MARKETING STRATEGY OF MOBILE PACKAGE SERVICE XYZ

WEST JAVA REGION

Aryoputro WICAKSONO1, INDRAWATI2

12Faculty of Economics and Business, Telkom University, Bandung, Indonesia

Corresponding author: 1aryoputro.wicaksono@gmail.com, 2indrawati@telkomuniversity.ac.id

Abstract. The underperformance of XYZ company in West Java Region in the year 2015 to
2019, despite the massive increase of broadband services usage, had initiated this research to
be undertaken. This research used a mixed-method approach to determine the variables used
for customer segmentation. There is a total of seven variables used as a parameter in segmenting
customer behaviour. Big data analysis based on Call Data Records (CDR) Summary is
processed using K-Means clustering method. The result of the data analytic will be used as the
basis for determining the main marketing target and the right product tactics of XYZ Mobile
package services in West Java. The results of this research illustrated the behaviour of XYZ
West Java Region customers and highlighted what should be optimized to provide the right
mobile marketing product tactics to improve ARPU customers and eventually increase revenue.

Keywords: mixed-method, segmentation, K-means, big data, product tactics, mobile plan
service

Introduction

Telecommunication technology is developing very rapidly. The increase in the use of

4G/LTE technology that offers broadband internet speed is one of the causes of loss of
revenue gained from legacy business (voice and SMS), which up until recently has been
the backbone revenue in the telecommunication business. Competition is getting tougher
because customers no longer need to be loyal to one operator, but rather otherwise change
according to needs. There are more demands for finding a mobile package service that is
tailored to suit customers’ needs (Arora, A., Vohra, R.,2014). Presently, XYZ customer
segmentation approach is based on two basic attributes, i.e. Average Rate per User (ARPU)
and Length of Stay (LOS). The value of ARPU is the average monthly expenditure spent
by customers to buy XYZ Mobile package service while LOS is the duration of which
customers are using XYZ services.
The customer segmentation approach conducted by XYZ as explained above has not
given any significant effect to XYZ Mobile package service sales, which ultimately affects
XYZ's revenue. Most of the year 2019, the target set by the company was not achieved.
The overall achievement of XYZ of West Java Region amounted to 94% of the target set
by Management. This of course affects the performance of the company in which the
revenue achievement of XYZ within 2015 to Q3-2019 had suffered a declining trend as of
22.29% (Annual Report Telkom Tbk. 2015 to Q3-2019).
Previous research by Sagala, F., Y., P., Ramantoko, G (2018) discussed the most
influential variables on Telkomsel broadband revenue. The goal of said research is to offer
the best data package for Telkomsel post-paid customers for chat/communication usage
(Whatsapp, LINE, Telegram), Social Media, and video/entertainment. Whereas Insani, R.,
Soemitro, H. L (2016) conducted customer groupings to identify loyal customers,
profitable customers, and customers who will churn, and then a special approach strategy
is done without discussing the details of the mobile package service offered. Research by
Setiawan, W., B., Tricahyono, D (2019) focused on Telkomsel's marketing strategy in
increasing penetration of 4G device associated with 4G network development; it is not a
discussion about the grouping of customers based on the purchasing power in a region and
also the mobile package service offered. The increased revenue of Internet Data Services
from the broadband service and Digital Business Services has not been able to provide the
optimal contribution to the loss of revenue from the Legacy business. Therefore, this
research is analyzing the customers’ behaviour in using XYZ Mobile package service in
the West Java region in order to increase the Average Rate Per User (ARPU).
This journal discusses XYZ mobile package service marketing strategy. Several phases
conducted in this research are segmenting customer's geography, based on the purchasing
power level of the community in a region; determining variables to be processed for
segmentation of customer behaviour, and analysing big data based on customers’ Call Data
Records (CDR) Summary in February 2020 using K-Means clustering and Phyton Jupyter
Notebook software. The results of this data processing are product tactics and the exact
target of XYZ mobile package service based on customer characteristics in each region of
West Java.
Literature Review

