One51 Success Stories
Understanding customer attitudes towards Brand, Products and Services with sentiment analysis for a fast-food industry giant
Overview
Market and customer understanding is essential and something that should be broken down and analysed. Business intelligence is a significant insight method for examining market interest and fulfilment. Since business intelligence requires intense analysis, sentiment analysis is an incredible method for analysing customer satisfaction and delivering business insights.
Sentiment analysis is a popular method for reviewing and analysing text reviews about products and services. Feature-based sentiment analysis is used to reveal customers’ sentiments not just at the overall product/brand level, but also at the product/service specific feature level.
Our client, a leading global fast-food brand, wanted to analyse and understand the customer reviews about their products and services in order to identify any issues and improve the overall customer experience.
Our team designed an end-to-end solution to deliver key insights from customer reviews using Microsoft Azure Cognitive Services Text Analytics. We then created a dashboard to present the findings.
Data and analysis
We used customer review comments from various sources such as the company app and website, as well as customer surveys and delivery partners.
Some of the requirements our client wanted included:
- Overall stats for the review sentiments
- Sentiment classification of customer reviews by time, store, channel, and category
- To identify the product names and services that have maximum positive or negative reviews
- An interactive dashboard for the end users
To analyse customer comments, we used the Azure text analytics API to classify each review into four categories — positive, negative, neutral, and mixed. Each review was also assigned a score for positive, negative, and neutral, adding up to 1 or 100%. The analysis models used by the API were retrained using an extensive body of text and natural language technologies from Microsoft. Moreover, we also used the Azure entity recognition API to identify keywords and their corresponding categories from the text. As a result, the reviews were classified into 13 distinct categories such as Product, Skill, Organisation, Person type and so on.
Results
After an initial discovery process, One51 set about creating a solution that would meet our clients’ needs.
The business requirements asked for a data platform that would:
The reviews classification by sentiments charts enabled users to compare
the sentiment across all channels.
The reviews by category visuals helped users to understand the categories
with the most positive or negative reviews and improve customer
experience based on the results.
The word cloud identified the most common issues raised by customers.
From a business perspective, negative reviews can be more valuable than positive or neutral reviews because stakeholders can evaluate these reviews to detect common concerns by customers and then fix the issue. Our solution helps to identify these opportunities for improvement.
The final product for this project was the implementation of an intuitive dashboard for the end users that allowed them to access the data and make key business decisions based on the results. Given the huge volume of data, the dashboard was cleverly organised in a user-friendly way. Our team collaborated with various stakeholders to provide feedback on the visualisations and key metrics to make sure they wouldn’t be overwhelmed and could interpret the results quickly.
