Digital Marketing Analysis of Mie Gacoan
Customer at Jakarta Using RFM and K-Means Clustering Methode
Fitra Alghifari
Suhardi1, Andi Saputro2, Amanda Sri Nugroho3,
Jerry Heikal4
Magister
Management, Universitas Bakrie, Jakarta, Indonesia
Amandasrinugroho@gmail.com1, Andisaputro2012@gmail.com2,
Fitra.alghifari14@gmail.com3, Jerry.heikal@bakrie.ac.id4
Abstract
PT. Pesta
Pora Abadi or Mie Gacoan is an Indonesian company
within the Food and Beverages industry that produces food using noodles. Mie Gacoan is currently thriving with hype and sales, and with
those momentum supporting them, they want to take advantage of this moment to
expand their business by analysing their customer to
market more their products, but they still don’t know Segmentasi
is a process of knowing customer characteristics to get profitable customers
for the company. Segmentasi will be done by grouping
the customers into x categories. This study aims to combine K-means
clustering and RFM to analyze the customer segmentation of Mie Gacoan. This customer characteristics will help Mie Gacoan in decision making to prioritize their energy and
resources to potential customers or profitable customers.
Keywords: Customer
Segmentation; RFM; K-Means Clustering
Introduction
PT. Pesta Pora Abadi is an Indonesian
company that is in the food and beverages industry with their product, Mie Gacoan (Fauzi, Permata, & Setiawan, 2023). Mie Gacoan is a franchise about spicy noodles that is currently
booming in Indonesia. It has a wide area restaurant that sells food, snacks,
and drinks that starts at a pretty low price which is around Rp. 10.000.
Development in the era of digitalization is getting faster especially in the
food and beverages industry (Demir & Dincer, 2020). Company must
be able to adapt with technological advancements, market dynamics, and customer
preferences in order to maintain their competitiveness in the industry (Kindström, Kowalkowski, & Sandberg, 2013). Mie Gacoan in making transactions, already records every
transaction digitally, but their transaction data has not been analyzed, so the
company does not know the criteria for customers who make transactions. The
company also still applies the same treatment to all customers. So that there
is no special treatment given by the company to customers who often buy mie gacoan, rarely buy, and
customers who often buy in large and small amounts. The characteristics of
every customer can be defined by RFM model (Recency, Frequency, Monetary) (Asmat, Suryadi, & Govindaraju, 2023). Recency is the
last time a customer buys or does a transaction, Frequency is the amount of
transaction done in a period of time, and Monetary is the amount of money spent
in the customers’ transactions (Widiyanto & Witanti, 2021).
Because every customer has a variety of
characteristics, sometimes company struggle in marketing their products. Which
is why it’s important for companies to know the possibility and characteristics
of their customers in buying their products (Womack & Jones, 2015). In this case,
customer segmentation is required for the company to understand their
customers' varying characteristics more comprehensively, thus the company can
design the correct approach to each different customer segmentation (Hiziroglu, 2013).
Method
The
following are the stages carried out in research based on images from the flow
chart of the research method:
1. Data Collection
At
the data collecting or data collection level, will be used data related to
research. The dataset used is survey
data with total data.
2. Data Analysis
In
this second stage, data understanding or initial investigation will be carried
out to recognize the data used such as what attributes and variables are in the
data. Data analysis is also one of the stages in data processing with the aim
of retrieving important information in the data.
3. Data Pre-Processing
Pre-processing
is the initial process in data processing in exploratory data analysis so that
it will produce data with the appropriate format and ready for use at the next
stage.
The
purpose of pre-processing is to make data easier to use when classifying. The
following are the steps carried out during pre-processing:
1)
Handle Missing Value: find empty data, then drop variables that have
missing values or replace those missing values with average, median or mode of
data.
2)
Handle Outlier: is a value that has an extreme value. The way to handle
data that has outliers is to calculate the inter quartile range.