1. Marketing
Marketing is the main reason a company expands its business. Without marketing, a
company could not thrive and grow. In a more modern concept, marketing has a broader
scope than sales, because sales activities only focus on the number of products that are
successfully sold and how to increase the next sales. While marketing activities
(marketing) is a comprehensive and complex system to support the business of the
company. Successful marketing activities will establish requests for products or services
(Kotler and Keller, 2016).
2. Marketing Strategy
The business strategy of a company is closely related to marketing. Whatever the
company's profit-related activities, they are inevitably executed based on the marketing
strategy. Marketing strategies define the target market and the related marketing mix. This
is a big picture of what the company will do in a specific market. It takes two interrelated
sections:
 • Target Market – A group of homogeneous customers have a resemblance, similar,
that will be approached to be offered product/service from a company.
• Marketing Mix – Multiple variables that can be controlled, united by the company
to satisfy a targeted group (Perreault et al., 2017).
3. Segmentation
It is a process to identify market conditions as well as sorting market conditions
according to their needs, characteristics, or purchase patterns. A market segment is a
collection of some customers or consumers who share their needs, have similar interests,
and certainly have a considerable amount of interest. Therefore, the market needs to be
grouped into smaller parts, where the similarities of characteristics, buyers, and needs or
behaviors become the basis of the grouping. According to Kotler and Keller (2016), it can
be distinguished into the following variables based on customer characteristics:
• Geographic segmentation: a segment that divides the market into different
geographical units such as cities, provinces, regional, country, or environment
where the customer resides.
• Demographic segmentation: segmentation by dividing the market into groups
based on variables such as age, gender, occupation, family cycles, and customer
income.
• Psychographic segmentation: a segment of market divided into different groups
based on classes such as motivation, perception, personality, lifestyle, social class
of customers.
• Behavioral segmentation is a segment of market divided into different groups
based on attitude, use, and response from customers.
4.Targeting
After doing segmentation by identifying opportunities that exist in the market, the next
step is to evaluate which segments will be used as the target market. The target market is
a group of customers who share specific criteria and are in the same segment, of which the
company is set to direct all marketing activities for profit to. There are 2 (two) types of
primary targets that are shared based on demographic information – who they are and
customer behaviour information – what they do (Kingsnorth, 2016). It was reaffirmed by
Chaffey and Chadwick (2016) that the customer's targeting process can be done based on
the customer journey behaviour, such as the content, products or services that are
commonly accessed, and what activities can be done in social media.
5. Positioning
Positioning is the company's position from the customer's point of view compared to
similar competitors. This includes the brand image, benefits and advantages promised to
the customer, the product excellence (competitive advantages) which is the reason why
consumers buy the marketed products. The main point to note is that this positioning is not
what has been done to a product, but rather about what is done to attach the brand to the
memory of the target customer, the product must be perceived and enjoyed by the target
customer to get a clear picture of the product compared to similar competitors (Tien and
Anh, 2017).
6. Marketing Mix
Marketing a product or service in a company required an easy and flexible approach,
commonly referred to as marketing Mix (Marketing). Marketing Mix is a classic tool used
to help the planning of "what is offered to the customer" and "how to offer the product to
customers"(Kotler et al., 2017).

Methodology

1. Respondent and Data Processing

Selected respondents are obtained from both XYZ and non-XYZ resources. The
internal respondents have experiences of more than 12 years in marketing and strategic
positions in running the company. External resources are academists and professionals in
the field of marketing, consisting of loyal customers of XYZ and non-XYZ customers. The
results of the interviews of the said experts are interpreted in the form of data to be analysed
quantitatively and qualitatively.

Table 1. List of Respondents

Category Definition List of Respondent

Internal Respondents Work in XYZ and have General Manager
experience of more than 12
(twelve) years in Marketing. Manager
Respondents have strategic Manager
positions to determine and
execute XYZ’s strategy. Manager
External Respondents XYZ’s Customer Telkom University
Non XYZ’s Customer Consumer Goods
Banking

In the first phase, secondary data processing from the Statistical Central Bureau for
regional grouping based on the purchasing power of the public was carried out. The second
stage, using the customers’ Call Data Record (CDR) summary as secondary data to classify
behaviours of customers with minimum monthly usage of IDR 2,500.00 (two thousand five
hundred rupiahs) according to XYZ's minimum package value in the market
(https://www.telkomsel.com/en/promotions, 25 May 2020). Data to be processed is as
many as approximately 8 million (eight million) Mobile Subscriber Integrated Service
Digital Network Number (MSISDN). Both stages are processed with K-Means clustering
using Jupyter notebook software.
2. Geographic and Demographic Segmentation
Based on the results of the literature review of several previous studies, a demographic
variable to facilitate the grouping of target markets is required, especially regarding the
needs, desires, and the consumption level of a product (Kotler, Keller, 2016). It is necessary
to know the purchasing power of people in one area as it will have an impact on the
marketing strategy of XYZ Mobile package service in the community. There are 4 (four)
indicators that are used as parameters of the Human Development Index (HDI), stating
how the population can access the development results in obtaining income, health,
education, and so on (Rory, 2019), (Nina, G.A., Rustariyuni, S.D., 2018). Per capita
consumption and expenditure of a month is the cost incurred for the consumption of all the
household members for a month either derived from the purchase, delivery, and production
itself divided by the number of household members in the household (Indikator
Kesejahteraan Rakyat, Welfare Indicators, 2015), (Nina, G.A., Rustariyuni, S.D., 2018).
Gross Regional Domestic Product (PDRB) is an important indicator to determine the
economic condition of an area within a certain period, either based on the prevailing price
or the basis of constant price (Giovanni, R., 2018), (Amanaturrohim, H, Widodo, J., 2016).
The percentage of poor people, poverty is seen as an inadequacy on the economic side to
fulfil the basic needs of food and not food consumption measured from the side of
expenditure (BPS West Java).

