C16 - Technology and data analytics

Visual Overview

Objective: To understand what cloud, mobile technology and Big Data are, the
benefits and risks associated with using them and how they can be used to develop
strategy.

1
 1.0 Introduction

This chapter covers the role of new and disruptive technologies in transforming the
nature of business analysis and transactions.
The technologies themselves, cloud, mobile and smart technologies are explained
first with an explanation of associated risks and benefits.
We will also cover information and data analysis, and, in particular, big data. Access
to high quality information in decision making at all levels of an organisation, is
crucially important, so this area of the syllabus has wide ranging implications.
The impact of disruptive technologies is an important consideration in strategic
decision making, and so could be part of any discussion about strategy.
Recent trends in IT include the use of cloud computing, mobile and smart
technologies. The most significant development has probably been the use of
artificial intelligence, which has the potential to significantly change the business
landscape.

1.1 Cloud computing

Cloud computing involves the sharing of a pool of physical and/or virtual resources
via the Internet rather than using locally based systems and networks:
 Cloud hosting services offer remote storage and data management solutions.
For example, Microsoft OneDrive, Dropbox and Google Drive allow users to store
their files online and access them from any device connected to the Internet.
 Software as a service (SaaS) provides access to software applications hosted
on a central server (rather than stored on the hard drive of the user's computer).
Users typically pay a monthly charge to use the software (i.e. on a subscription
basis) rather than an upfront license fee for "owning" the software on the users'
own devices. For example, Microsoft Office 365 provides access to a Web
version of Microsoft Office programs for a monthly fee.
 Infrastructure as a service (IaaS) provides a complete architecture service for a
customer's systems which are accessed over the Internet. For example, Amazon
Web Services provide the technology (websites, storage) to other businesses.

Example 1 Cloud Computing with AWS

Amazon Web Services (AWS) is the world’s most comprehensive and broadly adopted cloud platform,
offering over 175 fully featured services from data centers globally. Millions of customers—including the
fastest-growing startups, largest enterprises, and leading government agencies—are using AWS to lower
costs, become more agile, and innovate faster.
Source: aws.amazon.com/what-is-aws

2
 Platform as a service (PaaS) offers a platform for programmers to develop applications
very quickly, without the complexity of building the infrastructure typically associated with
developing and launching new apps. New apps may be offered over the Internet or
software sold out of the box.

1.1.1 The Cloud as an Alternative to a Traditional Network

Infrastructure
The basic infrastructure to support information systems includes the following:
 machines including Internet servers and clients (devices that connect to the
server such as PCs and mobile devices);
 software;
 network and communications; and
 physical facilities such as server rooms.

The nature of this expenditure is capital, although there will be some running costs.
Using cloud services will save some, if not all, of these costs:
 Data and software is stored on hardware owned by the cloud provider.
 The cloud provider also owns the software licences.
 Savings on physical facilities (e.g. no need for air-conditioned server rooms).

Instead, the cloud provider will charge a fee for the use of its hardware and software.

Factors to Consider
The following factors need to be considered when deciding whether to move to the
cloud as an alternative to owned hardware and software:
 Cost: The cloud is likely to be cheaper because service providers benefit from
large economies of scale. Many charge on a usage basis so cost savings can be
even greater for small and medium-sized businesses that do not use their full
capacity. (Costs of owning hardware, software and IT staff are fixed to a large
extent, regardless of usage.)
 Cash flow: Owning hardware and software requires upfront investment (although
some may be obtained under a lease agreement). Cloud services are revenue
expenditure rather than capital expenditure.
 Changing needs: If the business is growing very fast or business processes are
changing regularly, the flexibility of cloud computing is a particular advantage.
Businesses may become constrained and opportunities limited if owned
hardware and software is insufficient to support growing volumes of data or
changes in business processes.
 Security: Although many organisations have concerns about relying on the cloud
to store confidential data, this may not be justified, as cloud providers should
have much better security procedures in place.
 Selection of supplier: It is important to select an established provider of cloud
services. Their financial stability is critical to ensuring that they can continue in
business. The quality of services provided by a reputable provider should exceed
that which would be available if the systems are maintained in-house.

3
  Speed: For some industries, the speed of data-exchange or transaction
processing is critical (e.g. trading on the London Stock Exchange). Being closer
to the server geographically can mean transactions are processed nanoseconds
quicker. Using the cloud may not be appropriate if this would slow down the
processing speed.

Benefits
 Flexibility and scalability – Infrastructure as a service (IaaS) provides more
flexibility to organisations in setting up and configuring hardware. If the hardware
configuration is not right it can be changed, because the business is paying for
the right to use the hardware rather than buying it. Extra resource can be
accessed as and when required.
 Ease of setup – Hardware configurations can be set up more easily and quickly,
allowing the business to take advantage of new opportunities when they arise.
 Cost effectiveness – Rather than investing up front in hardware and software,
organisations typically pay a monthly fee for the right to use programs and data
access stored by the third party. This spreads the cost and allows more flexibility.
Cost is based on use.
 Reliability – Using cloud-based services effectively means outsourcing them. In
particular, as there are a number of available servers, if there are problems with
one, the resource will be shifted so that clients are unaffected.
 Managers are also freed from having to manage IT issues and can focus on the
main activities of the business.

Risks
 Data security – Data is held and processed on third-party devices. The use of
mobile devices also increases the risk of data being obtained by unauthorised
personnel (e.g. if a user leaves a mobile device on a train).
 Reliance on Internet service providers (ISPs) – Over time, ISPs may increase
their charges or provide unreliable services, which could disrupt operations.
 System outages – When devices on the cloud "crash". For example, storms in
the US that disrupted Amazon's cloud services affected Netflix, as AWS hosts
Netflix.
 Intellectual property right issues – Who owns the data: the client or the
company that provides the storage?

Hybrid Arrangements

Definition

Mission critical – an activity, device, service or system whose failure or disruption will cause a failure in
business operations.

Using cloud services does not necessarily mean that all systems and data will be
switched to the cloud. Many organisations switch on a piecemeal basis:

4
  "Mission critical" systems (e.g. communication systems of online retailers) are
kept in house, initially;
 Less critical systems are migrated to the cloud first.

Once the cloud had been tried and tested, the critical systems can be migrated later.

1.2 Mobile Technology

Definition

Mobile Technology – is any electronic communications device intended for portable use, and the
communication networks that connect these devices. Typical mobile technology examples are
smartphones, tablets and smart watches.

