Customer Review Analytics Solution for E-commerce

Client

The client is a large US E-commerce platform with rapidly growing online business. The company is constantly working on enhancing the customer experience by improving products and services offered. They needed to build an automated system which identifies sentiment in customers reviews about the products and services. They wanted to get insights into customer satisfaction to implement prompt marketing and customer care tactics towards poor customer satisfaction and offering upselling to the ones who are receptive to it. So, they asked the InData Labs team to develop a customer review analytics solution.

Challenge: detect sentiment and emotional tone in customer product reviews with ML

Sentiment analysis is important to uncover customer satisfaction and recognize their preferences through emails, comments, tweets, etc. To enable insight extraction, the InData Labs team developed an ML model based on customer reviews on products to detect opinions of customers.

To build an emotion intelligence model for product review analytics, we went through the following process of:

• Scraping customer review data from company website
• Pre-processing
• Feature Extraction
• Training and evaluation of Machine learning model

Solution: AI-powered sentiment analysis software

To enable sentiment classification, we made ML models and deep neural networks. We worked with:

• Transformers models like BERT.
• Logistic Regression, Support Vector machine (SVM) and Naïve Bayes – traditional ML models for document review analytics and sentiment classification.
• Deep neural networks used for sentiment analysis: Recurrent Neural Network, Long short term memory (LSTM) and Gated Recurrent Unit (GRU).

Take a closer look at the pipeline created:

To develop the solution, we used the following technologies:

Result: an automated system to identify sentiment in customers reviews

Automating customer satisfaction data collection, we enabled actionable information about consumer attitudes toward the client’s products and services at much higher response levels. Our sentiment analysis solution has helped the client determine if a particular customer segment feels more strongly about business or not. Besides, it’s made it easier for them to track how a change in product or service affects customers.

The key benefits of the product review analytics:

• Data collection and analysis of total customer reviews
• Tracking the company’s brand strength
• Tracking overall customer satisfaction
• Customer mood change detection
• Customer emotional trigger detection
• Customer churn prediction
• Live insights for prompt marketing tactics to support customer satisfaction strategy