3. Customer Behaviour Segmentation

The next stage is customer grouping based on behavior. There are several variables
based on telecommunication services used based on the literature review in Table 2.
Previous researches had shown 4 (four) variables used in determining the behavior of
customers based on the use of telecommunication services i.e. broadband, digital service,
voice, and SMS.

4. K-Means Clustering
K-Means is the most popular method of clustering and often used to process data
analysis or data mining in the form of modeling without supervision (unsupervised) by
grouping the data system on a partition. The K-means method is the simplest and most
common clustering method. K-Means has the advantage of being easy to implement and
run, and the time it takes to learn this method is relatively fast, easy to adapt as well as
commonly used (U. R. Raval and C. Jani, 2016)
In this study, the K-Means Hard Clustering method will be done in which each data is
only used for 1 (one) cluster. The steps in performing clustering with K-Means are as
follows:
1. Specifying the number of clusters.
2. The clustering process starts with identifying the data to be clustered, xij (i = 1, 2, 3,…,
n; j = 1, 2, 3,…, m) with n as the amount of data to be clustered and m as the amount
 of variable.
3. At the beginning of the literature, data will be inserted into the cluster randomly, ckj
(k = 1, 2, 3,…; j = 1, 2, 3,…).
4. Centroid calculation is performed on the existing data from each cluster using the
Euclidean method.
5.
𝟐
𝑑𝑖𝑘 = √∑𝒎
𝒋=𝟏(𝒙𝒊𝒋 − 𝒄𝒌𝒋 ) (1)

Where
dik = range between data
xij = first data point (center of cluster)
ckj = second data point
m = amount of data attribute
5. Allocate each of the data to the nearest centroid.
6. Redo the fourth step when there is data cluster movement or if there is a change in the
centroid value.
7. A fixed centroid value means the clustering process is complete.
8. Using elbow analysis, by looking at the percentage result of the comparison between
the number of clusters that will form an elbow at one point by calculating the Within
Cluster Sum of Square Error (WCSS) to ensure the most appropriate cluster count of
the processed data

𝐶𝑛 𝑑𝑚

𝑊𝐶𝑆𝑆 = ∑ ( ∑ 𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒(𝑑𝑖 , 𝐶𝑘 )2 ) (2)

𝐶𝑘 𝑑𝑖 𝑖𝑛𝐶𝑖

Where
C = cluster centroids
d = data point in each Cluster
9. Validate cluster accuracy by performing the index silhouette measurements for Lot
samples, the average result of the Silhouette Index on the Lot clusters, and the overall
average Silhouette Index for the dataset. The Silhouette Index is calculated by using
the Silhouette Coefficient (SC),

Silhouette Coefficient = (x-y)/max (x,y) (3)

where x is the average of the closest cluster distance (the average distance between the
samples with the closest cluster) and Y is the average distance of intra-cluster (the
average distance between the samples in the same cluster).
 Table 2. Telecommunication services variables based on literature review

Widodo, Amiri, Duenias, Ovando,C.,

Dasi, A Sugiyanto This
Variables et.al Setl. M et.al
etl. (2017) (2018) study
(2018) (2017) (2015) (2015)
Broadband
✔ ✔ ✔ ✔ ✔
Revenue
Digital
Service ✔ ✔ ✔
Revenue
Voice
✔ ✔ ✔ ✔
Revenue
SMS
✔ ✔ ✔ ✔
Revenue
Total
✔
Revenue
International
✔
Roaming
Others ✔