Mobile Cloud Computing (MCC) is a technique or model in which mobile apps are
built, powered and hosted using cloud computing technology. Cloud computing is
particularly suited to mobile devices such as smartphones and tablets because it
avoids the need to store apps and data on the devices.

Example 2 Salesforce

"Ever since we launched our first CRM solution, Salesforce products have run entirely in the cloud. That
means it's all online – no software, no hardware. There are no expensive setup costs, no maintenance,
your employees can work from any device with an internet connection – smartphone, tablet or laptop…."
Source: http://www.salesforce.com/eu/crm/what-is-salesforce/

Mobile networks include

 Radio networks using cell towers that enable mobile devices to switch
frequencies automatically and communicate without interruption across large
geographic areas.
 4G networking, which is the current service standard for most wireless
communications. This involves data being organised into packets for
transmission and being reassembled at the destination.
 Wi-Fi uses radio waves that connect mobile devices to the internet.
 Bluetooth is a method of connecting mobile devices over short distances using
short-wavelength radio waves.

Benefits of Mobile Technology in business

 more flexibility of service for customers, including the ability to make purchases at
any time
 increased ability to communicate at all times leading to greater efficiency and
productivity of staff
 access to apps and services

5
 improved networking capabilities

Risks of Mobile Technology

 Distraction – as the range of mobile technologies and services increases, so
does the potential for them to become distractions that adversely affect
productivity
 Wellbeing – the prevalence of mobile technologies in the work and leisure sphere
means that it becomes more difficult for employees to disengage from work, and
this may lead to unnecessary stress and burnout.
 Health – health risks associated with mobile phone use include damage to
eyesight, spread of germs and traffic accidents.
 IT security risks – mobile devices are vulnerable to security risks, especially if
they contain sensitive or critical business data.

1.3 Smart Technology

Smart technology is said to incorporate three main technologies – smart devices,

connected devices and Internet of Things (IoT) devices – but the terms tend to be
used interchangeably.
One view is that:
 Connected devices are products with a real-world function that is connected to
the Internet in order to transmit data or to be controlled remotely.
 Smart devices aim to make users’ lives easier. They have some kind of
intelligent behaviour that enables them to react to real-world situations and even
predict the needs of their users. Most are connected to the internet (although they
don’t have to be) but not all connected devices are smart.
For instance, imagine a light switch connected to the internet which could be
turned on or off form anywhere in the world – a connected product. A smart light
switch would turn lights on or off as you enter or leave the room.
 An IoT device is part of the massive network of connected devices and systems
that collect and exchange data.

The Internet of Things (IoT) refers to the billions of physical devices connected to the
internet, collecting and sharing data. Processors and wireless networks provide a
level of digital intelligence to devices that allows them to communicate real-time data
without human intervention.
The term IoT is mainly used for devices that would not normally be expected to
function in this way such as smart speakers, home security systems, smart fridges
and driverless cars. (PCs and smartphones are not considered IoT devices.)

Benefits of Smart Technology:

 Improving work processes: The oil and gas industry has been able to build
models based on seismic images to improve the precision of exploration for new
reserves. Data about the conditions that products are kept in can be used to help
predict with precision the life of individual products within a larger batch. This is

6
 particularly important in the food industry, where shelf life can vary according to
harvesting, holding, processing and distribution arrangements.
 Generating efficiencies: Information provided by smart devices used in
manufacturing operations can help prevent bottlenecks. Efficiencies also include
reduction in the use of resources (e.g. better energy management in buildings). In
agriculture, sensors on farms enable farmers to monitor soil conditions and
irrigation levels to optimise use of water and fertilizers.
 Monitoring for potential problems: Devices can give warnings to remote
locations of security problems (e.g. open windows) and enhance the
effectiveness of procedures designed to prevent problems occurring. “Predictive
maintenance” uses data relating to the condition of equipment to estimate when
maintenance should be performed. The timing of maintenance will be partly
determined by when it is cost effective, but also trying to ensure that it happens
before there is likely to be a loss in performance.
 Underpinning strategic development: In the automotive industry 3D printers
can be used to shape pressing tools by printing the shape directly onto
granulated steel. This helps enable car designs to be tailored to customer
requirements.

The IoT has already proved useful in a number of industry sectors:

 Health: Devices in the home can provide alerts that medical emergencies, such
as falls, have occurred. They can also enable continuous remote monitoring of
blood pressure and heart rate, and also responses to treatment.
 Transport: Traffic cameras, sensors and vehicles’ tracking modules can be used
for longer-term objectives, as well as warning motorists of traffic problems ahead.
They can identify the most-frequently taken routes and provide data that can be
used as a basis for decisions about simplifying those routes.

Risks of Smart Technologies:

 As always, security is an important concern. A significant criticism of smart
devices is that development of applications has not been accompanied by
sufficient development of security. In some instances the consequences of
hacking could be life-threatening.
 Privacy and breaching data protection regulations are also serious concerns. The
more information that devices provide, the more is known about what individuals
are doing, leading perhaps to a greater ability to predict behaviour. Individuals
may fear “big brother” and whether the inferences drawn from data collected
about them can be used to “control” their behaviour.
 Compliance with data protection regulations is further complicated by regulations
continuing to develop and differences in regulations between countries.
 Platform fragmentation (i.e. providing applications that can work with smart
devices that use a variety of hardware and software) means that obsolescence of
one device can disrupt the whole network of which that device forms a part.
 Data storage and processing costs may be high. Problems with connection,
gathering and understanding data may hinder IoT application, emphasising the
need for expert staff to be involved.
 Many smart device applications do not get beyond the pilot or development
stage which can concentrate too much on the technology and not establish, early
on, whether there is a convincing business case.

7
 o The time horizons for successful implementation may be felt to be too long-
term.
o Barriers to implementation may also include the limitations of the hardware
being used by businesses or the insufficiency of their information systems’
infrastructure.
o Old legacy systems may be a particular issue, as businesses may not want
to replace all computer equipment for new devices that can accommodate
the IoT and other smart technologies.

