Wednesday, February 19, 2020

Elevating online shopping experiences through machine learning

Sep 04. 2016
Facebook Twitter



TO SERVE today’s customers, accessibility and convenience through online channels are simply not enough.

Customer experiences have become indispensable for every successful business. According to IBM research, 70 per cent of the customers have stopped doing business with brands with which they have had poor experiences. Some 74 per cent even felt frustrated when they were not offered products related to their interests.

On top of these concerns, speed to fulfil customers’ needs is paramount, as they are prone to instant gratification and impulse purchases. Hence e-commerce needs a swift personalisation engine to improve customer experiences.

Each proposition to customers should be dynamic and personalised.

Machine learning has been used by e-commerce giants like Amazon to provide customised product recommendations and best-seller lists. Not only does this save time for customers, it also helps segment them, taking into account their associations and existing contracts in a more seamless and instantaneous way.

Most conventional algorithms utilise previous keyword searches, the customer’s characteristics, historical clicks etc, to rank the most related items on the first priority.

Beyond textual descriptions, image recognition has been used to elevate customer experiences. Visenze, one of the technology firms applying visual analytics and machine learning, has provided customer personalisation solutions for e-commerce. “Search by Image” allows customers to shop by uploading visuals instead of typing keywords. It also recommends visually similar products to customers. By looking into styles, patterns and colours, related products will be shown based on relevancy instead of popularity.

E-commerce businesses such as Lazada and Zalora have used Visenze solutions to improve their customer experiences.

Applications of machine learning on customer experiences even transcend keyword and image search. Chatbot can help e-commerce utilise the large historical data set integrated with some pattern-recognition techniques to interact with customers.

Connected with mobile messaging applications such as Facebook Messenger, WhatsApp or Kik via APIs (application programming interfaces), Chatbot is adaptive enough to replace human personnel.

Imagine personalised conversations with a multitude of individual customers simultaneously without a single sale representative or call centre.

How would this work? Once you start chatting with Chatbot, it initially asks you about your interests. After answering what you are looking for, based on a machine-learning algorithm, it will show pictures of what you are looking for and some statistically related items with the corresponding prices.

From there, users can go back and forth with the bot asking it to show different outfits and accessories. Then the customer can tap on a picture, and it is then linked to the e-commerce website where the customer can purchase it immediately.

Moreover, this bot can send emojis to make customers feel as if they are chatting with a close friend before making a purchase.

A recent study from IBM shows that e-commerce enterprises applying personalised recommendations can increase sales efficiency or conversion rate by up to 5.5 times.

These “mass personalisation” machine-learning tools are developed to elevate customer experiences by improving customer understanding and relationships. They can also enhance cross-selling and up-selling opportunities.

Make no mistake, this new innovation will soon transform retail businesses.

Views expressed in this article are those of the author and not necessarily of TMB Bank or its executives. Panawat Innurak is a specialist at TMB Analytics. He can be reached at

Facebook Twitter
More in Business
Editor’s Picks
Top News