3) Data
Transformation: converting data categorical and data numeric with one hot
encoded or feature scalling.
4. K-Means Clustering
Implementation
The
next stage is to implement a machine learning model using the K-Means
Clustering model which is one of the unsupervised learning algorithms. In this
stage, the results of clustering from the implementation and division of
clusters into customer segments will be evaluated. The customer segmentation
flow starts from the stage of preparing data then selecting the data to be in
the cluster and determining the cluster value by applying the RFM method. The
clusters for Mie Gacoan customers made by K-means clustering
are as follow:
Based
on table 1.x in cluster 1 the customers are on average 30 years old, live in jakarta ooo, have an income
ranging from 0 to 10 million rupiah, …
Rfm result -> Research Results?
5. Clustering Result
From
the cluster division process, it can be identified the results of customer
segmentation using K-Means. The result of clustering is in the form of groups
of customers who have contributed to sales which are divided into certain
segments according to purchasing behavior. Cluster 1 consist of customers who bla blabla in blablabla
work blablablabalblasb… Cluster 2 is a cluster that
consist of the core customer for Mie Gacoan who buys
several products in one transaction.
The
cluster that contributed to x% of the sales are from cluster 2 and 3, which
consist of 57% of the total respondents.
This cluster will be defined as the core customer for Mie Gacoan. There is also the Cluster 5 that even though
contributes only 20% of the revenue, it consists of many customers that buy Mie
Gacoan products in smaller quantities. They are also
customers who often utilize the promo or discount offered by Mie Gacoan.
Results and Discussion
This research uses K-Means and RFM
(Recency, Frequency, Monetary) analysis to determine company strategy (Rizki, Ginasta, Tamrin, & Rahman, 2020). The table used
in this research is the final cluster center resulting from SPSS data
processing (Kent, Jensen, & Kongsted, 2014). There are 14
categories used in this research, namely Gender, Age, Domicile, Type of work,
Location of Mie Gacoan, Mie Gacoan's
social media knowledge, Monetary, Last purchase (in days), Instagram, Tiktok, Promotions and discounts, Types of promotions and
discounts, Promotion and discount information. If the k-means determination
uses a scale then each category has a scale of 100%, but each item on the scale
has a different scale. However, if the data used is nominal, the results will
show direct calculations. The results of the k-means research used 5 clusters
and the category used to determine value was the highest expenditure of each
consumer (Kuo, Ho, & Hu, 2002).
RFM (Recency, Frequency, Monetary)
results also have an influence on consumer priorities (Safari, Safari, & Montazer, 2016). RFM plays an
important role in determining consumer groups and RFM divides them into three
groups, namely core customers, potential customers and casual customers (Han, Ye, Fu, & Chen, 2014). The
determination of each consumer group is based on total expenditure during the
period October to December 2023. The cluster with the most expenditure will be
classified as the main consumer.
