A leading manufacturer of Automobile industry headquartered in Europe wanted to launch a new model in the market. They are a global leader in designing, manufacturing, and selling of modern motor vehicles.
The last model they had launched was not doing well in the market and a lot of unhappy customers had started to pour in their complaints on the company’s Twitter handle. Their in-house team faced huge difficulties in extracting and classifying those negative comments. The manual processing caused a lot of errors which increased the time taken to resolve their problem.
The company approached Xtract.io to aggregate and analyze massive amounts of Tweets from various geographies. They were hoping to improve their customer service and launch a new model of their car with a better understanding of the market.
The tweets were in many languages like Chinese, Japanese, French, Spanish and more. The target market was spread across various regions around the world. The main challenge was to analyze each tweet and classify them based on the sentiments (positive, negative, neutral). Around 1 million tweets had to be aggregated spanning over a period of one year. They wanted to identify the root cause of the negative reception. This meant the sentiment behind each tweet had to be analyzed for specific features like engine, design, style, and door from all those aggregated tweets.
Xtract.io approached the whole need with an aspect-based sentiment analysis approach. We first translated all the non-English tweets using our proprietary translation bots.
Tags were created for client desired features like design, engine, style, and door. Our data analysts identified more features that affected the customer buying preference like price, ease of use, and headlamp. Sentiment analysis was performed on each tweet and polarity rating for each of these identified features was tagged.
With our sentiment analytics platform, our client was able to get the bigger picture displaying all the key metrics with their aspects and polarity. In a single window, they were able to filter the reports based on various categories like price, ease of use, headlamp, and more. It helped them identify what caused a negative impact in the past and thereby reducing the resolution time.
Our client was able to identify the emotions behind the purchase of their product and reduced 60% cost in R&D efforts. The time spent on research was reduced to weeks from months. They were able to plan their product improvements before the manufacturing phase with the new insights that helped in reducing the cost and time drastically. The new model released in the market aligned with customer expectations and increased the profit margin.
Enhanced customer satisfaction
Accelerated product planning
1 million multi-lingual tweets analyzed
Emotions behind each feature identified