1.4 Technology and Strategic Opportunities

New technologies such as cloud, mobile and smart technologies are a disruptive
force that can change entire industries. It is important therefore for strategic
managers to be aware of the trends and consider how they may provide new
opportunities.
Information technology can assist strategies in the following ways:
 Cost leadership – cloud computing has the potential to reduce the costs of IT
and therefore the underlying business processes and thus support a strategy of
cost leadership. It also enables businesses to move functions to other
geographical areas where salaries may be lower (e.g. shared service centres).
 Differentiation – Mobile Cloud Computing (MCC) can provide businesses with
opportunities for communicating with their customers in more effective ways (e.g.
by sending informational text messages to customers' smartphones).
 Develop new ways of doing business (i.e. develop new products or services,
identify new markets or niches or evolve the current business into a
fundamentally different structure through innovation). MCC presents new ways of
working (e.g. teleworking from home by accessing systems via the cloud).
 Expand production and services by identifying and entering new markets
(location and buyers) or producing new products (growth). For example, the cloud
provides opportunities for existing internal business functions and processes to
become "extensible" enterprises (i.e. internal business functions become
business opportunities in their own right). Amazon Web Services is a good
illustration.
 Forge strategic alliances – through integration of systems, shared systems,
joint ventures and mergers (alliance). Strategic alliances are easier to implement
where organisations have already moved their systems to the cloud, as it is much
easier to integrate the systems.

2.1 Data to Support Strategy

Business leaders require a range of information and data to help them develop and
implement strategy. Some data will be provided by the organisation's information
systems; some will be sourced externally.

8
2.1.1 Information for Developing Strategy
Much of the information required for developing strategies will come from external
sources and concern:
 the macro external environment;
 competitors and how they behave;
 regulations.

Research for informing strategies provided by the organisation's information systems

and data analysis will include:
 Historic financial information used to plan future activities (e.g. activity based
costing can be used to decide which products or services to pursue and which
customers to service).
 Financial and non-financial performance measures used to assess the
performance of different divisions and make decisions about them (e.g. whether
to invest in underperforming divisions or shut them down).
 Sales analyses used to identify market segments on which to focus. Big Data
(see s.3) enables the analysis of a wide range of data for this purpose.

2.1.2 Implementing Organisational Strategy

Business leaders require information to monitor the successful implementation of
strategies. The type of data required will be heavily influenced by the business
strategy:
 Differentiation requires non-financial performance measures relevant to the
basis on which the business has chosen to differentiate (e.g. product quality or
speed of delivery).
 Cost leadership will focus on actual costs compared with budgeted costs.
Detailed product costs will be required for this.
 Niche will need relevant performance indicators. For example, a manufacturer of
environmentally friendly cleaning products needs to know how satisfied
customers are with the effectiveness of the products.

Many organisations focus on critical success factors and use key performance
indicators (KPIs) to measure how well they perform against these factors
(see Chapter 21).

2.2 Business Information Systems (BIS)

A business information system (BIS) integrates computers, software,

communications technologies, data, business processes and people to supply
information that is accurate, up-to-date and useful for analysis and decision-making
at all levels in an organisation. A key feature of a BIS is that it brings together
information from many sources, both internal and external.
A BIS includes the typical functions of any information system: data input,
processing, storage and output. BIS can be considered an overarching term which in

9
practice may consist of several different systems that support the information
hierarchy of an organisation, including:
 Executive information system (EIS) – provides high-level, longer-term
information for the strategic apex (strategic planning level) of the organisation.
o Information includes key organisational metrics, often presented as a
dashboard and intelligence derived from external sources with an emphasis
on the future (e.g. forecasting and scenario planning).
o There may well be a bias towards non-financial, qualitative information
relating to the organisation's markets and competition.

 Management information system (MIS) – provides medium-term information for

the middle line (or tactical level) of the organisation for the purpose of
management and control.
o Information typically supports managers in relation to budgeting and
resource allocation as well as performance measurement and
management.

 Operational information system (OIS) – provides information to the operating

core (operations level).
o Information relates to the shorter term and might be very detailed.
o Data is often available to users almost as soon as it is generated (in real
time).

The actual systems that support these information levels have become somewhat
blurred, as has the terminology used. Organisations increasingly incorporate Internet
technologies, e-business and Enterprise Resource Planning (ERP) systems into a
single, highly integrated system. Such systems are shared across functional
departments and business units and often extend outwards to customers, suppliers
and business partners.

2.3 Data Analysis

The development of computerised information technology enables organisations to

collect huge amounts of data and to analyse it using detailed database queries. The
analysed data can assist in developing and implementing organisation strategies:
 Analysis of sales can identify changes in customer trends to inform the
development of new products.
 Analysis of costs (e.g. using activity-based costing) can identify opportunities for
businesses to reduce costs.
 Analysis of customers (e.g. using loyalty cards) can improve the effectiveness of
advertising by using more targeted adverts rather than traditional mass marketing
campaigns.

10
 Example 3 Data Mining

Many retailers operate customer loyalty schemes under which customers present a loyalty card when
paying for purchases. Details of their purchases are recorded and the data is "mined" to help identify
customer behaviour shopping patterns and trends.
Applications of the data mined include:
 Targeted advertising (e.g. promotional offers that a target group of customers may enjoy based on
their historic purchases);
 Developing effective customer retention programs – it is more economical to retain customers than
replace them with new ones;
 Identifying cross-selling opportunities; and
 Demand forecasting – more accurate predictions should reduce inventory losses (e.g. of perishable
items)

3.1 Big data

The increasing use of the Internet and, in particular, social network sites has led to a
huge increase in data that is available to organisations. If analysed appropriately, the
increased business intelligence can be used to improve the performance of an
organisation. Intelligence about consumer trends, for example, can be used to
identify new marketing opportunities, leading to improved sales.
Traditionally, the information available to organisations was mainly internal, relating
to transactions and operations of the business.
Today there are many additional external sources of data that may also be useful to
organisations. For example:
 discussions on social media about an organisation's products and brands; and
 the behaviour of visitors to the website.

Definition

Big data – extremely large data sets that may be analysed computationally to reveal patterns, trends and
associations, especially relating to human behaviour and interactions.

Characteristics of big data include the following, which are sometimes referred to as
the “5Vs” of Big Data:
 Volume; big data is enormous, with data combining to form huge data pools or
data lakes.
 Velocity; big data grows quickly, a massive and continuous flow of data.
 Variety; big data includes structured, semi-structured and unstructured data.
 Veracity refers to the inconsistencies and uncertainty which tend to be inherent in
big data.
 Value; to be of value, big data must be capable of being analysed to reveal
meaningful insight.

