient care path for elective surgery

Figure 1.6 shows a typical care path for elective surgery. First, a person may feel unwell and
has some symptoms to see a general physician (GP). If the GP cannot resolve the patient’s
problem, the GP would refer the patient to a specialist for further assessment. The specialist
may refer the patient to a diagnostic centre for diagnosis and, based on the diagnostic
result, decide on the next step of treatment. If the patient requires elective surgery, the
patient would be placed on a waiting list to be admitted to the hospital. The patient may
have to wait for a period before the surgery can be performed. At every stage of the care
path, there may be waiting involved – waiting to see the GP, waiting to see the specialist,
waiting to be diagnosed, waiting to be admitted to the ward, etc.

As a result, the dePnition of a patient’s waiting time is not clear. In particular, what should be
the legitimate starting and ending points of patient’s waiting time? DiGerent stakeholders
may have diGerent opinions.

From patients’ perspective, they may feel that they suGer from the moment they feel
unwell and therefore, the starting point is the moment when the symptoms arise.

From a health administrator’s perspective, patients’ waiting times should start from the
time when their conditions are conPrmed, either by the GP or the specialist, because
that is when patients’ need for health resource is warranted.

The starting point of a patient’s waiting time may also aGect healthcare providers’ KPIs.
Therefore, they may resist including the part of patient waiting within their jurisdictions
in the calculation of waiting time.

From this simple example, one can see that even the data speciPcations can be subject to
debate and controversy, let alone other tasks in the data collection. A capable data analyst
must manage diGerent parties involved in the data collection process eGectively.

As a premise for eGective data analytics, data collection needs to ensure reliability and
validity.

[Link] Page 20 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

Data reliability refers to the accuracy and consistency of data.

Data validity means that the data collected correctly measure what they are supposed
to measure.

Data can be valid but not reliable. For example, a temperature gauge continuously measures
a few degrees above the true temperature of a food item is considered not reliable, but the
data are valid because the gauge does measure food temperature.

Data can reliable but not valid. For example, the counts of email queries to a company’s
service centre may not be valid in assessing customer satisfaction, although the data are
accurately and consistently recorded.

2.7 Misconceptions of Data Analytics/Science

Misconceptions of data analytics or data science can arise. Dr Krishnaswamy, CTO and co-
founder of AIDA Technologies, claimed that data science is not just “buying a data analytics
platform, pushing your data through it, and pressing the button to generate a Linear
Regression Model” (Quek, 2017, para. 13) More skills, experiences and know-how are needed.
She argued that “the key is to understand the business problem, understand what data is
available, and map both to the right machine learning approach” (Quek, 2017, para. 14). As a
result, it is misleading to believe that coding and algorithms are all one needs for data
science. Instead, a key ingredient in the process of data analytics is the eGective
communication. The solutions developed by data and sophisticated algorithms would
become a useless piece of code if the solutions were not accurately and suFciently
communicated to customers.

Some assume that machine learning algorithm performs 80 percent of the work. However,
the truth is something diGerent – it really does just 20 percent. For example, Eugene Yan, a
data scientist in Lazada, shared that “50 percent of his time is spent on understanding data,
exploration, data cleaning, data preparation, and feature engineering. The remaining 30
percent of his time is spent on execution, like building a POC, validation/AB testing,
developing an API, deployment, and so on” (Quek, 2017, para. 16).

Data analytics is a science, but it is also an art. Making sense of the conclusions from data
analytics models and being critical of them requires experiences, intuition, industrial
knowledge, and common sense. One should not assume the neutrality and objectivity of the
outputs from a data analytics model unconditionally. As a matter of fact, a degree of
subjectivity is involved at every stage of data analytics, including but not limited to problem
identiPcation and formulation, data requirement, data collection and sampling, data
modelling, model validation, data analysis and interpretation, and translation of data
analysis results to actionable solutions. As a result, the intuition, industrial know-how,
experience, personal judgement, (sometimes) luck of the people involved in data analytics
projects also greatly aGect the outcome of the projects.

2.8 Four Realms of Analytics

Descriptive, diagnostic, predictive and prescriptive analytics are considered as the four
realms of data analytics. These realms target diGerent purposes (Vlamis Software Solutions,

[Link] Page 21 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

2015).

Descriptive Analytics describes past events.

Diagnostic Analytics assists one to develop an understanding of the causality of past

events.

Predictive Analytics develops a prediction of the most likely future events.

Prescriptive Analytics makes recommendations on actions that can inSuence outcomes.