Table 1
Final
Cluster Centers Table
|
Category |
Sub-Category |
Cluster |
||||
|
1 |
2 |
3 |
4 |
5 |
||
|
Gender |
Male |
,49 |
,56 |
,59 |
,64 |
,64 |
|
Female |
,51 |
,44 |
,41 |
,36 |
,36 |
|
|
Age |
|
30 |
27 |
25 |
27 |
24 |
|
Domicile |
Central Jakarta |
,33 |
,21 |
,19 |
,43 |
,17 |
|
South Jakarta |
,67 |
,21 |
,30 |
,00 |
,37 |
|
|
West Jakarta |
,00 |
,21 |
,12 |
,00 |
,11 |
|
|
East Jakarta |
,00 |
,36 |
,21 |
,43 |
,24 |
|
|
North Jakarta |
,00 |
,00 |
,18 |
,14 |
,11 |
|
|
Type of work |
Self-employed |
,33 |
,21 |
,11 |
,00 |
,16 |
|
Businessman |
,00 |
,21 |
,23 |
,43 |
,09 |
|
|
Student |
,00 |
,14 |
,14 |
,00 |
,28 |
|
|
Employee |
,33 |
,36 |
,42 |
,29 |
,41 |
|
|
Civil servants |
,33 |
,07 |
,11 |
,29 |
,07 |
|
|
Unemployed |
,00 |
,00 |
,00 |
,00 |
,00 |
|
|
Incomes |
0-10 Millions |
,33 |
,64 |
,75 |
,71 |
,90 |
|
10-20 Millions |
,67 |
,21 |
,25 |
,14 |
,10 |
|
|
20-30 Millions |
,00 |
,14 |
,00 |
,14 |
,00 |
|
|
#ERROR! |
,00 |
,00 |
,00 |
,00 |
,00 |
|
|
Location of Mie Gacoan |
Central Jakarta |
,67 |
,21 |
,18 |
,43 |
,23 |
|
East Jakarta |
,33 |
,43 |
,26 |
,43 |
,28 |
|
|
South Jakarta |
,00 |
,29 |
,35 |
,14 |
,35 |
|
|
North Jakarta |
,00 |
,07 |
,21 |
,00 |
,13 |
|
|
Mie Gacoan's social
media knowledge |
Instagram |
,67 |
,50 |
,40 |
,43 |
,46 |
|
Youtube |
,33 |
,43 |
,16 |
,29 |
,05 |
|
|
TikTok |
,00 |
,07 |
,37 |
,29 |
,35 |
|
|
Others |
,00 |
,00 |
,07 |
,00 |
,14 |
|
|
Monetary |
Purchase in one transaction |
1.072K |
456K |
177K |
788K |
50K |
|
Last purchase (in days) |
|
32,33 |
15,36 |
26,16 |
14,29 |
42,83 |
|
Instagram |
Do you know Mie Gacoan's
Instagram? |
1,00 |
1,00 |
,93 |
1,00 |
,87 |
|
Tiktok |
Do you know Mie Gacoan's
Tiktok? |
1,00 |
1,00 |
,82 |
1,00 |
,78 |
|
Promotions and discounts |
Knowledge about Mie Gacoan
promos and discounts |
1,00 |
1,00 |
,89 |
1,00 |
,78 |
|
Use Mie Gacoan promos
and discounts |
1,00 |
1,00 |
,86 |
1,00 |
,75 |
|
|
Types of promotions and discounts |
Buy 1 get 1 |
,00 |
,14 |
,12 |
,14 |
,16 |
|
Flashsale |
,00 |
,14 |
,19 |
,00 |
,16 |
|
|
Bundling package |
,00 |
,29 |
,33 |
,14 |
,19 |
|
|
Special day discounts |
,33 |
,36 |
,09 |
,29 |
,10 |
|
|
Discount for minimum purchase |
,67 |
,07 |
,18 |
,43 |
,16 |
|
|
Don't use discounts or promos |
,00 |
,00 |
,09 |
,00 |
,22 |
|
|
Promotion and discount information |
Instagram |
,00 |
,43 |
,37 |
,14 |
,33 |
|
Youtube |
,33 |
,36 |
,11 |
,29 |
,04 |
|
|
TikTok |
,67 |
,21 |
,28 |
,57 |
,34 |
|
|
OnTheSpot |
,00 |
,00 |
,18 |
,00 |
,15 |
|
|
Lainnya |
,00 |
,00 |
,07 |
,00 |
,14 |
|
Based on the company customer cluster results in table 1.1,
several personas from each cluster can be seen. Below are the personas for each
cluster:
● Cluster 1 or Wood Customer
Wood customers have criteria and
personas with male gender 49% and female 51%, age 30 years, for the percentage
of each domicile, namely Central Jakarta 33% and South Jakarta 67%. For type of
work, namely Self-employed 33%, Employees 33% and Civil servants 33%. Income in
this cluster is 0-10 Millions 33% and 10-20 Millions
67%. Based on the location where consumers make purchases in Jakarta, 67% of
consumers buy in Central Jakarta and 33% in East Jakarta. The process of
promoting Gacoan Noodles uses several social media
platforms, in promoting Gacoan Noodles using several
social media platforms and after conducting cluster 1 consumer research,
information was obtained regarding Gacoan Noodles
from Instagram 67% and YouTube 33%. Consumer expenditure in purchasing Mie Gacoan from each transaction is IDR 1,072,000. The last
purchase for cluster 1 consumers is 32 days. Mie Gacoan
has several social media as a means of promotion and after conducting cluster 1
consumer research, they found out that social media was 100% Instagram and 100%
TikTok. Consumer knowledge regarding 100% promos and discounts and the use of
100% promos and discounts. The types of promos used are special day discounts
of 33% and discounts for minimum purchases of 67%. Information obtained regarding
promotions and discounts from YouTube 33% and TikTok 67%.
