In today’s competitive market, understanding customer diversity is essential for tailoring marketing strategies and improving sales. ThisstudyfocusesoncustomersegmentationatTVS Sai Hemanth Motors, Srikalahasthi, using K-Means clustering, a machine learning technique.The aim is to classify customers based on demographic and behavioralfactorsandidentifymeaningfulpatternsin purchasing and non-purchasing behavior. Data was collected from both buyers and non-buyers, and the ElbowMethod wasappliedtodeterminetheoptimal numberofsegments. The analysisresultedinclearly defined clusters that revealed specific customer characteristics and preferences. This enables the dealer to make informed decisions regarding marketing strategies and customer relationship management. The study not only enhances understanding of customer profiles but also demonstratesthepracticalapplicationofdata-driven segmentation in a real-world context
Introduction
The study focuses on customer segmentation using K-Means clustering to improve marketing strategies in the competitive two-wheeler market, specifically at TVS Sai Hemanth Motors. Customer segmentation involves grouping customers based on shared characteristics such as demographics, behavior, and preferences to enable targeted marketing and improve customer acquisition and retention.
The literature review highlights previous research applying machine learning techniques, especially clustering, in the automotive sector to enhance customer targeting and sales strategies. The study aims to classify customers by demographics (age, income, profession), analyze purchasing behavior, and apply machine learning for automated segmentation.
Data were collected from 100 customers (50 buyers and 50 non-buyers) using surveys and dealership records. Python and visualization tools were used to perform clustering and determine the optimal number of customer segments.
Three distinct buyer clusters emerged:
Cluster 0: Young, price-conscious buyers (18-25 years, low income), focused on mileage and EMI options.
Cluster 1: Mid-income balanced buyers (26-35 years), with moderate interest in price, style, and speed.
Cluster 2: Affluent, feature-seeking buyers (36-45+ years), focused on premium features and style with low EMI dependence.
Non-buyer clusters showed similar but varied engagement levels, indicating different marketing needs. The study emphasizes the importance of understanding diverse customer segments for better resource allocation and targeted marketing in the two-wheeler industry.
Limitations include a small sample size, focus on one dealership, and reliance on demographic data without deeper behavioral insights.
Conclusion
K-Means clustering algorithm is successfully appliedto segment bothbuyers and non-buyers based on demographic and behavioral variables. The analysis revealed clear patterns in customer preferences and highlighted key differencesbetween similar customer groups who did and did not make a purchase. These insights help the dealers understand their customers better and craft targeted strategies for each segment. By bridging the gap between buyers and non-buyers through focused offers and engagement, the dealership can increase conversions and optimize its marketing efforts. Overall, this study demonstrates the practical value of machine learning in solving real- world business problems
References
[1] BagirathiIyer,(June2011),AStudyonReasons for Low Acceptability of Battery Operated 2 Wheelers (E-Bykes) in Nagpur Region with Special ReferencetoElectrothermIndiaLimited’s Yobykes, International Journal of Research in Finance & Marketing, Volume1, Issue2.
[2] PhilipKotler,(2003),MarketingManagement, 11thEdition, Pearson Eduction, Delhi, India, pp:278- 299.
[3] NareshK.MalhotraandSatyabhushanDash,(2007), MarketingResearch-AnAppliedOrientation, Fifth Edition, Pearson Education, New Delhi, India, pp. 636-661.
[4] M.N. Tuma, S. Scholz, and R. Decker, (2009), The Application of Cluster Analysis in Marketing Research: A Literature Analysis, Business Quest Journal, vol. 14.
[5] M MKodabagi, Savita S Hanji, Ravatappa A.B,(2013) “This is my paper”, A Novel Clustering Technique Using Similarity Measure, International Journal of Latest Trends in Engineering and Technology (IJLTET), Special Issue IDEAS-2013,ISSN: 2278-621X.
[6] H. N. Ramesh, Vinod S., Sanjay Hanji, (Jan-June 2012) “This is my paper”, Hybrid Segmentation AtTwo Wheeler Market In India, IEMS Jounal of Management Research, ISSN : 2249-569X,Volume1, Issue-1.
[7] RakeshKumarArora,Dr.DharmendraBadal,(April- June 2013), Evaluating Student’s Performance Using k-Means Clustering, International Journal of Computer Science AndTechnology(IJCST), ISSN(ONLINE):0976-8491ISSN(PRINT):2229-4333Vol.4,Issue-2.
[8] Hedayetul Islam Shovon, Mahfuza Haque, (2012), An Approach of Improving Student’s Academic Performance by using K-means clusteringalgorithm and Decision tree, International Journalof Advanced Computer Science and Applications(IJACSA), Vol. 3, No. 8.
[9] Chady El Moucary, Marie Khair, Walid Zakhem, (2011), Improving Student’s Performance Using Data Clustering and Neural Networks in Foreign- Language Based Higher Education, The Research Bulletin of Jordan ACM, Vol. I ( III ).
[10] Oyelade, O. J, Oladipupo, O. O, Obagbuwa, I. C, (2010), Application of k-Means Clusteringalgorithm for prediction of Students’ Academic Performance, International Journal of Computer Science and Information Security (IJCSIS), Vol. 7.
[11] Md. Hedayetul Islam Shovon,MahfuzaHaque,(July 2012), Prediction of Student Academic Performance by an Application of K-Means Clustering Algorithm, International Journal of Advanced Research in Computer Science and SoftwareEngineering,ISSN:2277128XVol2, Issue 7.
[12] Mahmud M.S, Rahman M.M, Akhtar M.N, (2012), ImprovementofK-meansclusteringalgorithm with better initial centroids based on weighted average, Electrical & Computer Engineering(ICECE), 7th International Conference, Pages:647- 650.