In a nutshell, descriptive and diagnostic analytics are considered as reactive approaches,

whereas predictive and prescriptive analytics are considered as proactive approaches. More
and more organisations are adopting proactive analytics approaches in their decision-
making processes.

For each realm of data analytics, there are skills and techniques to serve its purpose. In the
following sub-chapters, some of these skills and techniques are introduced. One should note
that the techniques discussed below are not exclusive to a unique realm; i.e., the same
technique can be applied in diGerent types of analytics.

Some of the commonly used data analytics technologies include the following:

Data Visualisation

According to Tableau Software (n.d., para. 1), “data visualization is the graphical
representation of information and data. By using visual elements like charts, graphs, and
maps, data visualisation tools provide an accessible way to see and understand trends,
outliers, and patterns in data. In the world of Big Data, data visualisation tools and
technologies are essential to analyse massive amounts of information and make data-driven
decisions”

Data Mining

According to Oracle Corporation (n.d., para. 3), “data mining is the practice of automatically
searching large stores of data to discover patterns and trends that go beyond simple
analysis. Data mining uses sophisticated mathematical algorithms to segment the data and
evaluate the probability of future events. Data mining is also known as Knowledge Discovery
in Data (KDD)”

“The key properties of data mining are:

Automatic discovery of patterns

Prediction of likely outcomes

Creation of actionable information

Focus on large data sets and databases” (Oracle Corporation, n.d., para. 4)

Data mining is useful for addressing problems that may not be accomplished by simple
query and reporting techniques.

Regression Analysis

Regression analysis is a type of statistical modelling that is used to quantitatively estimate

the relationship between a dependent variable (which can be perceived as the “outcome”)

[Link] Page 22 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

and one or more explanatory variables (which can be perceived as the “predictor”). Linear
regression is one of the most widely used forms of regression analysis which calculates the
linear equation that best Pts a set of observed data according to a speciPc mathematical
criterion.

Forecasting Methods

Forecasting process makes predictions on future events based on relevant information (e.g.,
data of past and present events). Trend analysis is commonly used for forecasting. There are
two types of forecasting: qualitative and quantitative forecasting.

Qualitative forecasting is useful when quantitative data are unavailable, limited, or

irrelevant. Decision-makers make forecasts based on their experiences, insights and
sometimes intuition. There are structured approaches that can help formalise the forecasting
process to yield satisPed outcomes.

Quantitative forecasting is useful when quantitative data about the past and present events
are available, and it can be assumed that patterns exhibited in the past and present events
are likely to be repeated in the future.

Simulation

A simulation attempts to mimic the way an existing or proposed system functions, such as
the operating room allocation process in a hospital, the distribution network of an e-retailer,
or the student recruitment process of a university. Simulation may be used to conduct the
what-if analysis and evaluate the costs and benePts of diGerent alternatives. Management
can then decide on the best alternative based on the output from the simulation model.

Optimisation Modelling

Optimisation modelling is useful in addressing decision problems that aim to achieve a

maximum or minimum value of a predetermined objective with constraints. For example,
achieving the maximum amount of output from a service process, given the limited amounts
of resources like manpower, capacity, and space, as well as constraints such as delivery time
and service level. Optimisation modelling is widely used to address problems arising in
various industries and business functions such as manufacturing, transportation, planning,
supply chain management, inventory management, service scheduling, healthcare, and
hospitality. Given the complexity of real-life optimisation problems, algorithms are developed
to search for the sub-optimal or near-optimal solutions.

More details and examples of these data analytics methods are shown in other study units.

Lesson Recording

Data-Driven Operations

[Link] Page 23 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

Summary
After reviewing this study unit, you should understand operations and operations
management from a process view of an organisation or a supply chain. Operations
management concerns diGerent aspects of the transformation processes of an organisation
– from product design, production, inventory to distribution. The objective of managing an
organisation’s operations is to achieve eFciency and eGectiveness. But in today’s business
context, an organisation should also consider the interests of the surrounding environment
and its stakeholders. As a result, the concept of sustainability is introduced.

With the increasing availability of data, operational decision-making can be more eGective
by leveraging on the power of data analytics. This study unit introduced the concept of data
analytics, the four realms of data analytics, challenges in data collection, as well as some
common misconceptions in data analytics and data science. To motivate the idea of data-
driven operations management, examples, and applications of data analytics in operations
management were also presented.