● Cluster 2 or Silver Customer
Silver customers have criteria and
personas with male gender 56% and female 44%, age 27 years, for the percentage
of each domicile, namely Central Jakarta 21% and South Jakarta 21%, West
Jakarta 21% and East Jakarta 36%. For types of work, namely Self-employed 21%,
Businessman 21%, Student 14%, Employee 36% and Civil servants 7%. Income in
this cluster is 0-10 Millions 64%, 10-20 Millions 21%
and 20-30 Millions 14%. Based on the location where consumers make purchases in
Jakarta, 21% of consumers buy in Central Jakarta, 43% in East Jakarta, 29% in
South Jakarta and 7% in North Jakarta. The process of promoting Gacoan noodles uses several social media platforms, in
promoting Gacoan Noodles using several social media
platforms and after conducting cluster 2 consumer research, information was
obtained regarding Gacoan Noodles from Instagram 50%,
YouTube 43% and TikTok 7%. Consumer expenditure in purchasing Mie Gacoan from each transaction is IDR 456,000. The last
purchase for cluster 2 consumers is 15 days. Mie Gacoan
has several social media as a means of promotion and after conducting cluster 2
consumer research, they found out that social media was 100% Instagram and 100%
TikTok. Consumer knowledge regarding 100% promos and discounts and the use of
100% promos and discounts. Promo types used are buy 1 get 1 14%, flash sale
14%, bundling package 29%, special days discount 36% and discount for minimum
purchase 7%. Information obtained regarding promotions and discounts from
Instagram 43%, TikTok 21% and YouTube 36%.
● Cluster 3 or Diamond Customer
Diamond customers have criteria and
personas with male gender 59% and female 41%, age 25 years, for the percentage
of each domicile, namely Central Jakarta 19% and South Jakarta 30%, West
Jakarta 12%, East Jakarta 21% and North Jakarta 18%. For types of work, namely
Self-employed 11%, Businessman 23%, Student 14%, Employee 42% and Civil
servants 11%. Income in this cluster is 0-10 Millions
75%, 10-20 Millions 25%. Based on the location where consumers make purchases
in Jakarta, 18% of consumers buy in Central Jakarta, 26% in East Jakarta, 35%
in South Jakarta and 21% in North Jakarta. The process of promoting Gacoan noodles uses several social media platforms, in
promoting Gacoan Noodles using several social media
platforms and after conducting cluster 3 consumer research, information was
obtained regarding Gacoan Noodles from Instagram 40%,
YouTube 16%, TikTok 37% and others 7%. Consumer expenditure in purchasing Mie Gacoan from each transaction is IDR 177,000. The last
purchase for cluster 3 consumers is 26 days. Mie Gacoan
has several social media as a means of promotion and after conducting research,
cluster 3 consumers found out that social media was Instagram 93% and TikTok
82%. Consumer knowledge regarding promos and discounts is 89% and use of promos
and discounts is 86%. Promo types are used, namely buy 1 get 1 12%, flash sale
19%, bundling package 33%, special days discount 9%, discount for minimum
purchase 19% and Don't use discounts or promotions 9%.
