B2C Brand Perception Analysis

Framework to measure a B2C brand's perception based on customer feedback and public complaints.

Short on time? Skip the scrolling and hit play- I'll walk you through it all in this video!

Prefer a quick read? Check out the highlights in this short presentation!

Brand Perception Analysis Presentation.pptx

INTRODUCTION

On an average, a company allocates approximately 9.1% of its revenue to marketing (source), with a significant portion directed towards campaign planning & content creation (40-50%), paid advertising (20-30%), and the rest on workforce marketing and events. (source). But what is the purpose of investing not only so much money, but also a lot of time and effort, into marketing?

The current market is vast, and regardless of the industry, every company faces thousands of competitors. In such an era, it is crucial to stay ahead and build a good perception among consumers to succeed. But, how can one measure a brand's perception?

In the past, measuring brand perception was challenging and required a lot of manual investigation. However, in today’s digital age, where customers' opinions are readily available all over the internet. It is easy to find both side's perspectives about any brand online. So, this project focuses on harnessing the power of these opinions by building a framework capable of collecting data about a brand across the web, analyzing them, and providing insights on the area of improvement. This framework allows a company to measure its customer perception and tailor their marketing strategies according.

HOW IT WORKS?

The reviews were collected from TrustPilot, grouped into related topics, and analyzed for sentiment.

Core Concepts

EDA: To analyze the basic statistics and generate insights from the data.
Topic Modelling: To group similar reviews based on the underlying discussion topics.
Classification: To predict the sentiment conveyed within each review.
Dashboarding: To analyze the sentiment across topics, dates, countries and other demographics.

Tools and Algorithms Used

Python, Tableau, SQL, Alteryx, Postgres Database, Excel
NLP, Latent Dirichlet Allocation (LDA), Sentiment Analysis, NaiveBayers, Decision Tree, Support Vector Machine

RESULTS & INSIGHTS

Global Reach and Local Voice

Uber operates in over 100 countries. Customers from various regions, including the United States, Canada, the United Kingdom, India, and Australia, shared their experiences on Trustpilot. The analysis focused on four key areas of customer feedback:

Convenience and Service: Customers used words like "booking," "app," and "pickup" to describe how easy it is to hail a ride through the Uber app.
Quality and Comfort: Terms like "clean," "smooth," and "dirty" reflected customer experiences with the vehicles themselves.
Driver Behavior: Feedback included words like "driver," "friendly," "respectful," and "professional" regarding driver interactions.
Ride Experience: Customers used words like "bumpy," "delay," "fast," and "convenient" to describe their overall journey.

Mixed Bag of Reviews

The analysis revealed a mix of positive and negative sentiments across all categories. Customers shared both compliments and complaints.

Convenience and Service, Ride Experience: These categories received a balanced mix of positive and negative reviews
Quality and Comfort: This category received the most positive feedback, suggesting that cleanliness and comfort are strengths for Uber.
Driver Behavior: This category had the most negative feedback, indicating a need for improvement in driver courtesy and professionalism.

Trends and Insights

The data showed a positive trend in customer perception of quality and comfort over the past three years. This suggests that Uber is delivering well in these areas. However, driver behavior reviews took a dip in late 2022. While there's a recent upward trend in positive reviews, it's important to address this concern.

Geographic Nuances

The analysis also revealed geographic variations in sentiment. Customers in India expressed more positive sentiment, likely due to their perception of good ride quality. In contrast, Australian customers leaned towards negative feedback, possibly linked to driver behavior issues.

RECOMMENDATIONS

Investigate Driver Concerns: A thorough investigation into the reasons behind negative driver behavior reviews in Australia and similar markets is recommended.
Take Action: Based on the investigation, Uber should implement targeted solutions, such as driver training programs or stricter monitoring mechanisms, to address these concerns and improve customer experience.

Explore insights on the Tableau Dashboard!