[Link] Page 24 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

Quiz
1. Operations management is applicable _______.

a. mostly to the service sector

b. to services exclusively

c. mostly to the manufacturing sector

d. to all Prms, regardless of service or manufacturing

Submit

2. An operations manager is NOT likely to be involved in _________.

a. the design of goods and services to satisfy customers’ wants and needs

b. the quality of goods and services to satisfy customers’ wants and needs

c. the identiPcation of customers’ wants and needs

d. work scheduling to meet the due dates promised to customers

Submit

3. Which of the following is the output of a higher educational institute?

a. Degree programmes

b. Educational services

c. Financial surplus

[Link] Page 25 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

d. Educated individuals

Submit

4. Which of the following is the output of a city’s transport system?

a. Transport service

b. Passengers at the destination

c. Number of ridership

d. Financial factor

Submit

5. Which of the following is NOT a key transformation process in the higher educational
institute?

a. Tuition payment process

b. Student recruitment process

c. Assessment process

d. Course development process

Submit

[Link] Page 26 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

[Link] Page 27 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

Formative Assessment
1. Which of the following is the best answer? Simulation can be used in _____.

a. descriptive and diagnostic analytics

b. diagnostic and predictive analytics

c. predictive and prescriptive analytics

d. descriptive and prescriptive analytics

Submit

2. Which of the following is the best answer? Data mining can be used in _____.

a. descriptive and diagnostic analytics

b. descriptive, diagnostic and predictive analytics

c. diagnostic, predictive and prescriptive analytics

d. descriptive and prescriptive analytics

Submit

3. Which of the following techniques can be used to estimate the price elasticity of
demand?

a. Regression analysis

b. Forecasting methods

[Link] Page 28 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

c. Simulation analysis

d. Data visualisation

Submit

4. In addition to the optimisation model, which of the following techniques is most likely to
be used to maximise proPt?

a. Data visualisation

b. Data mining

c. Regression analysis

d. Simulation

Submit

5. Which of the following is NOT a challenge when a data analyst collects data from a
restaurant chain?

a. The analyst lacks knowledge in the operations of restaurant chain

b. Data are unstructured

c. No direct access to database

d. Communication with internal staGs in the restaurant chain

Submit

[Link] Page 29 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

[Link] Page 30 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

References
Alicke, K., Glatzel, C., & Karlsson, P. M. (2016). Big data and the supply chain: The big-supply-
chain analytics landscape (Part 1). Retrieved from [Link]
functions/operations/our-insights/big-data-and-the-supply-chain-the-big-supply-chain-
analytics-landscape-part-1#

Dunn, S. (n.d.). Big data, predictive analytics and maintenance. Retrieved from
[Link]
[Link]

Evans, J. (2020). Business analytics: Methods, models, and decisions (3rd ed.). Pearson
Education Limited.

GitHub Incorporation. (n.d.). Kaggle Caterpillar tube pricing competition. Retrieved from
[Link]

Haurie, J. (2020). Data analytics is transforming supply chain management. Retrieved from
[Link]

Heizer, J., Render, B., & Chuck, M. (2019). Operations management: Sustainability and supply
chain management (12th ed.). Pearson Education Limited.

Infocomm Media Development Authority. (n.d.). IMDA's BDDB levels up data analytics for
smarter decision-making. [web log post]. Retrieved from
[Link]
data-analytics

Jacobs, F., & Chase, R. (n.d.). Operations and supply chain management (15th ed.). McGraw
Hill.

Jiang, Y., Abouee-Mehrizi, H., & Diao, Y. (2020). Data-driven analytics to support scheduling
of multi-priority multi-class patients with wait time targets. European Journal of Operational
Research, 28(3), 597-611.

KPMG Services Pte Ltd. Understanding the data and analytics landscape in Singapore: A
study of data and analytics adoption and practices in six sectors. Retrieved from
[Link]
publications/publications/occasional-paper/understanding-the-data-and-analytics-
[Link]

Lunden, I. (2014). Google conPrms purchase of satellite startup Skybox Imaging for $500M.
Retrieved from [Link]
satellite-startup-skybox-imaging-today/

Omata, S. (2019). 5 ways banks in Singapore are using big data. Fintech News Singapore.
Retrieved from [Link]
using-big-data/

Oracle Help Center. (n.d.). Data Mining Concepts. Retrieved from

[Link]

Quek, C. (2017). 4 Singapore-based data scientists share how data has been impacting lives.
Retrieved from [Link]

[Link] Page 31 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

SAS. (n.d.). Municipality puts wireless water meter-reading data to work. Retrieved from
[Link]