11
 Big data is variable and complex, making it difficult to analyse using traditional
database and software applications.

Example 4 New CDO Inspired by Complexity

Sainsbury’s has 191,000 staff, $34 billion revenue, 40,000 products and 1,300 stores – all creating an
incredibly complex business.
However, their Chief Data Officer, Andrew Day, is inspired by this.
"In the complexity of the business, what I started to realise from a chief data officer's perspective and from
an analytics perspective, there was just loads of opportunity, literally opportunity everywhere," Day said.
"Whether you're talking about the farm, or the process through to the fork, there was opportunity to apply
data and analytics to create value for our customers and for our business."
The opportunities of which Day speaks involve using data to manage every facet of the business. Retail
data can be used for growing better produce, producing better dairy products through livestock feeding
practices, improving how stock is brought to stores, measuring the effectiveness of merchandising
displays, and many more applications besides. The most ambitious of the opportunities is to capture the
300 transactions per second running through Sainsbury's checkouts, and turn them into a vast pool of
data which can then be leveraged to drive better buying- and customer experience-focused decisions.
"It's the unintended or the unseen applications which I think are going to be those that deliver the most
value to us."
Day is the first CDO in the company's 146-year history. His appointment reflects a growing recognition of
the power of the ever-growing volumes of data in the enterprise and the need for a strategy to harness it.

3.2 Big Data Strategies

Salvatore Parise, Bala Iyer and Dan Vesset describe various business big data
strategies. These strategies are based on two dimensions:
1. Data type: the nature of the data used, which can be either transactional data (e.g.
customer history) or non-transactional data (e.g. discussions on social media); and
2. Business objective: the use of the data – either measurement or experimentation.

12
3.2.1 Performance Management
This involves the use of transactional data that exists on the business's internal
information systems to measure performance using predetermined queries (e.g.
sales by customer or sales by market segment). Powerful analytical tools can be
used to obtain more complex queries.
Performance management may use business intelligence tools. These enable users
to generate their own reports and queries and to drill down to obtain more detailed
analysis of the data.
An example of performance management would be a sales analysis by sales person
and region. This would help the organisation identify which regions to focus on and
which sales people need additional training.

Activity 1 Customer analysis

Baked Co is a company that has a chain of bakeries in a single geographical

location. Ingredients are purchased centrally and distributed to the locations
according to orders placed by the shop managers. Products are baked in each shop.
The company uses scanners at the checkouts, but currently does not have a
reporting tool that can run reports from the sales system.

Required:

Prepare briefing notes for the next board meeting which recommend how big data
strategies can be used to improve their analysis of sales data.

13
 *Please use the notes feature in the toolbar to help formulate your answer.

The problem – Baked Co wants to improve our analysis of sales. There is no

reporting tool available, so reports are manually prepared. This is time consuming,
so there are delays in producing reports. The reports rely on a high degree of
manual intervention, so they may not be accurate. It is also difficult to get ad hoc
reports to investigate anomalies, pursue opportunities or increase understanding of
customer behaviour, due to the time required to compile a report.

The solution – Baked Co could invest in a data warehouse that automatically

extracts transaction data from the sales system. This would allow the use of a
reporting tool that managers could use to quickly create standard and ad hoc
reports, based on up-to-date data.

This will allow Baked Co to analyse of sales and identify trends, in terms of products
and locations. This will help to improve marketing decision making, such as where
advertising revenue should be focussed, or what promotions should be offered and
in which locations. We could also analyse the range of products purchased by each
customer to create attractive product bundles. Detailed analysis of sales can also
identify which sales and marketing activities are successful and which are not,
enabling corrections to be made to the plans rapidly.

The live analysis of sales trends will also assist in planning production, to ensure we
have enough stock of in demand products, moving products between locations if
needed. It will also reduce the waste of raw ingredients and products with lower
demand, which have to be discarded if they are not sold.

Professional skill

This question provides the opportunity to demonstrate the following professional skill:
Commercial acumen skills – by showing awareness of effective methods of data analysis.

3.2.2 Data Exploration

Data exploration also uses transactional data, but attempts to use it to predict future
behaviour by identifying attributes that managers may not previously have
considered.
Cluster analysis is a technique that may be used to identify:
 customer (market) segments that marketing departments may not even be aware
of;
 behaviours common to groups of people with similar attributes.

14
 Example 5 Cluster Analysis

An analysis of purchases of certain combinations of products, such as unscented lotions and vitamin
supplements, by women was used to determine a "pregnancy prediction score". Promotions on products
for pregnant women and clothes for new borns were then targeted at women based on the scores, leading
to an increase in sales of the advertised products.

 A danger of such a marketing approach is that people may feel that they are
being "spied on" and be put off buying products that are targeted at events in
their lives.

3.2.3 Social Analytics

Social analytics measures non-transactional data, particularly the exposure that
organisations achieve on social media sites (e.g. Facebook and X (formerly Twitter)).
Organisations may use social media to advertise new services and products. Social
analytics typically measures three aspects of a business's presence on social media:
1. Awareness – of social media content (e.g. the number of times a video posted
on Facebook is viewed);
2. Engagement – activity initiated by the users (e.g. the number of visits to the
page);
3. Reach or word of mouth – the extent to which content is sent by users to other
users (e.g. the number of times an item is shared on Facebook).

Recent developments in social analytics allow organisations to identify high "Klout"

individuals (i.e. people who have a large following on a social media) and whose
messages are shared with others. These individuals are targeted as "influencers" to
help market a brand.

Example 6 Clout

Virgin America offered 120 "influencers" free flights on a new route to Toronto in the hope that they would
comment on their flight experience on social media sites (they were not obliged to). The campaign
resulted in high brand awareness of the new route.

 The disadvantage of social analytics is that measuring awareness and

engagement does not measure the effect that these have on revenue and profits.

3.2.4 Decision Science

Decision science involves the analysis of non-transactional data for decision-making.
For example, releasing ideas about new products and discovering how successful
they might be by analysing discussions about them on social media.

15
 Decision science can be supported by several tools:
 Crowdsourcing – obtaining ideas, services or content by collecting contributions
from a large community or specific group of people (usually online).
 Textual analysis – a data-gathering process that examines and interprets the
characteristics of narrative or other visual messages.
 Sentiment analysis – a type of textual analysis that aims to analyse the mood of
a community.