Results and Discussions

1. Geographic and Demographic Segmentation

Data from Human Development Index (HDI), expenditure consumption index Per
capita a month (IKPpK), the gross regional domestic product based on applicable price
(PDRB) and percentage of the poor population (PPM) of 18 (eighteen) Cities/districts in
Telkomsel region of West Java province will be used in segmenting people purchasing
power based on geographic and demographic properties. The K-Means Clustering analysis
for geographical and demographical segmentation here uses the value of k = 3, which
means there will be 3 large groups from 18 areas that have resemblance or similarities
based on socio-economic conditions and the population of the community. The results of
this geographic segmentation are shown in table 3 below,
 Table 3 Geographic Segmentation
Source: Data Processing K-Means Clustering
Cluster 1 Cluster 2 Cluster 3
Region Cianjur Kota Kabupaten
(City/District) Garut Bandung Bandung
Tasikmalaya Indramayu
Ciamis
Kuningan
Kab. Cirebon
Majalengka
Sumedang
Subang
Kab. Bandung Barat
Pangandaran
Kota Cirebon
Kota Cimahi
Kota Tasikmalaya
Kota Banjar
Four variables are used for the geographical segmentation i.e. Human Development
Index (HDI), the expenditure consumption index Per-capita a month (IKPpK), PDRB
based on the prevailing price (PDRB), and the percentage of poor people (PPM) of 18
(eighteen) cities/districts in the Telkomsel Region of West Java. The results obtained are
divided into 3 where cluster 1 consists of 15 (fifteen) regions, cluster 2 consists of 1 (one)
region namely the city of Bandung, and cluster 3 consists of 2 (two) areas namely Bandung
District and Indramayu.

2. Customer Behaviour Segmentation

Based on the results of the interview with the respondents, some variables need to be
added in this study, shown in Table 4.
Table 4. Research Variable: Interview Result
Source: Interview Data Processing
Variable i1 i2 i3 i4 i5 i6 i7 Percentage
Broadband
✓ ✓ ✓ ✓ ✓ ✓ 86%
Revenue
Digital Service
✓ ✓ ✓ ✓ ✓ ✓ 86%
Revenue
International
✔ ✔ ✔ ✔ ✔ ✔ 86%
Roaming Revenue
Voice Revenue ✓ ✓ ✓ ✓ 57%
SMS Revenue ✓ ✓ ✓ ✓ 57%

Revenue Others ✔ ✔ ✔ ✔ 57%

 Total Revenue ✔ ✔ ✔ 43%

i = Interviewee /Informant
In the previous literature review for customer behavior segmentation, the variables are only
voice revenue, SMS revenue, broadband revenue, and digital service revenue. There are 3
(three) additional variables i.e. total Revenue, international Roaming Revenue, and Other
Revenues. Total revenue will be used as a value of Average rate Per User (ARPU) in
determining Customer Class (high or low-value customer) (Widodo and Ramantoko,2018).
In total there are 6 (six) variables used in determining segmentation based on customer
behavior:
1. Voice Revenue
2. SMS Revenue
3. Broadband Revenue
4. Digital Service Revenue
5. International Roaming Revenue
6. Others
Segmentation of customer behavior will be done in each cluster based on the geographic
segmentation resulted from table 3 above. Whereas Cluster 1 consists of 15 (fifteen)
cities/districts area, Cluster 2 consists of The City of Bandung, and Cluster 3 consists of
the district of Bandung and Indramayu. Telkomsel classifies customers based on the
monthly Average Rate Per User (ARPU) value. There are 4 (four) customer classifications:
Silver customers (ARPU <100K IDR), Gold customers (100K IDR – 299K IDR), Platinum
customers (300K IDR – 999K IDR), and Diamond customers (ARPU ≥1000K IDR) for the
average use of 3 months. (https://www.telkomsel.com/en/diamond,
https://www.telkomsel.com/en/platinum, https://www.telkomsel.com/en/gold, July 01
2020). With the launching of various new packages containing nominal value below IDR
50,000.00 with 30 days validity period such as As OMG! 4G
(https://www.telkomsel.com/en/node/152489, July 01, 2020), Kartu As Perdana Internetan
Puas 5K (https://www.telkomsel.com/en/kartu-as/kartu-perdana-kartu-as, July 01, 2020),
SimPati Internet OMG! 4G (https://www.telkomsel.com/en/node/152473, July 01, 2020),
and also SimPati daily retail package (https://www.telkomsel.com/en/simpati/paket-
simpati?page=2, July 01, 2020), daily As package
(https://www.telkomsel.com/en/node/149657, July 01, 2020). To refine the analysis in this
study, Silver class customers will be divided into two groups with ARPU of 2.5K IDR –
IDR 49.99K IDR and ARPU of 50K IDR – 99.99K IDR.
The following Fig. 1 describes the number of users (MSISDN) based on the Customer
class (ARPU) of each cluster,
 Fig. 1 ARPU Comparison per Cluster