Singapore News Center. (n.d.). Singapore SMEs who embrace digital transformation expect to
see average revenue gains of 26%: ASME-Microsoft study. Retrieved from
[Link]
transformation-expect-to-see-average-revenue-gains-of-26-asme-microsoft-study/

Tableau Software LLC. (n.d.). Data visualization beginner's guide: a dePnition, examples, and
learning resources. Retrieved from [Link]
visualization

UPS StaG Writer. (2016). ORION: The algorithm proving that left isn't right. Retrieved from
[Link]

Vlamis, T. (2015). The four realms of analytics [Web log post]. Retrieved from
[Link]

Xu, X., Shen, Y., Chen, W., Gong, Y., & Wang, H. (2020). Data-driven decision and analytics of
collection and delivery point location problems for online retailers. Omega, 100, e102280. http
s://[Link]/10.1016/[Link].2020.102280

Yeo, S. (2021). How product analytics helped ShopBack adapt to changing user demands
amid the pandemic. Retrieved from [Link]
shopback-adapt-changing-user-demands-pandemic

[Link] Page 32 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

Appendix 1.1 Data Analytics in Operations

Management Decisions
This appendix presents examples of how data analytics can be implemented in some of the
ten operations management decisions.

Location Decision

Xu et al. (2010) presents an example of how data mining models are used to analyse internet
customers and optimise online retailers’ collection and distribution points (CDPs). The
demand data of online shoppers inSuence the location of CDPs, which imposes a major
eGect on online retailers. Unlike other more traditional models that propose distribution
points, the authors present new approaches that integrate with analysis of consumer
behaviour data. In these new approaches, data mining is incorporated into facility location
models to optimise online retailers’ CDPs. The authors explore a real consumer behaviour
data set that comprises of 257,685 users’ online shopping records. Customers’ purchase
likelihood is estimated by using Pve data mining models. These estimate results allow the
authors to develop two facility location models that aim to minimise the costs of the
attended and unattended CDP locations, respectively. With the aid of numerical experiments,
the authors quantify the service level and cost at each location. The insights derived from
this analysis provide useful information with respect to the trade-oGs between customer
service level and total logistic cost.

Supply Chain Management Decision

According to Haurie (2020, para. 9), “data analytics can improve supply chain eFciency
through the following ways: validating data, detecting anomalies, benchmarking operations,
allowing for mobile reporting and visibility into global logistics’ oGering real-time route
optimisation, improving demand forecasts and inventory management, and providing for
responses to government audits”

In Alicke et al. (2016), supply-chain analytics turn data collected at diGerent stages of a
supply chain into real insights. According to McKinsey, “supply chain analytics uses data and
quantitative methods to improve decision making for all activities across the supply chain. In
particular, it does two new things. First, it expands the dataset for analysis beyond the
traditional internal data held on Enterprise Resource Planning (ERP) and supply chain
management (SCM) systems. Second, it applies powerful statistical methods to both new and
existing data sources. This creates new insights that help improve supply chain decision-
making, all the way from the improvement of front-line operations, to strategic choices, such
as the selection of the right supply chain operating models” (para. 4)

For instance, many businesses usually gather statistics on procurement quantities and
vendors over a small range of operations in the procurement phase. However, these supply-
side data can be used for more than analysing annual spending and reviewing a supplier’s
performance. Adverse sales variability can be identiPed by monitoring supply processes in
real time at the transactional level. There are opportunities in predictive insurance. The key is
to analyse data from production mapping, Google trend-style information, and social media
with respect to strikes, Pres, or bankruptcies. These data can be used to monitor supply
disruptions so that a company may take strategic actions ahead of its competitors.

Data analytics may be utilised to derive proactive supply chain decisions. For instance, a

[Link] Page 33 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

pharmaceutical Prm has produced a database of all packaging bids received. The data were
analysed in order to thoroughly appreciate the cost structure of these manufacturers so that
comprehensive costing models of various forms of packing can be established. With up-to-
date details on product rates, factor costs and plant usage, the Prm can use these models to
identify the best vendors for its new packaging programmes. In the same manner,
Caterpillar organised a competition on Kaggle's crowd-data research platform to model
quoted rates for industrial tube assemblies (GitHub Incorporation, n.d.).

Such “clean sheet costing” calculations also apply to the procurement of shipping and
warehousing. Cost breakdown data of trucking and warehousing operations across the globe
can be exploited. With these evidence-based results, companies have a better understanding
of the “should cost” and are in a better position to negotiate with carriers and logistics
solution providers (LSPs).