Example 7 Blurb

A San Francisco based book-making platform and creative community, Blurb had repeat purchase rates
of 35–40%, but heavy competition from large competitors for advertising keywords on Google and social
sites made it expensive to attract new business via search. Blurb used textual analysis to identify the most
valuable customers and those most likely to convert to a purchase, and segmented customer lists by
lifetime value. They also used SmartBidding and dynamic advertising to improve the outcomes from paid
searches.

3.3 Big Data Risks

There are numerous risks associated with the use of big data:
 The underlying data may be inaccurate or become out of date very quickly. Big
data uses many sources, some of which may not be reliable especially if the
origins are not consistently monitored and tracked. (However, inaccuracy may not
be a significant problem as it may not affect the general trends.)
 The data may not be representative of the population as a whole. Although whole
data sets can be analysed rather than just a sample, they may exclude important
people (e.g. those who do not use social media).
 The analysis may identify spurious correlations between variables which are
coincidental rather than causal links. Caution (and scepticism) must be exercised
in using correlation for predictions.
 The amount of data held about individuals raises concerns about privacy.
Although many countries are legislating on this, much personal data is gathered
without permission or notification.
 Data security is an even greater concern when large amounts of personal
information are being stored. Organisations risk civil actions if there are breaches
of such data.

Example 8 Prediction Errors

In 2009, during a flu epidemic in the US, Google was able to track the spread of flu across the US a week
faster than the government's Centers for Disease Control and Prevention, by analysing searches relating
to "flu remedies" in its search engine.

16
 Example 8 Prediction Errors

In 2013, the same analysis overestimated the incidence of flu by 100% due to people without flu searching
for flu remedies. Thus, the original correlation between people searching for flu remedies and people
actually having flu had changed. One explanation is that the auto-complete function in Google search may
have led to people accidentally searching for flu remedies without intending to.

3.4 Data Analytics

3.4.1 Techniques
There are unlimited ways in which data can be analysed, depending on the
questions which can be asked by business leaders. These questions may relate to
strategic decisions about new product developments, marketing and pricing.
Essentially the data analytics process should always begin with posing a specific
business question such as:
 Which products or customers are the most profitable and why?
 What new features and functions do customers want and why?
 How can sales be predicted if the weather or other factors change?
 What is the cheapest way of distributing goods from the warehouses to the
stores?
 At what price should a new product be sold to maximise profit?
 Is money or inventory being lost due to fraud and error? If so where, how much
and why?

Essentially there are three types of data analytics:

1. Descriptive
2. Predictive
3. Prescriptive

3.4.2 Descriptive Analytics

Descriptive analytics is the analysis of data to observe what has been and is
currently happening (e.g. the analysis of sales data by product, by outlet, by
customer, etc).
Observing past trends and analysing and classifying data in different ways helps
business leaders to draw conclusions about the data which might be relevant to
informing strategy or to support or make more effective decisions (e.g. minimising
risks or exploiting opportunities).
Descriptive analytics is often enhanced by the use of graphs or charts which are
commonly available in spreadsheet packages.
Pivot tables are particularly useful for presenting information in different ways. A
pivot table is a spreadsheet tool used to classify, filter, summarise, analyse and

17
 present information in a variety of ways. Better visualisation of comparisons, patterns
and trends provides insights into the data.

3.4.3 Predictive Analytics

Predictive analytics use techniques to evaluate strategies and to anticipate strategic
outcomes depending on which scenarios play out. For example, in Excel:
 Scenario Manager is a what-if analysis tool which creates a summary report for
each of a group of scenarios (e.g. “best”, “worst”, “most-likely” case, etc).
Scenarios can be modified and even merged.
 The Analysis ToolPack calculates and displays the results of complex statistical
analysis including multiple linear and non-linear regression. This allows the
business analyst to test the relationship between independent or input variables with
a dependent or output variable.

Example 9 Multiple Regression

Weather conditions such as hours of sunshine, amount of rain and temperature levels can affect sales of
certain items, such as barbecues or cold drinks. Knowing how well these factors are associated with the
dependent variable in the past, such as the sales of certain products, allows business leaders to predict
future sales more accurately when the independent variables are established.

Particular care must be taken in interpreting results where independent variables

appear to be good predictors of the dependent variable but are not the causes.

Example 10 Cause and effect

x, the independent variable = the price of petrol

y, the dependent variable = sales of motorcycle helmets
x is a good predictor of y, as there is a very high positive correlation between the two, but x is not the
direct cause of y. This is because as the price of petrol rises, the sales of motorcycles increase because
they use less fuel. In most countries, it is a legal requirement for motorcyclists to wear a helmet for safety
reasons, so sales of motorcycle helmets, as a complementary product, also increase.

3.4.4 Prescriptive Analytics

Prescriptive analytics is potentially the most powerful form of data analytics for
decision makers and business leaders. It is the field of analytics which is concerned
with optimisation such as:
 maximising sales;
 minimising costs; or
 maximising profit.

It draws on techniques which are commonly available in spreadsheets such as:

18
  Goal Seek: this feature finds the input value needed to achieve the desired
results from a formula (e.g. the maximum interest rate that can be afforded).
 Solver: whereas Goal Seeks works only with one variable input value, Solver can
be used, for example, to determine an optimal product mix (i.e. problems with
multiple variables and constraints).

Activity 2 Data Analytics

Gossip Co is an online news service that primarily covers celebrity news. It

developed out of a blog written by the founder. It has evolved into a news website
with thousands of regular subscribers.

Required:

Prepare a presentation slide, together with accompanying notes, for the board which
discusses the benefits and the costs of investing in big data analytics.
*Please use the notes feature in the toolbar to help formulate your answer.

Exam Advice

Candidates will not be expected to carry out analyses that would normally be done by expert data
analysts on the behalf of someone in the role of a business leader.
However, candidates will be expected to be aware of how big data and data analytics can be used in
leading or managing a business more effectively. For example:
 If a business may be losing or gaining competitive advantage because of its data analytics capabilities,
candidates may be required to identify these issues and consider ways to improve or exploit these
capabilities.
 A business may face threats from disruptive technologies and need to safeguard against them. Such
threats might be from competitors (e.g. Air BnB), security threats or the risk of using data irresponsibly
(e.g. Facebook).
 Other scenarios could include:
o inappropriate use of data analytics;
o leaders failing to exercise sufficient scepticism about information obtained;
o unethical use of such information.
 An exam might include information in the pre-seen material or the exhibits related to data analytics
(e.g. a spreadsheet, charts or multiple regression outputs) and require candidates to exercise
scepticism in interpreting or applying the findings.