Fig.1 shows that the Silver Customer Class with ARPU of IDR 2.5K IDR – 49.99K
IDR dominates in all clusters 1, 2, and 3. Also, when viewed from each class of customers
(Silver, Gold, Platinum, Diamond), there is a resemblance to the number of
customers/users of Cluster 1 and Cluster 3 in each of ARPU magnitude. There is a
similarity in the purchasing power levels of XYZ mobile package services on these two
clusters. In Silver Customer Class (50K IDR – 99.99K IDR), Gold, Platinum, and
Diamond, the number of subscribers/users is dominated by customers from Cluster 2. This
indicates that in cluster 2 the purchasing power level of the mobile package service is
higher than in both cluster 1 and cluster 3. The dominant mobile package services used by
each customer class in each cluster can be seen in Table 5 below.
 Table 5. Customers’ Dominant Services
Cluster 1 Cluster 2 Cluster 3
ARPU Dominant Dominant Dominant
% % %
(IDR) Service Service Service
2.5K - 49.9K Voice 45,20% Voice, SMS 51,58% Broadband 46,82%
(SILVER)
Broadband 36,10% Broadband 38,67% Voice 38,12%
Voice, SMS 10,41% Digital 9,64% Voice, SMS 7,88%
Service,
Voice
Digital 8,02% International 0,11% Digital 7,12%
Service Roaming Service,
Voice
International 0,05% International 0,05%
Roaming Roaming
50K - 99.9K Broadband 66,46% Broadband 75,30% Broadband 73,18%
(SILVER)
Voice 24,10% Voice 16,84% Voice 18,89%
Digital 4,70% Digital 4,19% Digital 4,25%
Service, Service, Service,
Broadband, Broadband, Broadband,
Voice Voice Voice
Voice, SMS 4,25% Voice, SMS 1,90% Voice, SMS 3,03%
Others 0,46% Others 1,72% Others 0,61%
International 0,03% International 0,05% International 0,03%
roaming roaming roaming,
Voice,
Broadband
100K - Broadband 74,10% Broadband 62,70% Broadband 74,98%
299.9K
Voice, SMS 16,66% Voice, 21,94% Voice, 12,58%
(GOLD)
Broadband Broadband
Digital 5,34% Others 8,11% Digital 5,25%
Service, Service,
Broadband, Broadband,
Voice Voice
Others 3,82% Digital 5,81% Others 4,72%
Service,
Broadband
International 0,08% International 0,14% Voice, SMS 2,37%
Roaming Roaming,
International 0,10%
Broadband
Roaming,
Broadband
300K - Broadband 63,63% Broadband 47,44% Broadband 60,23%
999.9K
Voice, 20,45% Voice, 33,90% Voice, 22,46%
(PLATINUM)
Broadband Broadband Broadband
Digital 9,32% Digital 8,08% Digital 9,82%
Service, Service, Service,
Broadband Broadband Broadband
 Cluster 1 Cluster 2 Cluster 3
ARPU Dominant Dominant Dominant
% % %
(IDR) Service Service Service
Voice, SMS, 2,74% Others, 5,40% International 2,82%
Broadband Voice Roaming,
Broadband
International 2,07% International 3,69% Voice, 2,35%
Roaming, Roaming, SMS,
Broadband Broadband Broadband
Others 1,78% Voice, 1,49% Others, 2,32%
SMS, Voice,
Broadband Broadband
≥ 1000K Digital 46,16% Voice, 30,26% Digital 47,48%
(DIAMOND) service, Broadband Service,
Broadband Broadband
Voice, 19,27% Digital 27,34% Voice, 19,09%
Broadband Service, Broadband
Broadband
Broadband 12,20% International 15,49% Broadband 14,67%
Roaming
International 10,57% Others, 15,33% International 9,78%
Roaming Voice Roaming
Others, Voice 9,30% Broadband 10,25% Others, 7,26%
Voice
SMS, 2,49% SMS 1,32% SMS 1,74%
Broadband

On the ARPU of 2.5K IDR – 49.99K IDR (Silver Class), Cluster 1 is dominated by
customers using voice service (45,20%), Cluster 2 by voice and SMS Services (51,58%),
and cluster 3 by broadband service (46,82%). On the ARPU of 50K IDR – 99.99K IDR
(Silver Class), the use of broadband service dominated as much as 66,46% in Cluster 1,
75,30% in Cluster 2, and 73.18% in Cluster 3. On ARPU of 100K IDR – 299.99K IDR
(Gold Class) is also dominated by customers who use broadband service as much as
74.10% in Cluster 1, 62.70% in Cluster 2, and 74.98% in Cluster 3. Broadband service also
dominates on the ARPU of IDR 300K IDR – 999,99K IDR (Platinum Class), 63.63% in
Cluster 1, 47.44% in Cluster 2, and 60.23% in Cluster 3. While on ARPU ≥1.000K IDR
(Diamond Class), is customers who use Digital service and broadband service dominated
in Cluster 1 (46.16%) and also on cluster 3 (47.48%), while voice and Broadband Services
use is 30.26% in Cluster 2.
Table 6 displays Mix Marketing Matrix, which results in customer segmentation in each
Cluster 1, 2, 3, the customer value, and also the mobile Service package used in Telkomsel
Regional West Java area.
 Table 6. Marketing Mix Matrix

