Exploring Insights in the Automotive Sector through AI Analytical Capabilities

In the highly competitive and ever-evolving automotive sector, companies must evolve consistently to distinguish themselves. Constant updates are vital to anticipate changing consumer needs and adapt to new technologies and market trends.  

At Nethodology, we leverage new technologies and the integration of Artificial Intelligence (AI) with our experience in the automotive industry to develop multiple social listening projects. This approach allows us to gain valuable insights for the strategy of companies in the sector, enabling them to make relevant and accurate strategic decisions based on data and cutting-edge technologies.

 Partnership with SEAT/CUPRA

A tangible example of this evolution is our recent collaboration with SEAT/CUPRA. This Spanish car brand, present in the market since 1950, is a leader in technological innovation, unique design, sports performance, and commitment to sustainability. Our recent collaboration is a clear demonstration of this evolving landscape, where data-driven approaches are combined with Artificial Intelligence for the development of innovative projects.  

Here, our primary goal was to analyze thousands of consumer reviews of 13 car brands in five principal European markets. We collected 420,000 Google reviews from 900 dealerships, carefully selected to ensure a representative geographic diversity of the five markets of interest for SEAT: Spain, Italy, France, Germany, and the United Kingdom.
The goal was to gather relevant information regarding consumer opinions of each brand compared to its competitors.

Implementation of AI

Artificial Intelligence (AI) played a crucial role in the development of the project, allowing us to identify and categorize thousands of data points. Initially, we classified opinions into two main categories using machine learning methods: retail/sales experience vs after-sales/maintenance experience. Then, we applied statistical methods to obtain a Net Sentiment Score (NSS) for each brand, evaluating customer opinions by assigning numerical values based on their emotional tone. This process provided us with relevant information on sales-related ratings, an aspect not shown on Google My Business, which only presents total scores. This way, we could differentiate between sales and after-sales ratings for each brand.

 Advanced Classification Techniques

Through advanced text classification techniques, we identified over 30 recurring categories of purchase experiences, such as accessibility, comfort, courtesy offers, customer service, vehicle availability, location, vehicle condition, online purchasing experience, and discounts, among others. Notably, multiple opinions about different categories were detected within each review. To discern the positive and negative aspects of the consumer retail experience, we employed micro-sentiment analysis.  

Our team with expertise in data science took care to carefully review and ensure the reliability of each algorithmic step, to prevent the model’s “hallucinations.” This term refers to a common problem with Large Language Models, as they might return plausible but incorrect answers. Leveraging GPT-4’s capabilities, and our experience in the automotive sector, we obtained several relevant insights:

Net Sentiment Score (NSS) by Brand – (NSS) is calculated as: [% of 5-star reviews]  –  [% of 1-star reviews]**

 Detailed and Comprehensive Results

What was the outcome? Thanks to our process and methodological development, we achieved a comprehensive view of the retail experience for each brand, offering detailed insights. Our analysis, supported by categorized data and real-world examples from Google My Business ratings, allows our recommendations on strengths and weaknesses to be more precise and insightful. This has supported informed decision-making related to SEAT/CUPRA’s retail strategy. 

Among the most relevant findings, we discovered: 

– The best and worst practices in the purchasing process, ranked by customer importance.
– A comparison of 13 brands across 5 countries in 34 relevant categories like “Testdrive” or “Model Display”.
– Detailed case studies, such as some brands performing poorly in certain markets but excelling in others. For instance, Volkswagen has a low NSS in Italy but is above average in Spain and Germany.  

All findings are supported by qualitative analysis and bolstered by real examples from Google Maps reviews.

In our pursuit to enhance AI-driven analysis, we adopted a pragmatic approach by distinguishing between sales and after-sales analyses and automating consumer comment classification. As we move forward, our commitment to innovation remains steadfast, focusing on delving into user emotions, refining sentiment analysis, and expanding AI use in future Social Listening studies.

Benchmark Germany

Nethodology and Our Commitment to AI

At Nethodology, we demonstrate our commitment to excellence in data analysis for the automotive industry. We combine new technologies and industry expertise to derive valuable insights that support companies’ market research findings. Our collaboration with SEAT-CUPRA is a prime example: our rigorous methodology and innovative use of artificial intelligence have led to significant outcomes.  

We collected and analyzed reviews, classified opinions in advanced ways, and leveraged the data to obtain detailed insights that go beyond traditional metrics. At Nethodology, we are committed to further improving AI-driven analysis and to staying at the forefront of innovation in the field of social listening.