19
 Big Data investment
 Available data
 Use in existing operations
 Use for identifying strategic opportunities
 Costs

Available data
Big Data analytics can provide information about the amount of online traffic to the
Gossip Co website and peaks and troughs in access. It can also provide information
about which articles are attracting attention and comment, and the uptake on other
media channels.
 Use in existing operations
Additional data about which articles gain the most reader and comment should
help us to be able to identify which celebrities and what types of articles, images
and headlines attract the most readers, and target and adapt our content to these
popular items and approaches.
 Use for identifying strategic opportunities
If Gossip Co has greater insight into the items and celebrities that attract site
visitors, we can use this to both improve content and attract new subscribers.
We may also be able to provide more information to potential advertisers which
will enable them to provide more value and higher advertising revenues.
 Costs v benefits
Costs will not only include set up and security costs, but also the costs of
expertise to interpret the data. Any investment in Big Data capabilities will also
need to be compatible with any current technology initiatives.

Summary
The Board needs to determine whether a large investment in Big Data will be
justified given the anticipated increase in subscription and advertising revenue.

Professional skill

This question provides the opportunity to demonstrate the following professional skill:
Communication skills – in presenting the points in the slides concisely and for supporting notes
which will make the issues clear.

4.1 Decisions

Businesses need data to support decisions about new products. For example:
 Should the product be developed? Can revenue from selling the product be
expected to exceed the costs of making it?
 Which market segment should the product be aimed at?
 What pricing strategy should be used?
 How to apply the 7Ps of marketing (see Chapter 13).

Much of this data will be collected by the marketing department.

20
 4.2 Market Research
Definitions

Market research – the process of systematically gathering, recording and analysing data and information about
customers, competitors and the market.
– American Marketing Association
Marketing research – a function that links the consumer, customer and public to the marketer through information.
– Philip Kotler

 In practice, the terms may be used interchangeably, without any distinction.

Market research is vital for every organisation because the focus is on questions
about:
 environmental scanning – competitors, market structure, government
regulations, economic trends, technological advances and numerous other
factors that make up the business environment;
 customers – their needs and wants and their buying behaviour;
 products – what products can be produced with available technology, and what
new product innovations near-future technology can develop (NPD – new product
development);
 marketing – the efficiency and effectiveness of advertising; and
 finance – the financial situation of companies, industries or sectors. In this case,
financial analysts usually carry out the research and provide the results to
investment advisers and potential investors.

4.2.1 Data Sources

Market/marketing research data is available from a number of sources which may be
classified as "primary" (field research data) or "secondary" (desk research data).

Primary Research Secondary/Desk Research

 Company itself
 Interviews  Government information
 Questionnaires  Press
 Observation  Trade associations
 Simulation or experimental method, e.g.  Specialist publications
advertisement, testing  Internet

21
 4.3 Analysing Customer Behaviour

In marketing, a market is defined in terms of its buyers (customers) or prospective

buyers. The following distinctions are typically made:
 Consumer markets
 Industrial markets
 Government markets
 Reseller markets
 Export markets.

4.3.1 Consumer Goods

Consumer goods do not need further commercial processing and are usually bought
by individuals for their own or their family's use. These goods are further classified
according to the purchasing method.

Toothpaste, make-up, washing powder, canned

Convenience goods (everyday purchases) foods

Refrigerator, vacuum cleaners, lawnmowers,

Shopping goods (bought less frequently in specialist outlets) motor cars

Specialty goods (associated with a particular manufacturer or

brand) Rolls-Royce

Services (intangible goods) Hairdresser

4.3.2 Industrial Goods

Industrial goods are used to make other goods/services.

Raw material (extracted from the natural state) Iron ore, timber, crude oil

Processed materials and components (transformed into a condition ready to

incorporate into finished goods) Textiles, packing materials

Machine tools, computers,

Capital goods (items used in manufacture and selling but are not consumable) buildings

Supplies (industrial consumables) Stationery, fuel

Services Insurance, debt-factoring

22
The processes involved in analysing customer behaviour can be explained using the
following diagram:

 Inputs – information about products, services, brands and alternatives available.

Include both rational (factual) and emotional elements.
 Behavioural determinants – include personality, culture, social status and the
importance of the purchase decision.
 Perceptual reaction – information from inputs is not accepted at face value but
interpreted. Factors to consider:
o Information that has been actively sought may be valued more highly than
that passively received.
o The credibility of the source of information, its perceived authority and
content.
o Information that is perceived to be unimportant or lacks credibility will be
filtered.
 Processing determinants – these personal criteria include purchase motivation,
available satisfaction, past experience and judgemental criteria.
 Inhibitors – include price of product and availability, also financial and time
constraints.
 Outputs – include attention, understanding, attitudes, purchase intentions and
the purchase decision.
 Ultimate outcome – may be revised feelings about the product, a purchase
decision, or to delay buying.

Industrial buyers will follow a similar process as that of consumers but will be more
rational and objective about the decision because they face more complex buying
decisions. Purchases often involve large sums of money, complex technical and
economic considerations, and interactions among many people. Often it is a
decision-making unit (DMU) at many levels of the buyer's organisation that will
decide what and when to buy in terms of:

 quality of product/service;
 price – may be the main purchasing motivation;
 reliability – of goods and delivery dates;
 credit terms – importance depends on financial status of the buyer;

23
 quantity – volume for current and future needs;
 flexibility; and
 delivery.

5.1 Terminology

 Artificial intelligence (AI) – using computer systems to copy the cognitive

functions of the human brain in learning and solving problems. Applications
include video games, fraud protection, cyber-security, virtual assistants
(“suggestions for you”), online customer support (using AI “chat bots”) and in
autonomous (“self-driving”) vehicles.
 Robotics – robots that have enhanced dexterity, senses and intelligence. Robots
have been used on industrial assembly lines for more than half a century
(General Motor’s Unimate was introduced in 1961). Robotics applications include
collaborative robots (“co-bots”), which are flexible, easily reprogrammable and
can learn by watching a demonstration of a task. Other applications include
surgical robots, self-driving vehicles and drones used in military and public safety.
 Machine learning – a subset of AI that is generally understood as the ability of a
system to make predictions or draw conclusions based on the analysis of a large
historical data set using mathematical algorithms. It is the ability of machines to
exhibit human-like capabilities in areas related to thinking, understanding,
reasoning, learning or perception. Applications include image recognition (used in
automatic friend tagging in Facebook, for instance), traffic prediction (as used in
Google Maps, say, to provide the shortest route to a destination) and automatic
language translation. Through advanced machine learning, robots are now able
to work autonomously on skilled tasks such as simple news article writing, trading
stocks and shares and undertaking case research for legal firms.

