ABSTRACT
Market segmentation approaches applied by small businesses in Kenya have mostly been based on very limited customer factors. This includes geographic, demographic, behavioral, and psychographic characteristics of the customer. These approaches have not entirely brought out the nature of the customers in the business. In some cases, the approaches have been based on incorrect assumptions and have also led to challenges like potentially ignoring new markets and difficulty in keeping up with changing customer needs. This research seeks to use a clustering algorithm to carry out market segmentation based on more variated and integrated data that could give more information on customer habits. The dataset that was used in this research contained integrated data on various customer and business facets. The research used the spectral co-clustering algorithm to bring out various traits on the business customers that could then be used to segment the customers into more effective markets. This research eventually brought out different market segments with varying characteristics all based on the best features in the integrated business data.
This study showed that there is more data available in a business that can be used for market segmentation other than just geographical, behavioral, demographic and psychographic data. It also brought out the importance of feature selection in a dataset since different features may have different effects on sales on a business and the overall performance of a market segment.
This study also contributes to research by identifying other features other than geographical, demographic, psychographic and behavioral factors that could be used to identify market segments in a small retail business. This approach is also more effective since the use of a clustering algorithm enables the discovery of patterns in business data that would not have been easy to spot with the naked eye, and the use of up-to-date business data aids the business in keeping up with customer habit changes and prevents missing out on potential markets.
TABLE OF CONTENTS
DECLARATION 2
DEDICATION 3
ACKNOWLEDGEMENT 4
Definition Of Important Terms 5
TABLE OF CONTENTS 6
LIST OF FIGURES 8
ABSTRACT 9
CHAPTER 1: INTRODUCTION
1.1 Background 1
1.2 Problem Statement 2
1.3 Objectives 3
Overall Objective 3
Research Questions 3
Research Objectives 3
1.4 Justification 3
1.5 Significance 4
1.6 Assumptions And Limitations 4
CHAPTER 2: LITERATURE REVIEW
2.1. Introduction 4
2.2. Existing Work 6
2.2.1. Tweet-Based Target Market Classification Using Ensemble Method 6
2.2.2. Use of K-Means on Business Data for Efficient Customer Segmentation: A Strategy for Targeted Customer Services 6
2.2.3. Customer Segmentation and Profiling using Demographic Data Obtained from Daily Mobile Conversations 7
2.3. Clustering 8
2.4. Gap 10
2.5. Linking Existing Work to My Proposed Approach 10
2.6. Process Model 11
CHAPTER 3: METHODOLODY
3.1. Introduction and the General Methodological Approach 12
3.2. Research Design 12
3.2.1. Data Collection and Analysis 12
3.2.2. Feature Selection 13
3.2.3. Algorithm Selection 17
3.2.4. Model Development and Validation 18
CHAPTER 4: RESULTS AND DISCUSSION
4.1. Data Collection and Analysis 19
4.2. Feature Selection 20
4.3. Customer Segmentation Results 23
4.4. Discussion 27
CHAPTER 5: CONCLUSION
5.1. Achievements 27
5.2. Contributions 29
5.3. Challenges 29
References 30
Appendix
Sample Source Code
LIST OF FIGURES
Figure 1: Process Model 11
Figure 2: Model development process design 18
Figure 3:Conceptual Framework 23
CHAPTER 1
INTRODUCTION
1.1 Background
The performance of 68.1% of small businesses are affected to a large extent by the marketing strategies they adopt (Muola, 2017). By selecting the right marketing strategy for its products, a company can manage to correctly utilize its resources and increase its profitability. (Muhammad Adi Khairul Anshary*, 2016) Companies in Kenya have recognized that they cannot appeal to all consumers in the same way since consumers are widely scattered in their needs and practices. (Njoroge, 2015) This has brought about the need to segment their markets based on various characteristics to enable them to reach their customers more effectively.
The market segmentation approaches used have mostly focused on the geographic, demographic, behavioral, and psychographic characteristics of the customer. These approaches have been fixed to only customer data on geographical location or demographic information (gender, age, etc) or lifestyle and personality characteristics. The consideration of one facet or feature of customer data to create market segments for a business is not enough. This is because the other features and factors held in business and customer data hold more insights which could create more elaborate and effective customer segments when they are all considered during customer segmentation. This research aims at carrying out a multi-faceted approach to market segmentation where we shall use various data sources and data features together to come up with customer segments for small retail businesses.
Chloride Exide company carried out market segmentation based on customer behavior data like user status, user rates, benefits, and geographical location. These factors, though important, were not enough for the company to create concise market segments bringing about the challenge of ignoring potential market segments and not managing to keep up with the dynamic changes to their customers’ needs. The company however highlighted the following aspects to be important in influencing the customers’ purchase decisions: education, gender, income levels, and use of additional services like warranties and discounts. These factors if included during the market segmentation process could lead to a better view of the market clusters for the company. (Yabs, 2014)
Kenya Commercial Bank in Mombasa carried out market segmentation through the geographical, demographic and behavioral segmentation approaches. In the end, they identified geographical segmentation as a very effective approach. Demographic and behavioral segmentation was also quite significant but psychographic segmentation was not effective for the organization. (Wayabila, 2020). Considering only one of the segmentation approaches was highly effective, it was important for the organization to consider alternative approaches to segment their customer markets. The use of more data for market segmentation was highlighted as channel that would enable the bank to carry out its customer segmentation goals in a more effective manner.
