Sentiment Analysis of Online Customer Reviews to Optimize Marketing Strategies in the E-Commerce Industry

Abstract

The present study observes the ideas, feelings, and shopping experiences that shapes consumers decision through online reviews within the fast-expanding e-commerce space. These days, evaluations have a significant impact on consumer decisions, brand perception, and customer trust. The authors wanted to look beyond surface-level star ratings and try to understand the mood, the frustrations, and the small wins that shape how consumers feel about brands. Two major e commerce players were studied for this analysis viz, Flipkart, Meesho representing affordable items. Both of this player operates in a slightly different niche, yet both share the same challenge: earning trust in a market where loyalty shifts quickly. For the practical side of this study, author pulled product reviews and star ratings from a publicly available Kaggle dataset. The reviews which were focused on covered electronics, fashion, lifestyle products, and cosmetics. These categories were analysed because written feedback tends to influence how buyers make decisions in these segments- a single negative review about a phone\'s battery life or a foundation shade mismatch can genuinely push someone toward a competitor. Authors used Python for the sentiment analysis of reviews collected from Kaggle for the analysis. For data cleaning and manipulation, Pandas library was used to filter rows, handle missing values and reshape the columns. A lot of the early preprocessing work dropping duplicates, normalizing the review text, and cutting columns that weren\'t going to matter for the analysis happened directly in Pandas. TextBlob was used to generate polarity scores and then split each review into positive, neutral, or negative based on where the score landed. Matplotlib and Seaborn were used for analysis of sentiment percentages across five platforms. This study reinforces that sentiment analysis is genuinely useful online shoppers and retailers- who want to stay close to what their customers actually feel. Going through customer evaluations helps a business see what is working, where the product or service is falling short, and how marketing messages can be shaped to speak to real concerns rather than assumptions.

Introduction

The text examines sentiment analysis of customer reviews in Indian e-commerce, particularly focusing on platforms such as Flipkart and Meesho. It explores how Natural Language Processing (NLP) and machine learning can help understand customer opinions and improve marketing strategies.

• The e-commerce industry has expanded rapidly due to increased internet and smartphone use, technological development, digital payments, improved logistics, and quick-commerce services. Major business models include B2C, B2B, C2C, and D2C.
• Previous research has applied machine learning and deep learning methods such as Naïve Bayes, SVM, Random Forest, CNN, LSTM, XGBoost, BERT, and Transformer models to analyse customer sentiment and predict purchasing behaviour.
• Research shows that customer sentiment, ratings, trust, discounts, and review transparency can influence purchase intentions and consumer behaviour.
• More advanced models, particularly BERT and Transformer-based approaches, have achieved high sentiment-classification accuracy, often above 90%. However, many existing studies focus mainly on classification accuracy rather than connecting sentiment results to actual business outcomes.
• Important research gaps include limited real-time deployment, multilingual capability, sarcasm detection, cross-platform validation, revenue prediction, customer lifetime value analysis, and integration of sentiment with marketing performance.
• The study therefore aims to:
  1. Analyse online customer reviews using sentiment-analysis techniques.
  2. Determine how customer sentiment insights can be used to improve marketing strategies, customer targeting, and service quality.
• The research uses a descriptive and analytical design, combining qualitative interpretation of customer reviews with quantitative analysis of sentiment patterns and relationships with marketing variables.
• Secondary data were collected from Kaggle, with reviews covering products, services, delivery, and overall customer experiences. Convenience sampling was used because of the accessibility of online review data.
• The analysis uses Python, Pandas, TextBlob, and Matplotlib. Reviews are processed using basic NLP techniques, and TextBlob is used to generate sentiment polarity scores.
• For the Flipkart dataset, the results show approximately:
  • 13,500 neutral reviews, representing the largest category.
  • 10,000 positive reviews, representing a substantial proportion of customer feedback.
  • 1,000 negative reviews, representing the smallest group.
• The Flipkart results indicate that neutral sentiment dominates, but positive reviews substantially exceed negative reviews. This suggests generally favourable customer perceptions and relatively satisfactory service experiences.
• The large number of neutral reviews also represents an opportunity for businesses to investigate what drives neutral experiences and potentially convert them into positive customer experiences.

Conclusion

The study found that sentiment analysis is an effective tool for understanding customer opinions and identifying factors that influence customer satisfaction, including product quality, pricing, delivery services, customer support, and overall shopping experience. Positive customer sentiment is mainly driven by timely delivery, affordable pricing, a wide variety of products, secure payment options, and efficient customer service. These factors help improve customer satisfaction, loyalty, and repeat purchases. Negative customer sentiment is generally associated with delayed deliveries, poor product quality, refund issues, misleading product descriptions, and inadequate customer support. These issues can reduce customer trust and negatively affect business performance. By analysing online customer reviews, businesses can identify customer expectations, evaluate service quality, improve product offerings, and make better marketing decisions based on customer feedback.

References

[1] Sharma, R., & Verma, P. (2023). Sentiment analysis of Flipkart product reviews using machine learning. International Journal of Advanced Computer Science and Applications, 14(2). [2] Iyer, K., & Reddy, M. (2021). Opinion mining on Amazon India reviews using hybrid model. [3] Patel, D., & Singh, V. (2023). Social media sentiment analysis for Indian brands. IEEE Xplore Conference Proceedings. [4] Kulkarni, S., & Mehta, R. (2020). Product recommendation system for Indian online retail using collaborative filtering. Springer LNCS, 121. [5] Rao, P., & Thomas, J. (2024). Deep learning-based sentiment analysis for Myntra reviews. Journal of Big Data Analytics, 8(1). [6] Chatterjee, S., & Das, A. (2022). Customer behavior modeling in Indian online grocery platforms. Data in Brief, 39. [7] Menon, L., & Krishnan, R. (2021). Sentiment classification of Indian e-commerce reviews using CNN. IJCA, 44. [8] Joshi, N., & Bansal, T. (2023). AI-driven marketing intelligence in Indian e-commerce. International Journal of Marketing Analytics, 11.

Copyright

Copyright © 2026 Mohammed Arvaz, Dr. Richa Das . This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.