ANL203 Data Mining Project: Digital Transformation for HeartlandMart

This project report presents an analytical solution to aid HeartlandMart, a supermarket chain, in its digital transformation journey. The assignment requires recommending a data mining approach to transition the business from a brick-and-mortar model to a mixed e-commerce model. The proposed initiative involves using data mining techniques like clustering and association analysis to personalize the online shopping experience, bundle products, and improve customer engagement. The report outlines the business objectives, data mining objectives, and the primary data mining techniques to be used. It also identifies limitations of the project, such as potential customer unfamiliarity with online shopping and initial costs, and suggests solutions like loyalty programs and financial budgeting. The report also details how data analytics can be applied to marketing campaigns and logistics to achieve the business objective of increasing sales revenue. The project includes considerations for implementation, including the roles of involved personnel and technical requirements, aiming for a successful digital transformation for HeartlandMart.

ANL203
Analytics for Decision Making
________________________________________________________________
Group Based Assignment
TG04
Names PI. Number
Tan Si Yuan Z1780624
Edwin Lim B1780525
Mohammad Rashidi Bin Sulaiman E1571461

Date of Submission:13/08/2019

In this assignment, you need to recommend an analytical solution to help an organisation in their
digital transformation journey. In each question, you will be required to complete a set of
activities that are related to the final transformation roadmap. Your report should have the
following structure:
______________________________________________________________________________
1. Executive summary - Fictitious Name: HeartlandMart
2. Current state of the organisation: Only 2 Supermarkets islandwide, one in Ang Mo Kio one in
Hougang
3. Proposed initiative: Transform the business from Brick and Mortar to a mixed e-commerce
business. Users of the website can opt for delivery or self collection.
4. Plans of action and criteria for success: Data mining tools used to bundles products frequently
bought together. First criteria of success is to encourage users to sign up for the online store. At
least 30 sign ups in a month. After 6 months of online store, introduce new data mining tools
(Clustering) to discover what items are frequently bought by people in the same demographics.
Use this data to personalise their shopping experience by showing these items as
recommendations. Offer 5% discount on items that are frequently bought together. Criteria for
success in the next 6 months, at least 60 customers per month using online store.
______________________________________________________________________________
Digital transformation requires rethinking business processes. It’s about using digital
technologies and data to put the customer at the center of your business; drive more value for the
customer

Summary
E-commerce is set to become the world’s largest retail channel by 2021. Enterprises of all sizes,
big and small, have scrambled to set-up their online presence to stay ahead. In order to remain
relevant and competitive, organizations can no longer confine themselves to traditional brick and
mortar store concept. This is the case for HeartlandMart, a neighbourhood supermarket that has
operations in Ang Mo Kio and Hougang vying to establish a presence in the online grocery
shopping scene. HeartlandMart considers putting together a team of experts to digitally
transform its business operations so as to deliver more value for its customers. It aims to improve
customer’s shopping efficiency and improve sales revenue by 15% in the first year.
The use of data analytics can help HeartlandMart achieve this goal. Clustering, a data mining
technique, can be used to analyse store transactions to reveal the relationship between products.
This enables HeartlandMart to bundle products frequently bought together and sell them at a
discount. Moreover, clustering also allows the personalization of e-commerce pages according to
customers’ demographics, increasing customer engagement and retention as it appeals to all
shoppers. However, this requires an abundance of customer data which may not be readily
available and the initial cost of establishing an e-commerce platform is hefty. HeartlandMart can
utilize loyalty programmes to capture customers personal and transactional data. For example,
the use of loyalty cards to track customer expenditure. Budgeting for resources helps reduce the
impact of unforeseen expenditure which reduces sales revenue.
As with every new business, the number of customers is low. HeartlandMart can actively
promote the platform to raise awareness among existing and new customers. And at the same
time design sales promotion to incentivise shoppers to spend recurrently. This will groom
customer loyalty and ensure recurring profits. In order to meet the business objective, expenses
have to be kept at bay. One significant expense is transportation; the delivery of goods to
customers. Poor of transportation cost may eat

