Q3 Technologies helped the client in developing a machine learning based application that segments the customer information through the data collected from point-of-sale devices. The executives can provide better shopping recommendations and unique offers to each customer based on this segmentation and further analysis.

Client Background

The client is a global company with manufacturing in China and India and sales operations in North America, and the UK, and is a leader in the import, export, and manufacturing of high quality gems and jewelry across the globe. The client manufactures a large variety of diamonds that includes various shapes, colors, and clarity. Today, the client is one of the largest exporters of colored gemstones from India and also one of the largest exporters of studded jewelry.

Industry Landscape

Similar to other industries, online jewelry industry has started the race for bigger market share, albeit much later than their traditional counterparts but are indeed running much faster.

The global jewelry market is growing at a steady pace of 5% CAGR and is expected to reach USD 257 billion by 2017. Asia Pacific has emerged as the largest jewelry market fueled by surging demand growth in India and China, while USA still continues to be the single biggest jewelry market country. Currently, online jewelry market constitutes 4-5% of the total global jewelry sales but is growing at a much faster rate and is expected to capture 10% of the total jewelry sales by 2020. Candace Cheung, senior manager at eBay, says online shopping is the emerging trend in luxury goods, with jewelry and watches among the top three items with fastest growth in e-commerce. This is reflected by the fact that jewelry is the No. 2 category of sales for eBay.

Future predictions are also in favor of massive growth for online jewelry, especially in India and China. E-commerce in China is expected to grow at 25%, much faster than the traditional retail, and US headquartered Blue Nile, the largest global online jewelry provider is capitalizing on this opportunity. China is company’s second biggest market after US and clocked 37% YoY growth with almost USD 30 million in sales.

Opportunity

The client wanted to get away with the method of consolidating the customer reviews; performing some manual cleaning and structuring of data in excel files. Then going over each review and tagging them with various departments or service lines. In addition, categorizing whether the feedback was positive, negative or neutral.

To accomplish this, the client required a digital channel system for the sales/marketing executives to gather customer requirements and gain customer satisfaction. The application developed guided the revenue lifecycle from customer acquisition to upsell and retention, and tracks each consumer experience from start to finish and predict the kind of engagement.

Additionally, automating customer service processes with the help of machine learning enabled ChatBot’s to provide quicker answers to customers.

Solution

Q3’s data science team provided various functionalities of the application to meet the specific requirements of the client.

Natural language processing to model the various topics, extract the named entity recognition and apply machine-learning algorithms to detect the sentiment and classify the department the review is intended towards.

In addition, various readily available API’s like the AWS Comprehend can be used. However, in the current scenario, Python was used for extracting feedback data through web scraping and social media feeds, open source libraries NLTK, Gensim and Spacy were used for text processing and modelling.

Business Impact

Increase in ROI by 30% due to reduced manual effort in reviews extraction, cleansing and topic identification.
The Machine learning based application helped properly segment the customer through data collection from Point of Sale. The executives can provide better shopping recommendations and unique offers to each customer.
Their data-driven algorithms in the applications has helped to compare each shopper’s search history and past purchases to promote relevant products with targeted recommendations.