100K - 300K -
2.5K - 49.9K 50K - 99.9K ≥ 1000K
299.9K 999.9K
(SILVER) (SILVER) (DIAMOND)
(GOLD) (PLATINUM)
Single Package
√ √ √ √ √ √
Voice
Single Package
√ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Broadband
Bundling
Package Voice, √ √ √ √ √ √ √ √
SMS
Single Package
√
Digital Service
Single Package
International √ √ √ √ √ √ √
Roaming
Bundling
Package Digital √ √
Service, Voice
Bundling
Package Digital
Service, √ √ √ √ √
Broadband,
Voice
Single Package
√ √ √ √ √ √ √
Others
Bundling
Package
International
√
Roaming,
Voice,
Broadband
Bundling
Package Voice, √ √ √ √ √ √ √ √
Broadband
Bundling
Package Digital
√ √ √ √ √ √
Service,
Broadband
Bundling
Package Voice,
√ √ √
SMS,
Broadband
Bundling
Package Others, √ √ √ √ √
Voice
Bundling
Package
International √ √ √ √ √
Roaming,
Broadband
 100K - 300K -
2.5K - 49.9K 50K - 99.9K ≥ 1000K
299.9K 999.9K
(SILVER) (SILVER) (DIAMOND)
(GOLD) (PLATINUM)
Bundling
Package Others,
√
Voice,
Broadband
Bundling
Package SMS, √
Broadband
Single Package
√ √
SMS
Increasing customer's Average Rate Per User (ARPU), broadband and digital
Objective
service penetration in Telkomsel West Java Region
Marketing Mix
Product The mobile package offers should be customizable; customers should be able to
make their selections according to their needs for voice, broadband, and digital
services (co-creation).
The largest market (56,81%) is the silver class customer's ARPU 2.5K-49.9K.
Price Need more mobile packages services variance less than 50K to attract customers
for increasing voice traffic, broadband, and digital usage services.
1. Conducting digital channel campaigns regarding XYZ Services on social media
platforms (Instagram, Facebook), Youtube Ads, celebgram endorsement, SMS,
pop up messages, outbound calls, and on an online media platform.
Promotion 2. Conducting traditional campaign regarding XYZ services by utilizing
Distributors Agents, Reseller Outlets, Mobile Grapari to bring services closer to
customers, using above the line promo media (billboards, billboards, video
profiles, TV and Radio Talkshow).
Utilizing both traditional channel distribution channels (Distributor Partners,
Place Reseller Outlets, Mobile Grapari) and digital media platform on social media, and
also MyTelkomsel application for improving customer satisfaction.
Improving big data analysis capability for the marketing team and hospitality for
People frontliner. Agility, adaptability, and technology update capability need to be
improved periodically for all XYZ's teams.
Process Do a periodic review process related to mobile packages service on the market.
Optimizing big data analysis and data mining for personalizing mobile package
service promotion to each customer.
Physical Focus on optimizing network quality and commercially developing 5G networks to
Evidence provide service excellence. Maintaining the quality of distribution channels, online
media campaigns (social media platforms, SMS), and offline as well as virtual
assistant services.

There are 17 mobile package services offered in XYZ West Java Region based on
customer behavior. Not all of the mobile package services are available in each cluster.
Some are single packages and some come in bundling packages available exclusively only
in Cluster 1, 2, or 3.
 Conclusion