Question 1
(a) Discuss a potential business objective and the data mining project that can help to meet this
objective. The illustration (which should not exceed 250 words) should include the following:
(i) State the business and data mining objectives
● Must be quantifiable and measurable. Example:
○ Business objective: Increase book sales to existing customers by 10% in Quarter 3
○ Data mining objectives: Association: Identify books that can be bundled in order
to increase sales
(ii) Discuss the primary data mining technique(s) to be used
(iii) Explain how the project will help achieve the business objective
HeartlandMart is a neighbourhood Brick and Mortar style supermarket that is embracing the
digital age by wanting to incorporate an e-commerce platform that offers online shopping
options for their customers. It operates in The objective is to improve customer’s shopping
efficiency and improving sales revenue by 15% in the first year.
HeartlandMart should first encourage customers to sign up for their online platform and begin
shopping with ease. Customers can choose to self-collect their groceries or arrange for delivery
through the platform. The initial business objective is to achieve 50 sign-ups per month. After 6
months of introducing the platform, HeartlandMart should engage the clustering data mining
technique to cluster their customers based on their demographics (Age, Gender, Occupation).
Clustering these customers together will allow HeartlandMart to identify what are the items
frequently bought by each cluster and curate a personalised shopping page for each cluster. Items
that are bought frequently will be recommended to customers according to the cluster that they
belong to.

At the end of the first year, HeartlandMart should achieve a 15% increase in sales revenue and
have at least 100 e-platform customers per month. The online platform helps HeartlandMall to
achieve their business objectives by reducing the number of customers present in the mart, thus
increasing the efficiency of making purchases, and attracting customers who are too busy to shop
physically. The ease of shopping from a mobile device allows customers to make purchases on
the go and have it delivered to their houses, therefore saving them some time amidst their busy
schedules.
(255 words)
(15 marks)
(b) Carefully study the current state of the organisation, and identify two (2) limitations of the
data mining project stated in Question 1(a).
HeartlandMart currently operates on a Brick and Mortar model in the heartlands of Singapore.
The location of these two marts could mean that a huge portion of their customers could be
elderlies or homemakers who are not familiar with the concept of online shopping. This could
potentially be a problem because they may find it difficult to understand the processes behind
shopping online.
Insufficient data
The lack of usage by the elderly will then reduce the amount of data available when employing
the clustering technique. The lack of information on the elderly will potentially create a less
meaningful experience for the elders who are tech-savvy and uses the online platform. Their
shopping page may not be exquisitely curated to their age because of insufficient data, the
products recommended to them could be based on their gender and occupation instead.
Initial outlay cost
The business objective of this project is to increase sales revenue by 15% by the end of the first
year of introducing the online platform; however, the introduction of this platform may incur
costs that exceeds the increase in sales revenue. The development of the e-platform and data
mining tools requires hiring a new team of experts to run the project. Extensive preparation and

technical planning is also expected during this phase. The costs that are incurred to hire these
talents and to run the online store should be factored in to ensure that the cost is not
astronomical. It is often argued that an initial outlay is necessary to help a business grow in order
to make more profits. However, if the cost of embarking on this project exceeds the 15%
increase in sales revenue, then the business objective would not be met in the first year.
(10 marks)
(c) Discuss with details on one (1) possible way to overcome each of the limitations stated in
Question 1(b).
Introduce loyalty card
In addition to the online platform, HeartlandMart and introduce a loyalty card that allows them to
accumulate points for vouchers. The customers would be required to fill in their personal details
at the time of registering for their loyalty card. Whenever a customer makes a purchase and
presents his/her card, their personal details (including their demographics) along with the details
of their purchases will be sent to a database. Thereafter, clustering techniques can be used on this
database to curate a more reliable and personalised shopping page for all the online customers.
This way, users who belong to the minority cluster will also enjoy the same benefits as the other
customers groups.
Proper financial budgeting and introduce bundle discounts
The cost of starting the online platform could be higher than the potential income it’s going to
bring the organisation in the first year. Therefore, proper financial budgeting must be done to
reduce costs and increase revenue whenever possible. A way to increase revenue using the online
platform is to introduce bundle sales. The data mining process would be able to identify what the
items frequently purchased together and HeartlandMall can offer a 5% discount on these
complementary items when they are purchased together through their online platform.
The introduction of these discounted bundles will attract more customers to make their purchases
online because they can save money and time by using the online platform. This promotion
technique incentivise customers to purchase more products and increase their average spending

