Customer Sentiment

Insights from sentiment analysis increases customer satisfaction and retention for US based technology company

Business Outcomes

95%

increase in customer feedback analysis accuracy

100%

rise in customer satisfaction levels

100%

visibility of customer sentiment for quick insights

The company is a multidisciplinary technology outsourcing and engineering consulting provider with presence across USA, UK and Netherlands. They cater to varied sizes of customers ranging from SMB’s to Fortune 500 companies.

Business Need

In order to enhance their services and customer experience, the company actively sought existing client satisfaction levels of all customers spread across regions.

For this, they regularly conducted six-monthly surveys for structured and detailed feedback. They also received quick and colloquial feedback in meetings, chat messages and emails that was documented.

However, due to a lot of unstructured customer sentiment data lying in various siloes, the company found it difficult to manually decipher, identify and ascertain actual levels of customer satisfaction to make the required changes. They partnered with Hitech Analytics for automated analysis of these unstructured text data and get valuable insights on customer feedback.

Challenges

Tapping in various sources of unstructured feedback data stored in survey results or emails, chat messages.
Understanding the context of the feedback received across the sources and bucketing it based on categories like positive, negative and neutral
Managing an erratic volume of customer sentiment data
Scanning through hordes of records with generic feedback that made it difficult to judge satisfaction levels
Preparing an updated and structured text database of feedback to serve as a training database automated analysis using a ML- model

Want to derive userful information from large volumes of unstructured text data?

Get in Touch with us »

Solution

Hitech Analytics delivered a Machine Learning based model for automated analysis of textual data in database. The process included scanning emails and documents to determine the exactly expressed customer sentiment. This model was used to deconstruct the text, tag, parse and classify the data based on emotion and opinion to make sense of the feedback.

Along with this, an intuitive dashboard provides web-based notifications and email alerts every 12-hours to respective stakeholders about existing customer sentiment.

Approach

The systems of the respective stakeholders were synced with a plugin to collect feedback data from emails, survey forms, chat messages and save it in a centralized database.
Hitech’s text annotators and specialists got secured access to this centralized database to pre-process the data.
Business sensitive information like client name, company name, designation, location or revenue related details were not fetched by the plugin. Only the customer feedback data was displayed here.
Our team designed a specific template to structure the existing feedback data received from various sources.
Next, the existing data records were cleansed by removing the extra content and jargons. This cleansed data was used as a training baseline for ongoing classification and labelling of feedback content for the Machine Learning model.
The data in the database was run under the Machine Learning model to classify and label the feedbacks under positive, negative and neutral sentiment buckets. Using text annotation, our specialists were able to decipher the millions of feedback records.
Annotated thousands of feedback records under strict timelines while maintaining 95% accuracy.
Accurate, quick and data-driven insights on customer satisfaction have helped the company to drive better customer engagement.

Results

Enhanced feedback labeling accuracy
Quick identification of feedback sentiments
Boost response to customers that resulted in higher retention and satisfaction

Let's get started.

Talk to us to build fresh perspectives and leverage smart innovation for winning business outcomes.