5.2 Benefits

By exploiting existing and emerging AI and robotics, organisations can benefit by:
 Automating routine, repetitive and labour-intensive tasks and processes
 Reducing operating costs and increasing efficiency
 Providing 24/7 service via fixed and mobile devices
 Developing innovative new products and services
 Ensuring products and services meet customer needs
 Scaling up operations with fewer and cheaper resources
 Extracting more value from existing investments in technology

A survey of executives reported in a Harvard Business Review article ‘Artificial

Intelligence in the Real World’ reported the following business benefits of AI:

24
 Percentage of executives
Benefit citing the benefit

Enhance the features, functions and performance of products 51%

Optimise internal business operations 36%

Free up workers to be more creative by automating tasks 36%

Make better decisions 35%

Create new products 32%

Optimise external processes like marketing and sales 30%

Pursue new markets 25%

Capture and apply scarce knowledge where needed 25%

Reduce headcount through automation 22%

5.2.1 Strategic decision making

Whilst a recent survey by ACCA showed that only between one tenth and one third
of respondents thought that a machine algorithm could lead or be given full reliance
in situations requiring complex judgement and interpretation (“Machine Learning –
More Science than Fiction”), benefits of applying AI to strategic decision making that
have been put forward include:
 Rapid assessment of different outcomes based on alternative strategies
 Increased speed of decision making, especially in response to new data
becoming available or competitive threats emerging, allowing organisations to
capture the benefits of stronger market positions earlier
 Identification of missing data
 Increased rationality, particularly via the removal or reduction of cognitive bias by
decision makers
 Creation of a common basis for decision making
 Incorporation of learning from experience

5.2.2 Pursuing corporate objectives

AI, robotics and machine learning can help organisations:

25
  Improve productivity – from more accurate demand forecasting, predictive
maintenance, hyper-personalised manufacturing, optimising manufacturing
processes and automated materials procurement
 Increase competitive advantage – for example AI has allowed Ocado, the
online supermarket, to deliver a personalised service to customers, tailored to
their individual needs, with the ability to adapt their user interface depending on
whether they are new or experienced users
 Deliver innovation – by accelerating the development of new products and, for
instance, adding machine learning capabilities to innovation and idea
management ensure corporate memories for ideas are much longer, leading to
sustainable and ongoing innovation
 Increase profitability – through intelligent automation (creation of a new virtual
workforce), labour and capital productivity (complementing and enhancing the
skills and ability of existing workforces and physical capital, by predicting and
preventing machine failures say) and driving innovation.

Example 10 Using Robotics

Union Bank Becomes the First Nigerian Bank to Introduce Robotics into its Operations (2018)

“In an avant-garde move which sets it apart from its counterparts in the Nigerian banking industry, Union
Bank has announced the deployment of the innovative Robotic Process Automation (RPA) technology in
its operations; a first in the Nigerian Banking industry.”
“RPA technology makes use of robots which are software tools developed to simplify business process
delivery. The software robots offer improved business efficiency and data security by automating repetitive
tasks across multiple business applications without altering existing infrastructure and systems.”
“The Bank’s adoption of RPA technology is expected to enhance staff productivity, reduce process
turnaround time and improve accuracy and compliance. With the new technology in place, employees are
better able to focus on other value adding and customer related functions, significantly improving the
overall quality of customer experience.”
Source: www.unionbankng.com

5.3 Risks

Risks can arise from:

 Data difficulties: As the amount of unstructured data from sources such as the
web, social media, IoT and so on increases, its use becomes more problematic.
This could result in pitfalls such as inadvertently utilising or revealing sensitive
data hidden among anonymised data, leading to organisations falling foul of
privacy rules such as the European Union’s General Data Protection Regulation
(GDPR) or reputation risk.
 Technology troubles: Technology and process issues across an organisation’s
operations can negatively impact the performance of AI systems.
 Security issues: If security precautions are insufficient, it is possible to stitch
together seemingly non-sensitive marketing, health and financial data that

26
 organisations collect to feed AI systems to create false identities, leading to
consumer backlash and regulatory repercussions.
 Unstable/inappropriate models: AI models can deliver biased results (if, for
example, a sample of a population is underrepresented in the data used to train
the model), can become unstable or produce outcomes for which those affected
by its decisions have no recourse (such as someone denied a loan with no
knowledge of what they can do to reverse the decision).
 Unintended interaction problems: There is a risk when people and machines
interact, for example in automated transport, manufacturing and infrastructure
systems when operatives might not recognise when systems should be overruled
e.g. in self-driving cars. On the other hand, human error (in data management
say, or in model training data.) can compromise fairness, privacy, security and
compliance. Interactions intended to cause problems by disgruntled employees,
the competition and so on are also a risk.

Example 11 Unintended consequences of AI for organisations

In “Confronting the risks of artificial intelligence” (in McKinsey Quarterly, April 2019) the following
unintended consequences of AI were set out:
Financial performance – algorithms unable to adapt to new trading circumstances leading to sudden
financial losses.
Non-financial performance – use of complex algorithms in hiring and promotion of staff that unintentionally
lead to biased decisions.
Legal and compliance – disclosure of protected personal data.
Reputation – advertising algorithm using personal information causes the public to view the organisation
as intrusive or dishonest.
Economic stability – automated trading algorithms increase volatility in trading markets.
Political stability – manipulation of national institutional processes such as elections through false
messaging.
Infrastructure integrity – intelligent systems lead to overuse/misuse of infrastructure (e.g. GPS cause
unprecedented traffic jams).