Information obtained regarding promos and discounts from Instagram 37%, YouTube
36%, TikTok 28%, on the spot 18% and others 7%.
● Cluster 4 or Bronze Customer
Bronze customers have criteria and
personas with male gender 64% and female 36%, age 27 years, for the percentage
of each domicile, namely Central Jakarta 43% and East Jakarta 43% and North
Jakarta 14%. For types of work, namely Businessman 43%, Employee 29% and Civil
servants 29%. Income in this cluster is 0-10 Millions
71%, 10-20 Millions 14% and 20-30 Millions 14%. Based on the location where
consumers make purchases in Jakarta, 43% of consumers buy in Central Jakarta,
43% in East Jakarta and 14% in South Jakarta. The process of promoting Gacoan noodles uses several social media platforms, in
promoting Gacoan Noodles using several social media
platforms and after conducting cluster 4 consumer research, information was
obtained regarding Gacoan Noodles from Instagram 43%,
YouTube 29% and TikTok 29%. Consumer expenditure in purchasing Mie Gacoan from each transaction is IDR 788,000. The last
purchase for cluster 4 consumers is 14 days. Mie Gacoan
has several social media as a means of promotion and after conducting cluster 4
consumer research, they found out that social media was 100% Instagram and 100%
TikTok. Consumer knowledge regarding 100% promos and discounts and the use of
100% promos and discounts. Promo types are used, namely buy 1 get 1 14%, bundling
package 14%, special days discount 29% and discount for minimum purchase 43%.
Information obtained regarding promotions and discounts from Instagram 14%,
YouTube 29% and TikTok 57%.
● Cluster 5 or Gold Customers
Gold customers have criteria and
personas with male gender 64% and female 36%, age 24 years, for the percentage
of each domicile, namely Central Jakarta 17%, South Jakarta 37%, West Jakarta
11%, East Jakarta 24% and North Jakarta 11%. For types of work, namely Self
Employee 16%, Businessman 9%, Employee 41%, Student 28% and Civil servants 7%.
Income in this cluster is 0-10 Millions 90% and 10-20
Millions 10%. Based on the location where consumers make purchases in Jakarta,
23% of consumers buy in Central Jakarta, 28% in East Jakarta, 35% in South
Jakarta and 13% in North Jakarta. The process of promoting Gacoan
noodles uses several social media platforms, in promoting Gacoan
Noodles using several social media platforms and after conducting cluster 5
consumer research, information was obtained regarding Gacoan
Noodles from Instagram 46%, YouTube 5%, TikTok 35% and Others 14%. Consumer
expenditure in purchasing Mie Gacoan from each
transaction is IDR 50,000. The last purchase for cluster 5 consumers was 42
days. Mie Gacoan has several social media as a means
of promotion and after conducting research, cluster 5 consumers found out that
social media was 87% Instagram and 78% TikTok. Consumer knowledge regarding
promos and discounts is 78% and the use of promos and discounts is 75%. Promo
types used are buy 1 get 1 16%, flash sale 16% bundling package 19%, special
days discount 10% and discount for minimum purchase 16% and Don't
use discounts or promotions 22%. Information obtained regarding promos and
discounts from Instagram 33%, YouTube 4%, TikTok 34%, on the spot 15% and
others 14%.
Table
1.2
RFM Segmentasi Table
Identifying the core customer is
essential for businesses to tailor their products, services, and marketing
strategies to meet the specific needs and preferences of this key demographic.
●
Core Customer
The term "core customer"
typically refers to the central or primary target audience for a business or
product. It represents the group of individuals or organizations that are most
likely to benefit from and engage with the offerings of a particular company.
a.
Cluster 2 atau
Entusiast Customer
b.