The Kenya Commercial Bank in Kisii county mainly used geographical market segmentation for products that were needed by a large and diverse group of customers. This was with the assumption that the entire geographical region comprised of customers who had same behaviors towards the products, an aspect that was not entirely the case since customers in the same geographical region also had different product preferences. (Nyabwari, 2017)
It was then important for the bank to incorporate more variated data in their customer segmentation process to eliminate these assumptions and in the process come up with data driven market segments that were more effective for the organization.
Customer segmentation through clustering is the “use of a mathematical model to discover groups of similar customers based on finding the smallest variations in data”. (Optimove, 2020) Clustering enables the use of a large variety of data on a given market since it has the capability to learn from diverse data sources like information from loyalty cards, point of sale systems and existing datasets to better categorize the different customers into more homogenous segments that can be easily targeted during marketing. It also addresses the challenge of needing to keep up with the changing customer needs in a company since the algorithm constantly learns from the data fed into the business’ system, thus giving up to date market segments.
1.2 Problem Statement
Organizations in Kenya have been limiting their consideration to only a few factors during market segmentation. These have been either geographical or behavioral factors and, in some cases, demographic data has also been used. (Wayabila, 2020) These approaches have mostly been based on incorrect assumptions and have not been completely effective. (Nyabwari, 2017) The use of this approaches has also led to challenges like potentially ignoring new markets and difficulty in keeping up with changing customer needs.
There is thus the need to come up with a market segmentation approach that is more effective by using more variated and dynamic data factors that could give more information on customer habits.
1.3 Objectives
Overall Objective
The aim of this project was to create a machine learning model that would use a variety of sales and customer data to create effective market segments.
Research Questions
The following research questions were used to attain the overall objective of the project.
1. What are the sources of data that hold information on customer data and business sales?
2. What are the features that influence sales in a given business?
3. What is the best algorithm to use to cluster customers into effective market segments?
Research Objectives
The above research questions then advised on the following research objectives
1. Identify the sources of data that hold information on customer data and business sales
2. Identify the features that most affect sales in a given business
3. Identify an optimal algorithm that can be used to cluster customers into effective market segments
4. Develop and validate a machine learning model that will use these different sources of data to design market segments for a given business.
1.4 Justification
The success or failure of a business has been greatly attributed to the marketing strategy that the business employed (Muthee Janet, 2014). The market segmentation approaches currently used by organizations have not been completely effective due to the limited nature of data used. It has brought about challenges in bringing out the actual structure of the market thus raising the risk of ignoring potential market segments that would have been captured with the use of more variated data. It is therefore important to use a wide variety of data factors to create customer segments.
1.5 Significance
Small businesses in Kenya will benefit from this research since it will provide a more effective approach to come up with market segments. Through the use of integrated data, the small businesses will manage to consider a wider array of factors regarding their customers and the business during this process. This will help to alleviate any assumptions made during market segmentation processes.
The use of a machine learning algorithm on the integrated business data will enable the business to identify patterns in their data that could not have been possible with only one source of data. It will also help the business in keeping up with the dynamic changes in customer needs through the different results given by the algorithm at different circumstance. This will lead to the creation of more effective market segments in small businesses.
1.6 Assumptions And Limitations
a) This research will focus on small retail business in Kenya
b) There may be other factors or conditions that may affect the marketing of a given product that may not be represented by the data collected during the project.
Click “DOWNLOAD NOW” below to get the complete Projects
FOR QUICK HELP CHAT WITH US NOW!
+(234) 0814 780 1594
Buyers has the right to create
dispute within seven (7) days of purchase for 100% refund request when
you experience issue with the file received.
Dispute can only be created when
you receive a corrupt file, a wrong file or irregularities in the table of
contents and content of the file you received.
ProjectShelve.com shall either
provide the appropriate file within 48hrs or
send refund excluding your bank transaction charges. Term and
Conditions are applied.
Buyers are expected to confirm
that the material you are paying for is available on our website
ProjectShelve.com and you have selected the right material, you have also gone
through the preliminary pages and it interests you before payment. DO NOT MAKE
BANK PAYMENT IF YOUR TOPIC IS NOT ON THE WEBSITE.
In case of payment for a
material not available on ProjectShelve.com, the management of
ProjectShelve.com has the right to keep your money until you send a topic that
is available on our website within 48 hours.
You cannot change topic after
receiving material of the topic you ordered and paid for.
Login To Comment