n8n templatesn8n templates

Trustpilot Customer Review Intelligent Analysis Workflow

This workflow aims to automate the scraping of customer reviews for specified companies on Trustpilot, utilizing a vector database for efficient management and analysis. It employs the K-means clustering algorithm to identify review themes and applies a large language model for in-depth summarization. The final analysis results are exported to Google Sheets for easy sharing and decision-making within the team. This process significantly enhances the efficiency of customer review data processing, helping businesses quickly identify key themes and sentiment trends that matter to customers, thereby optimizing customer experience and product strategies.

Workflow Diagram
Trustpilot Customer Review Intelligent Analysis Workflow Workflow diagram

Workflow Name

Trustpilot Customer Review Intelligent Analysis Workflow

Key Features and Highlights

This workflow automatically scrapes customer reviews of specified companies from Trustpilot, utilizes the Qdrant vector database to store and manage review data, applies the K-means clustering algorithm to automatically identify thematic clusters of reviews, and then leverages OpenAI’s large language model (LLM) to perform in-depth analysis and summarization of each cluster. Finally, the insights are exported to Google Sheets for easy team sharing and decision support.
Highlights include:

  • Automated multi-page review data scraping and structured extraction
  • Efficient similarity search and clustering analysis using a vector database
  • Intelligent insight generation combining machine learning clustering with large language models
  • Flexible time-range filtering supporting periodic or on-demand analysis
  • Direct output to Google Sheets for convenient subsequent data processing and presentation

Core Problems Addressed

  • Manual collection and analysis of large volumes of customer reviews is time-consuming and inefficient
  • Review data is unstructured, making it difficult to summarize genuine customer feedback and pain points
  • Inability to quickly identify trending topics and sentiment trends among customers
  • Lack of automated tools to transform massive review data into actionable insights

Application Scenarios

  • Brand and product management teams regularly monitoring customer satisfaction and feedback trends
  • Market research departments analyzing competitors or industry reputation
  • Customer service teams quickly identifying key complaints and improvement opportunities
  • Data analysts building business reports driven by customer review data
  • Any enterprise or individual seeking to extract structured insights from Trustpilot reviews

Main Process Steps

  1. Initialization: Configure target company ID and clear historical review data in Qdrant to ensure data freshness.
  2. Review Data Scraping: Automatically scrape the latest three pages of review HTML pages from Trustpilot via HTTP requests.
  3. Review Content Extraction: Parse HTML nodes to extract fields such as reviewer, rating, title, body, date, and country.
  4. Data Structuring: Organize extracted review information into array format for subsequent processing.
  5. Text Vectorization: Use OpenAI Embeddings API to convert review texts into vector representations.
  6. Vector Data Storage: Insert vectors and their metadata into the Qdrant vector database to support efficient similarity search.
  7. Trigger Analysis Sub-Workflow: Filter review vectors by time range and execute clustering and analysis processes.
  8. Clustering Analysis: Run K-means algorithm via Python code node to automatically group review vectors into multiple thematic clusters.
  9. Retrieve Cluster Review Details: Filter valid clusters (containing 3 or more reviews) and fetch corresponding review content from Qdrant.
  10. Intelligent Insight Generation: Use LLM to summarize each cluster’s reviews, generating insights, sentiment orientation, and improvement suggestions.
  11. Export Results: Write insights and original review counts into a Google Sheets document for easy viewing and sharing.

Involved Systems or Services

  • Trustpilot: Source of customer review data.
  • Qdrant: High-performance vector database for storing and querying review vectors.
  • OpenAI API: Generates text embeddings and intelligent text summaries (GPT-4 model).
  • Google Sheets: Stores final analysis results, supporting team collaboration and data sharing.
  • n8n: Low-code automation platform used to build the entire data scraping, processing, and analysis workflow.

Target Users and Value

  • Brand managers and marketers, aiding in understanding customer reputation and product feedback.
  • Data analysts and business intelligence teams, enhancing the efficiency and depth of customer review analysis.
  • Customer service leaders, quickly pinpointing major complaint areas and improvement directions.
  • Any enterprise or individual user seeking to automate extraction of valuable insights from Trustpilot reviews.

This workflow significantly lowers the barrier to processing customer review data by transforming massive unstructured reviews into structured, thematic, and insight-rich information, empowering businesses to improve customer experience and product competitiveness.