5.4 Controls

Controls should have:

 Clarity – Organisations should use a structured approach to identify the highest
priority risks, which will vary by industry and organisation. A healthcare
organisation might focus on issues such as patient misdiagnosis.
 Breadth – Organisation-wide controls should guide the development and use of
AI systems, ensure proper oversight and strong policies, procedures, staff
training and contingency plans should be put in place.
 Nuance – Organisations should ensure specific controls are in place depending
on the nature of the risk. Controls will depend on the complexity of algorithms, the

27
 nature of human-to-machine interaction, the potential for intended exploitation
and the extent to which AI is embedded in a business process.

Controls should be put in place across the entire life of an AI solution, from its
conception, to when its used and monitored. For example:
 At conception – independent review of the purpose of the solution, the proposed
analytic methods and the intended use
 Data management – minimum data and model access requirements, including
prevention of sensitive data download, to secure ‘protected’ data
 Model development – statistically significant input variables reviewed to validate
usability, to remove bias and discrimination
 Model implementation – user training to ensure users do not follow
recommendations without question
 Model use and implementation – systematic tracking, reporting and root cause
analysis of errors, near misses and overrides to guard against failures to override
wrong decisions or overriding correct decisions.

5.5 Ethical Issues

AI raises a range of ethical issues including:

 Job losses: According to a McKinsey Global Institute report in 2017, by the year
2030 about 800 million people will lose their jobs to A1-driven robots. It could be
argued that such jobs are too menial for human beings if they are taken by
robots, and that AI can be responsible for creating better jobs that take
advantage of unique human abilities involving higher cognitive functions, analysis
and synthesis. AI may also create jobs – creating the robots and managing them.
 Wealth inequality: AI could impact the economic flow that ensures economies
grow. Robots do not get paid and do not pay taxes. They can contribute at a level
of 100% with low ongoing costs. Shareholders and senior management could
keep higher proportions of company profits generated by their AI workforce,
leading to greater wealth inequality
 AI is imperfect, and could make a mistake: If well trained, using good data, AI can
perform well. But if AI is given bad data or there are errors in internal
programming, AI can be harmful.

Example 12 AI is imperfect

In 2016 Microsoft’s AI chatbot, Tay, was released on Twitter. In less than 24 hours, because of the
information it was receiving and learning from other Twitter users, it learned to produce racist slurs and
Nazi propaganda. Microsoft shut down the chatbot immediately given huge reputational risk.

 “Technological singularity”: This is the point when technology growth surpasses

human intelligence, potentially making human beings obsolete
 How to treat AI: Should robots have rights, especially if they can be evolved to
the point that they are capable of “feeling”?

28
  AI bias: AI is used increasingly in facial and voice recognition systems which are
vulnerable to biases and errors introduced by those developing the systems. The
data used to train the AI systems can also be biased. For instance, facial
recognition algorithms made by Microsoft and IBM had biases when detecting
gender.

Example 13 Ethical considerations of machine learning

In 2019 ACCA released a report “Machine learning: more science than fiction” which set out the following
ethical considerations:
Professional accountants need to consider, and appropriately manage, potential ethical compromises that
may result from decision making by an algorithm. They must remain engaged in AI and its component
parts, including machine learning.
Dealing with bias: This is one of the biggest ethical challenges for ML. The algorithms, both supervised
and unsupervised, may need to be properly interpreted in order to avoid confusing correlation with
causation.
Strategic view of data: Data is the single most important and non-negotiable requirement for powering the
use of ML. In order to take advantage of data in a sustainable way, an organisation needs a coherent data
strategy.
Assigning accountability: Who takes responsibility for the consequences of decisions made, the human
professional accountant or the algorithm? Dealing with this clearly and consistently will be a key focus for
the years ahead.
Looking beyond the hype: AI has become a ‘buzzword’ in recent years, and ML, as part of AI, has often
attracted similar attention. Unrealistic expectations and the vested interests of those selling this
technology mean that there is also a real risk of the misrepresentation of what is on offer.
Acting in the public interest: Technology can raise universal questions about public good and public value
and professional accountants may find themselves being pulled in different directions as a result.
Defending the public interest requires an ability to go beyond the basic minimum that is required for legal
compliance.

29
Syllabus Coverage

This chapter covers the following Learning Outcomes.

E. Technology and Data Analytics

1. Cloud, mobile and smart technology
a. Discuss, from a strategic perspective the need to explore opportunities for
adopting new technologies such as cloud, mobile and smart technology
within an organisation.
b. Discuss key benefits and risks of cloud, mobile and smart technology.
c. Assess and advise on using the cloud as an alternative to owned hardware
and software technology to support organisation information system needs.

2. Big Data and data analytics

a. Discuss how information technology and data analysis can effectively be
used to inform and implement organisation strategy.
b. Describe Big Data and discuss the opportunities and threats Big Data
presents to organisations.
c. Identify and analyse relevant data for strategic decisions on new product
developments, marketing and pricing.

3. Machine learning, AI and robotics

a. Explain the potential benefits of using artificial intelligence (AI), robotics and
other forms of machine learning to support strategic decisions and the pursuit
of corporate objectives.
b. Assess the risk, control and ethical implications of using AI, robotics and other
forms of machine learning.

Summary and Quiz

 Cloud computing involves the organisation's information systems, applications
and data being held by a third party and accessed via the Internet.
 New technologies such as cloud computing can support business strategies such
as cost leadership and differentiation, and help to find new ways of doing
business.
 Cloud computing offers cost effectiveness and flexibility but increases security
risks and relies on third parties to operate business-critical systems.
 Smart technology includes smart devices, connected devices and Internet of
Things devices.
 Business information systems bring together data from many internal and
external sources and supply up-to-date, accurate, relevant information for
decision making at all levels of the organisation.
 The analysis of big data to reveal patterns and trends can be used for
performance management, forecasting future trends or obtaining qualitative
information about attitudes of potential customers.
 Businesses need to collect and analyse data about the needs of customers to
help make decisions about new product development, marketing and pricing.

30
 Artificial intelligence, robotics and machine learning can bring benefits to
organisations but carry risks and have ethical implications.

Technical Articles

ACCA provides technical articles and other resources to guide and help students.
This chapter includes the relevant content of the following related technical articles
available at the time of writing (November 2022):
 Application of new technology - part 1 (s.1)
 Application of new technology - part 2 (s.1)
 Applications of new technology – part 3 (s.1)
 Getting connected to the Internet of Things (s.1)
 Applying Big Data and data analytics in Strategic Business Leader (s.3)
 A world of intelligent agents (s.5)

For more recent articles and other resources please visit the ACCA global website.

31