The different purchasing power levels of customers in each cluster can be distinguished by
comparing the number of MSISDN (Mobile Subscriber Integrated Service Digital Network
number) with the ARPU (Average rate Per User) value of these clusters. The four indicators
of the BPS affect the behavior of customers in the selection of Mobile package services
used in their activities. Segmentation by region/area will enable analysis and finding of
any indicators that eventually affect the winning in sales of an area. Some areas have high
rates of non-adoption of internet access, some of non-adoption is linked to demographic
features of users, such as older age, low income, and less education (Greenstein, 2020).
Sales transactions and promotion methods offered will be more directed, effective, and
focused. Geographical Segmentation will also facilitate mapping of which region/area will
be prioritized in product marketing.
Additional variables of customer behavior in this research i.e. International roaming
and others are used as differentiators and as a complement for previous researches. This
will refine customer grouping analyses based on their behavior.
By looking at the results of processing big data analysis of 3 (three) clusters in the XYZ
region of West Java area, it can be concluded that the broadband service becomes a service
of choice in almost all classes of customers. But even so, the legacy service is still the
favorite for some Silver class customers (ARPU 2.5K IDR – 49.99K IDR) on cluster 1
(voice) and cluster 2 (voice and SMS). Customers with ARPU value of 2.5K IDR – 49.99K
IDR are the customer group that uses XYZ Mobile package service the most. It is also
worth considering to add product variations to this customer class segment. Adding
bonuses and promo to keep them interested in increasing monthly usage. Digital services
need to be bundled with other services such as broadband and voice. This is needed to
increase the use of digital services. International roaming and others are very special
services and have their specific customers. Users of international roaming services are
dominated by Platinum class customers (300K IDR – 999,99K IDR) and Diamond
customers (ARPU ≥1.000K IDR). While another service is dominated by Silver customers
(ARPU <100K IDR) and Gold customers (100K IDR – 299K IDR). Looking at the results
of data analysis for this service, bundling offers with broadband service can be done. The
customer segment of this service user can be categorized as a niche market that requires
special treatment to be further explored.
Another product tactic is conducting digital channels and traditional campaigns
regarding XYZ Services on social media platforms (Instagram, Facebook) SMS, pop-up
messages, outbound calls, and utilizing Distributors Agents, Reseller Outlets, and Mobile
Grapari to bring services closer to customers. The mobile package offers should be
customizable by optimizing big data analysis and data mining. More and more
telecommunication businesses are ruined due to tight competition leading to a price war.
So, instead of falling in the price war with competitors, the writer suggests XYZ focuses
on optimizing network quality, commercially developing 5G networks to provide service
excellence, maintaining the quality of distribution channels, and online media campaigns
(social media platforms, SMS), and offline as well as virtual assistant services. In the
broadband internet industry access to broadband and the broadband quality are important
for consumer and social welfare (Boik, A., Takahashi, H., 2020).