Data analytics can be applied to transportation. For example, in the trucking operations,
analysis of fuel consumption is used to identify ways to improve driving eFciency,while GPS
technologies are used to support the real-time allocation of warehouse bays so that truck
waiting times can be reduced.

Many couriers begin real-time routing of packages dependent on position and traFc
statistics. For example, UPS spent almost 10 years designing its On-Road Optimized
Optimisation and Navigation system (Orion) to maximise the network. The latest technology
would save the organisation many million a year (UPS, 2016).

The usage of big analytics often allows shipping Prms to distribute packages less frequently
without spending more capital. Strategically speaking, businesses can save resources and
lower pollution by choosing the correct means of transit. Logistics companies will benePt
immensely from analytics so they can realise what products need to be transported
eFciently by rail, and when there is an option to transport by barge or rail (McKinsey &
Company, 2016).

Inventory Management Decision

Brick and mortar stores have realised how data-driven optimisation will provide them with
strategic advantages. It is often used nowadays by many retailers such as for markdown
selling and shelf-space management. In addition, advanced analytics will allow companies
to realise how long to retain goods in high value regions and encourage them to discover
sales advantages with clustering analysis of goods.

A satellite imagery technology (provided by Skybox) is used to track vehicles in the car park
to predict in-store demand (Lunden, 2014). In a similar fashion, drones with cameras can be
used to monitor on-shelf inventory levels.

Retailers are dealing with numerous challenges. One of these challenges is how to detect and
prevent out-of-stock situations. Both manual checks and RFID can be costly, and therefore
are often not practical in monitoring individual items. Unlike in the past, retailers now track
sales data for out-of-stock situations. If an object that is frequently sold is not on the store,
an alarm is activated so that someone can verify if the item is simply not accessible. Smart
systems are being utilised, including sensor technology used to search shelves for
malfunction products and cameras that track on-shelf stock levels.

Related systems may be utilised immediately at the place where or the time when a product
or service is used. Amazon’s Dash is an Internet shopping service that allows customers
purchase same items without repeating the ordering process (McKinsey & Company, 2016).

[Link] Page 34 of 35
SUSS iSmartGuide 23/2/26, 10:54 PM

Stores can gather user data via the Internet of Things (IoT) – e.g., customers’ internet-
connected refrigerators – and use these data to forecast demand.

Scheduling Decision

Jiang et al. (2020) presents the use of data analytics in the planning of health services. In
this research, the authors Prst collect the healthcare data from over 3.7 million records of
medical scans and use the data to determine the best performing factor, and help predict
daily arrival of patients at a particular hospital which aims to serve 90% of its patients within
their appointment wait times. This article recommends two simple scheduling policies:
namely, priority policy and Prst-in Prst-out (FIFO) policy, despite the higher risk of delay and
waiting in the system.

Under the priority policy, diGerent priority levels (coded as diGerent weights in the system)
are assigned to patients when they arrive at the clinic initially and then accumulate as a
function of the duration of the waiting period. Under priority promotion rule, patients are
promoted to a higher priority level after waiting for a predetermined amount of time. The
performance measure is the amount of time customers waited before they were moved to a
waitlist, and the proportion that received no treatment (i.e., the proportion of patients with
waiting time exceeding the target). The result shows that enhancing the waiting time
management would greatly shorten the wait time for all patients. Implementing clinical
scheduling can result in signiPcant improvement in the waiting times without the need for
capacity expansion.

Maintenance Decision

Predictive maintenance is powered by predictive analytics. The Prst goal of this technology is
to track and control device irregularities and faults, which avoids the risk of crucial loss and
downtime. This technology facilitates the deployment of restrained capital, the improvement
of facilities and life cycles, the development of quality and supply chain systems, and the
general satisfaction of stakeholders.

Dunn (2020) presents the usage of statistical analysis in maintenance operations. From a
maintenance point of view, data from various systems within an enterprise are maintained
and examined in silos. However, the emergence of the Internet of Things and other
advancements in information technologies allow administrators to develop the capacity to
save and analyse a more complete view of asset condition, with more data collected from
diGerent sources.

Watch

Siemens. (2019, October 9). Asset Analytics and Predictive Maintenance with PlantSight
[Video]. YouTube. [Link]

This video presents an example of how Siemens uses predictive analytics to help its clients
improve the eFciency of their maintenance operations.

noormohamed004@[Link]
[Link] Page 35 of 35