Cluster 3 atau
Flexible Customer
●
Potential Customer
a.
Cluster 5 atau
Loyal Customer
●
Casual Customer
a.
Cluster 1 atau
Luxury Customer
b.
Cluster 4 atau
Situational Customer
Conclusion
From the data that we have gathered
and analyzed, it can be concluded that the shared value for social media is
“Instagram”, and the shared value of promotions and discounts is “Bundling
Package”. Core customers group that consist of 33% of
Mie Gacoan’s total customer contributes for 52% of
the revenue for October until December 2023. The shared value of cluster is
cluster 2.
It is recommended for Mie Gacoan to do promotions for the core customers group first
by making some new Bundling Package followed by special day discount, and after
that innovate their “buy 1 get 1” promos for the potential customers group, for
example like “buy 2 get 1” or “buy 1 get 2”. For social media, Mie Gacoan should first focus their digital marketing in Tiktok, and Instagram after that.
REFERENCES
Asmat, Fikri, Suryadi, Kadarsah,
& Govindaraju, Rajesri. (2023). Data mining framework for the
identification of profitable customer based on recency, frequency, monetary
(RFM). AIP Conference Proceedings, 2508(1). AIP Publishing.
Demir, Yeliz, & Dincer, Fusun
Istanbullu. (2020). The effects of Industry 4.0 on the food and beverage
industry. Journal of Tourismology, 6(1), 133–145.
Fauzi, Rizal Ula Ananta, Permata,
Zakia Intan Dyah, & Setiawan, Hendra. (2023). The Influence Of Taste And
Product Varıance On Purchase Decısıon Of Mıe Gacoan
Madıun. Asian Journal of Management, Entrepreneurship and Social
Science, 3(03), 1072–1088.
Han, Shuihua, Ye, Yongjie, Fu, Xin,
& Chen, Zhilong. (2014). Category role aided market segmentation approach
to convenience store chain category management. Decision Support Systems,
57, 296–308.
Hiziroglu, Abdulkadir. (2013). Soft
computing applications in customer segmentation: State-of-art review and
critique. Expert Systems with Applications, 40(16), 6491–6507.
Kent, Peter, Jensen, Rikke K., &
Kongsted, Alice. (2014). A comparison of three clustering methods for finding
subgroups in MRI, SMS or clinical data: SPSS TwoStep Cluster analysis, Latent
Gold and SNOB. BMC Medical Research Methodology, 14, 1–14.
Kindström, Daniel, Kowalkowski,
Christian, & Sandberg, Erik. (2013). Enabling service innovation: A dynamic
capabilities approach. Journal of Business Research, 66(8),
1063–1073.
Kuo, R. J., Ho, L. M., & Hu,
Clark M. (2002). Integration of self-organizing feature map and K-means
algorithm for market segmentation. Computers & Operations Research, 29(11),
1475–1493.
Rizki, Bayu, Ginasta, Nava Gia,
Tamrin, Muh Akbar, & Rahman, Ali. (2020). Customer Loyality Segmentation On
Point Of Sale System Using Recency-Frequency-Monetary (RFM) and K-Means. Jurnal
Online Informatika, 130–136.
Safari, Fariba, Safari, Narges, &
Montazer, Gholam Ali. (2016). Customer lifetime value determination based on
RFM model. Marketing Intelligence & Planning, 34(4), 446–461.
Widiyanto, Agus Tri, & Witanti,
Arita. (2021). Segmentasi Pelanggan Berdasarkan Analisis RFM Menggunakan
Algoritma K-Means Sebagai Dasar Strategi Pemasaran (Studi Kasus PT Coversuper
Indonesia Global). KONSTELASI: Konvergensi Teknologi Dan Sistem Informasi,
1(1), 204–215.
Womack, James P., & Jones, Daniel
T. (2015). Lean solutions: how companies and customers can create value and
wealth together. Simon and Schuster.