at the store, therefore increasing the revenue of the organisation. Additionally, it also helps to
meet the secondary business objective of improving the shopping experience of the customers
and improving efficiency of the business.
(20 marks)
Question 2
Consider the working environment that you have identified in Question 1. Identify two (2)
organisational processes (some examples include marketing campaign and ticketing operation)
that are related to the business objective stated in your answer to Question 1. Discuss how data
analytics can be applied to meet the business objective. (Use up to 150 words per process.)
Marketing Campaign
HeartlandMart’s strategy of expanding and diversifying its revenue stream via an ecommerce
platform is viable only if customer acquisition, engagement and retention are high. User
acquisition begins with the company actively promoting the new platform so that existing and
new customers become aware of it, and at the same time incentivising them to shop online. With
data analytics association analysis can be performed on store transactions to identify products
that are often bought together so that online sales promotion can be better curated to entice
customers’ purchase. For example, buying both apple juice and cereal bar in a bundle online will
entitle customers to receive a $5 off on the next purchase. This serves as a call to action for
customers, and incentives customers to repatronize HeartlandMart physically and/or
electronically. Hence, boosting customer engagement and retention as they are more inclined to
make purchases using the sales promotion.
Logistics

Implementing an ecommerce platform would require HeartlandMart to manage its delivery
logistics. However, as with all delivery service it is timely and costly if poorly coordinated.
Transportation cost may skyrocket eating into sales revenue and unnecessary manpower hours
may be wasted. In other words, managing delivery route and time is key to reducing shipping
cost. Data analytics can be utilized to schedule common delivery timings to reduce the number of
delivery trips. This is done by clustering orders and preferred delivery timing according to
various districts in Singapore (e.g. Tampines, Pasir Ris etc.). For example, if the majority of the
orders in Tampines opted for the 6.00pm – 8.00pm delivery time slot, HeartlandMall can
effectively and efficiently dispatch resources to complete the delivery. This reduces the need for
multiple trips or re-route compared to if it was uncoordinated.
(30 marks)
Question 3
Based on the details covered so far, answer the following questions:
Who should be involved in implementing the proposed initiative, and what should the scope of
their roles be?
How does the proposed initiative help fulfil the business strategy and IT strategy?
What are the possible technical considerations and requirements for implementation?
Give your recommendations in no more than 500 words.
(25 marks)
There are four different groups of people who should be involved with the development of the
proposed initiative. The first and foremost group is the sales manager who will provide key
insights for the project based on his understanding of how customers shop in their stores. The
next group is the project management team who would be responsible for managing the project.

This consists of the project manager and any personnel who he or she deems fit for assisting. The
next group would be the system developers. These would include people who are familiar with
architecting, designing and developing of the proposed system. Afterwards, the third group
would be the unit testers who would be responsible for carrying out the testing tasks of the
project. Once the testing of the project is done and made live for the public, there would be three
new groups with distinct roles in each group. The first group would be to handle the content part
of the website. This would involve uploading products, writing descriptions, adding offers,
preparing newsletters among others. The next group would be in the technical side of the
website. These would include people who are optimizing database, who are running analytics on
the website, also who are involved with the security aspect of the site among others. The third
and final group would be customer-oriented group who would be responsible for handling the
marketing and customer relationship building. These would include things such as creating
marketing campaigns, running Ads, understanding analytics generated by the customer’s digital
footprints and so on.
The proposed initiative would help create an online presence for the HeartlandMart. Customers
would be able to purchase the grocery as they intended and would also be able to subscribe
newsletter, make account, add products to cart and purchase the products later on. The shop-
owners would be able to fulfill customer’s purchases through the use of inventory management
system, delivery management system as well as through the use of 3rd party logistic providers.
The system would be robust enough to handle customers during peak times during the day such
as late evening, weekends as well as during festival times. As and when customers make
purchases or even browse products as well as respond to digital advertisements, it will create a
trail of the customer’s activity on the web as well as help in making a digital identity or persona
of the product. This data then can be used to provide personalized recommendations to the
customers while they are surfing on the web, bundle products together, provide personalized
recommendations on email and SMS as well as serve the customers in a better way.
A robust cloud-scalable server would be needed to host the website. An SSD based and at least
an 8GB RAM based Linux server should be capable enough to sustain the load of the website.
Furthermore, the server would be configured with Linux, Apache, MySQL and PHP
technologies. Additionally, the WordPress would be used as the main content management

system along-with WooCommerce addon that would help in creating the e-commerce part of the
website. Lastly, the site will be configured with Google Analytics to learn from the user
behaviour as well as complemented with several addons and optimizations to make the site fast,
secure and trustworthy.