References
Amanaturrohim, H., & Widodo, J. (2017). Pengaruh Pendapatan dan Konsumsi Rumah Tangga
Terhadap Kesejahteraan Keluarga Petani Penggarap Kopi di Kecamatan Candiroto Kabupaten
Temanggung. Economic Education Analysis Journal, 5(2), 468. Retrieved from
https://journal.unnes.ac.id/sju/index.php/eeaj/article/view/13579
Amiri,S., Woodside,J.M (2017). Emerging markets : the impact of ICT on the economy and society.
Emerald Insight, Digital Policy,Regulation and Governance, Vol.19, Issue:5,pp.383-396. Retrieved
from https://doi.org/10.1108/DPRG-04-2017-0013.
Annual Report PT. Telkom tahun 2015, 2016, 2017, 2018, Q3- 2019.
Arora, A., Vohra, R.(2014), Segmentation of Mobile Customers for Improving Profitability Using Data
Mining Techniques. Retrieved from https://www.ijert.org/segmentation-of-mobile-customers-
using-data-mining-techniques.
Badan Pusat Statistik Jawa Barat, Sosial dan Kependudukan. Retrieved from
https://jabar.bps.go.id/subject/26/indeks-pembangunan-manusia.html#subjekViewTab1.
Boik, Andre, and Hidenori Takahashi. 2020. "Fighting Bundles: The Effects of Competition on Second-
Degree Price Discrimination." American Economic Journal: Microeconomics, 12 (1): 156-87.
Retrieved from DOI: 10.1257/mic.20180303.
Chaffey, Dave & Chadwick, Fiona Ellis (2016). Digital Marketing, Strategy, Implementation and
Practice, 6th edition. United Kingdom: Pearson Education Limited.
Dasi, A., Elter, F., Gooderham, P.&Pedersen, T (2017). New business models in the making in extant
MNCS: digital transformation in a telco. Emerald Insight, Breaking up the Global Value Chain:
Opportunities and Consequences Advances in International Management, Vol.30,pp.29-53.
Retrieved from https://doi.org/10.1108/S1571-502720170000030001.
Duenias, M (2015), The key to driving 4G profit: sell value, not bandwidth. Retrieved from
http://bswan.org/driving_4g_profit.asp
Giovanni, R. (2018). Analisis Pengaruh PDRB, Pengangguran dan Pendidikan Terhadap Tingkat
Kemiskinan di Pulau Jawa Tahun 2009-2016. Economics Development Analysis Journal, 7(1), 23-
31. Retrieved from https://doi.org/10.15294/edaj.v7i1.21922
Greenstein, S (2020). "The Basic Economics of Internet Infrastructure." Journal of Economic
Perspectives, 34 (2): 192-214. Retrieved from DOI: 10.1257/jep.34.2.192
https://www.telkomsel.com/en/node/152489, July 01,2020
https://www.telkomsel.com/en/node/152473, July 01,2020
https://www.telkomsel.com/en/node/149657, July 01,2020
https://www.telkomsel.com/en/diamond, July 01,2020
https://www.telkomsel.com/en/gold, July 01,2020
https://www.telkomsel.com/en/kartu-as/kartu-perdana-kartu-as, July 01,2020
https://www.telkomsel.com/en/simpati/paket-simpati?page=2, July 01,2020
https://www.telkomsel.com/en/platinum, July 01,2020
https://www.telkomsel.com/en/promotions, 25 Mei 2020
Indikator Kesejahteraan Rakyat, Welfare Indicators 2015 yang dikeluarkan Badan Pusat Statistik
Indonesia. Retrieved from
https://www.bps.go.id/publication/2015/11/10/9b1a211623201b8577993639/indikator-
kesejahteraan-rakyat-2015.html
Insani,R.,Soemitro,H.L(2016). Business Intelligence for Profiling of Telecommunication Customers.
Asia Pacific Journal of Contemporary Education and Communication Technology. Retrieved from
https://apiar.org.au/journal-paper/business-intelligence-for-profiling-of-telecommunication-
customers/.
Kingsnorth, Simon (2016). Digital Marketing Strategy, An Integrated Approach to Online Marketing.
Philadelphia: Koganpage.
Kotler, Philip & Keller, Kevin Lane (2016). Marketing Management Global Edition 15e. England:
Pearson Education Limited.
Kotler, Philip, Kartajaya, Hermawan & Setiawan, Iwan (2017). Marketing 4.0 Moving from Traditional
to Digital. Hoboken, New Jersey: John Wiley & Sons.
Nina,G.A., Rustariyuni,S.D.(2018). Pengaruh Gini Ratio, Pengeluaran Non Makanan dan Belanja
Modal Terhadap Tingkat Kesejahteraan Masyarakat di Provinsi Bali. Retrieved from
https://jurnal.uns.ac.id/jiep/article/view/23330.
Ovando,C.,et.al.(2015). LTE techno-economic assessment: The case of rural areas in Spain.
Telecommunications Policy, Vol.39, Issues :3-4, May2015,pp. 269-283.Retrieved from
http://dx.doi.org/10.1016/j.telpol.2014.11.004
Perreault, William, D. Jr., Cannon, Joseph, P. & McCarthy, E. Jerome, (2017). Essential of Marketing,
A Marketing Strategy Planning Approach 15e. New York: McGraw-Hill Education.
Rory (2019), Analisis Pengaruh Komponen Indeks Pembangunan Manusia Terhadap Tingkat
Kemiskinan di Indonesia. Retrieved from https://www.researchgate.net/publication/336988187
Sagala, F., Y., P., Ramantoko, G (2018). Strategies to Increase Purchase of Data Packages for
Telkomsel Kartuhalo Customers. Retrieved from
https://www.academia.edu/38281923/PROCEEDINGS_OF_THE_9TH_SCBTII.pdf.
Setiawan, W., B., Tricahyono, D.(2019), Optimization Strategy of Mobile Cellular Network Based on
Customer Smartphone Penetration. Retrieved from DOI: 10.5220/0008432904980505.
https://www.researchgate.net/publication/339081559_Optimization_Strategy_of_Mobile_Cellular
_Network_based_on_Customer_Smartphone_Penetration.
Tien, Nguyen Hoang & Anh, Dinh Ba Hung, (2017). Global Strategic Marketing Management.
Warsaw: Wydawnictwo EMENTON.
U. R. Raval and C. Jani, “Implementing & Improvisation of K-means Clustering Algorithm,” Int. J.
Comput. Sci. Mob. Comput., vol. 55, no. 5, pp. 191–203, 2016. Retrieved from
https://www.ijcsmc.com/docs/papers/May2016/V5I5201647.pdf
Widodo, A, W., Ramantoko, G (2018), Optimizing Market Segmentation for The Telecommunications
Industry: A Contextual Marketing Based Approach (Case Study at PT. Telkomsel). Retrieved from
http://www.ajmse.leena-luna.co.jp/ajmsevol8n1.php