Download PDF

Customer Profile

One of the biggest e-commerce marketplace in India offering consumer products like books, media, electronics, clothes and accessories, footwear, lifestyle, and healthcare products. Their product assortment is more than 80 million products across more than 70 categories and caters to more than 100 million registered users via their mobile, desktop, and web applications. It boasts more than 150 million registered users and more than 200 million app downloads.

Business Challenge

The client’s marketplace offers a vast collection of products across categories for their millions of users daily. With the widely spread user and visitor base, the searches performed on the platform were also high in numbers and varied in nature. The success of the client’s platform relies on delivering the right results for the searches performed on the platform by the visitors and users.

The biggest challenge impacting the client’s growth plans was the increasing Search Pain.

What is Search Pain?

Search Pain is the term that denotes the dissatisfaction of the users when they are unable to get the right results for the searches performed by them.

Typically a higher search pain denotes a deteriorating customer experience which in turn reflects in lower conversions. This was alarming for several categories on the client’s platform.

Project Objectives

As search pain was the major concern for the client, our team aligned with their product team and derived the following objectives to achieve:

• Understand how the search functionality and algorithm works with the client’s platform
• Analyze and find out the major factors contributing to the increasing Search Pain
• Check the relevancy of search results based on the searches performed across categories
• Understand the correlation between the availability of the products searched for and their non-appearance in search results
• Provide suggestions to improve the search results for products searched across categories
• Deliver actionable insights to improve the algorithm behind the search functionality of the client’s platform

The Approach

With the objectives clear, our e-commerce team charted out an approach to achieve them and improve the search pain for the client. The highlight of the approach was the design of a mechanism where the client’s product team was able to receive regular actionable insights on the improvement of their algorithm for better performance.

The project was designed into 3 phases with each one of them addressing the challenges faced by the client’s platform. The phases of the project included:

Phase I

In this phase, the search experts in our e-commerce team aligned themselves with the client’s product team. In-depth technical and demonstration sessions were organized by the client’s team for the team to understand how search functionality operated on the platform. Our team performed a comprehensive study of the factors impacting the searches performed on the platform.

The most important factors impacting the searches were discovered to be:

• Long tail and customized keywords used by visitors were unable to deliver the right products, despite being available on the platform
• Incorrect spellings in keywords used for searches impacting the products displayed among search results
• Vernacular language searches were not delivering the right products resulting in lower conversions
• Product catalog issues resulting in non-discoverability of products available on the platform

These factors that were detrimental to the customer experience were studied and their impact was understood in detail by our team.

Phase II

The team was trained by our internal team and the client’s team on the search functionality of the platform. They were accustomed to the factors impacting the search results and contributing to search pain. This process knowledge helped the team to develop their plan of action and methodology to tackle the challenges.

Phase III

Our team performed an in-depth analysis of the search terms contributing to the search pain that was provided by the client. The purpose of this analysis was to identify the key reasons why the searches were not delivering relevant products that were searched for.

This comprehensive analysis helped the team to develop an LMS for the client’s team for improvement. This LMS included

• Spelling / Synonym Dictionary: For Hinglish, Hindi terms, and Fat Finger spell issues. This helped us strengthen the knowledge base for the analysts (the LMS), and our clients can use the data for recurring algorithmic machine learning
• Vernacular Repository: A translation repository was created for frequent searches that helped accomplish:

1) Infusion of the same into the system regularly – so that search starts reacting to those terms and

2) Easier analysis for a layman (who doesn’t know a specific language) – which eliminates the barrier of hiring only regional resources

Suggestions were included in the insights on how these key changes can be incorporated in algorithms behind the search functionality of the client’s platform. Over some months, the platform had started displaying intuitiveness towards regional keywords and leveraging autocorrect functionality, the search pain was brought down considerably.

The Results