Customer Segmentation and Targeting by Data Science Methods
Moon, Inwook (2020)
Moon, Inwook
2020
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2020090820248
https://urn.fi/URN:NBN:fi:amk-2020090820248
Tiivistelmä
The objective of this thesis is performing a segmentation analysis as well as classifying target segment members with a given survey data. With the performance of this customer survey data analysis, the purpose of this research is to confirm the usefulness of data science methods in marketing and sales. By this study, it is expected to further understand data-driven approach in business. The development task was firstly grouping customers in different segments and secondly spotting the members of the target group, respectively with the use of available data science methods.
The theoretical background used for the customer survey data analysis includes marketing theories about segmentation and targeting. The author’s own experience in marketing and B2B, B2C sales tasks were helpful as well. For the data science methods, different algorithms have been studied since the time before starting this research. The research methods used in this study are quantitative analytics methods of data science. For the task of segmentation problem, unsupervised learning methods of clustering were utilized with tools such as K-means algorithm. For the task of predicting target members, supervised learning algorithms were used including an ensemble method and a neural network.
By performing of the analysis, research problems were solved. Firstly, the basic statistics and distributions of the survey results were summarized and visually presented. Secondly, by the use of relevant segmentation criteria, individuals were grouped into 5 distinctive segments and the target segment was chosen according to the predefined standards. The observed characteristics of the segment members were described. Then, the evaluation about the segmentation effect was performed by the comparison of the profits between the ‘data science use’ case and the opposite. Thirdly, the individuals that belong to a target group were identified from the survey data. The data was split into training and test data and the accuracy of the latter was measured by metrics.
In conclusion, it was confirmed that data science methods are applicable on business problems in segmentation and targeting. By using data science methods, a company is more likely to achieve the higher profits with the same budget given for the marketing and sales. This research tried to demonstrate the findings by both the volume of impact on business and the technical measurement. Based on this study, the author would like to recommend the small and medium size companies to consider applying data science methods on their business problems.
The theoretical background used for the customer survey data analysis includes marketing theories about segmentation and targeting. The author’s own experience in marketing and B2B, B2C sales tasks were helpful as well. For the data science methods, different algorithms have been studied since the time before starting this research. The research methods used in this study are quantitative analytics methods of data science. For the task of segmentation problem, unsupervised learning methods of clustering were utilized with tools such as K-means algorithm. For the task of predicting target members, supervised learning algorithms were used including an ensemble method and a neural network.
By performing of the analysis, research problems were solved. Firstly, the basic statistics and distributions of the survey results were summarized and visually presented. Secondly, by the use of relevant segmentation criteria, individuals were grouped into 5 distinctive segments and the target segment was chosen according to the predefined standards. The observed characteristics of the segment members were described. Then, the evaluation about the segmentation effect was performed by the comparison of the profits between the ‘data science use’ case and the opposite. Thirdly, the individuals that belong to a target group were identified from the survey data. The data was split into training and test data and the accuracy of the latter was measured by metrics.
In conclusion, it was confirmed that data science methods are applicable on business problems in segmentation and targeting. By using data science methods, a company is more likely to achieve the higher profits with the same budget given for the marketing and sales. This research tried to demonstrate the findings by both the volume of impact on business and the technical measurement. Based on this study, the author would like to recommend the small and medium size companies to consider applying data science methods on their business problems.