by John Mai
Approximate read time: 3 minutes

We live in a world where technology is currently advancing and evolving faster than ever. Who would have thought we would program computers to recognize images, learn patterns and make decisions themselves? Advances in ecommerce have been extraordinary – most would not have anticipated 20 years earlier that we would be ordering everything from toilet paper to cannabis tinctures via the web. Machine learning, which is a subset of Artificial Intelligence, has a great deal of application in the ecommerce industry that transcends far beyond just analytics. Here are just a few significant applications of how machine learning can add intelligence to ecommerce.

We order practically everything online - yes that includes recreational drugs

Let’s break down these buzzwords, Artificial Intelligence (AI) and Machine Learning. Machine Learning is a subset of AI where an algorithm is trained using an existing dataset. Let’s take housing prices as an example. Given an existing dataset with various house prices, we can accurately determine which features of a house affect the price and by how much. Machine Learning figures out the weight/impact of each feature. The more rooms and bathrooms there are, the higher the house price will be - giving these factors a heavier weight. Similarly, zip codes and location will greatly affect the price of the house, hence will be weighed higher. The learning process continually updates weights of each feature until it is able to predict a price as close to the actual price as possible. By the end of the process it creates the most optimal weights to provide the most accurate predictions as possible. “But wait a minute, I sell clothing/accessories/groceries/electronics, not houses. Does this apply to me?”. Yes! All of this applies to different verticals of ecommerce. In the clothing vertical for example - using deep learning, a branch of machine learning, we can identify patterns, colors and shapes in images to determine which product items are similar. Combine the automated product discovery with intelligent recommendations based on past sales data and you can deliver an optimal shopping experience that is personalized to each shopper.

Machine Learning is a process to train an algorithm from the data

In any ecommerce business, the vast inventory of products must be filtered down to showcase only the most relevant product to each shopper. As a merchant, you have a very short window of time to capture the user’s attention and interest. Successful businesses are the ones which are able to quickly surface the most relevant products from their inventory and translate that interest into conversions. Vast majority of your (irrelevant) inventory will remain unseen by each shopper. And that is how it should be. This is why personalization is absolutely critical, and where machine learning shines. Machine learning enables businesses to leverage their existing data and reveal notable insights about trends and customer preferences. A study by Janrain showed that “73% of customers are disinterested by being presented with irrelevant content”. AI based tools trained using Machine Learning offer merchants the ability to personalize shopping experiences for each of their customers and deliver a more optimal shopping experience. A customer-centric experience is something that a customer will not only appreciate but is likely to remember, increasing the likelihood of those customers becoming loyal and repeat customers.

You have to capture the shopper's attention with relevant products before they lose interest and wander off

Such powerful tools are no longer exclusive to giant ecommerce corporations with dedicated Data Science teams. Please drop us a line about the challenges you are facing and we will be happy to brainstorm solutions with you.