Fascinating report from Indian e-commerce startup Flipkart showing how they assess your online shopping cart before you click to buy.

 

It’s being done in response to the trend of easy returns for fashion items. The much higher return rate for online retailers (vs. brick and mortar) has been a disaster. Figuring out how to minimize those returns is crucial for online retailers’ survival.

Through machine learning, they’ve been able to accurately determine what you will return based on the patterns of what you’ve looked at online, what you’ve put in your shopping cart, plus your actual size and fit (more accurate than you are even aware of yourself). With that information, they can change your returns behavior.

 

Here are the major highlights of the study:

In less than 70 milliseconds, the computer can decide, how much of a risk of a return you are.

Based on your past shopping/return patterns, they decide on reward or punishment, e.g., increasing your shipping charges, as a deterrent, or offering you a coupon as an incentive to make your purchase non-returnable.

To get real-time predictions, the researchers put together a “fully-connected” deep neural network, which is trained on numerous factors about products and customers. That trained model will then produce the instantaneous assessment of the customers’ cart to predict the probability of returns e.g.

• How many times a given article of clothing has been returned by anyone
• How many similar things you’ve put in your cart, e.g., the same shirt in different colors (doubling-up of items is a leading indicator of higher returns).
• The more items a person has in a cart, the higher the return rate. More than five products in the cart? Return rate goes up to 72%, whereas a cart with one product has return chances of 9%.

Because sizing is the #1 reason for returns, the researchers have developed something called a “personalized sizing latent feature.”

• They put the info from all your online purchases, over time, into the network which allows them to examine what you’ve got in your cart and compare your intended purchases to “sizing vectors which explain your body shape & fit for different brands & products.”
• Running all this info through the neural network, the program creates a probability score of potential returns.
• The model predicts the return probability for a cart as well as the exact number of products that will be returned from that cart. Based on that prediction, the decision is made on what rewards and punishments to implement, if any.

If you shop online for apparel and footwear, this ZDNet article is a must-read. Absolutely fascinating!

 

Read on below for more on how companies are growing their businesses by cultivating a data culture.

 

From Tableau Software (the company just bought by Salesforce).

Companies that are data-driven across the organization are growing at an average of more than 30 percent annually.

• And yet, a McKinsey report shows that only 8 percent of companies are achieving analytics success across the business.

 

Here’s what companies are doing to become more data-driven:

#1: Creating a data literate workforce confident in the language of data is vital to achieving business success.

• According to Gartner’s fourth annual Chief Data Officer survey, poor data literacy was rated the number one roadblock to creating a data-driven culture and realizing its business benefits.

#2: Cultivating a passionate community of data thinkers is pivotal to building a data culture that drives business value while generating collective excitement about data.

• How do they achieve this?  By encouraging an environment of sharing, e.g., lunch-and-learn sessions, data visualization competitions, and internal user-groups where people can discuss data challenges and debate solutions.
• As soon as a few people share what they can do with data, it creates a momentous force where people become empowered through learning and achievement and want to tell others.
• Keep in mind that even small changes can lead to a major cultural shift within companies.

#3: The right balance of control and freedom.

• If businesses give employees too much freedom with data, then they risk creating an environment of chaos, but if they enforce too much control, then they stifle innovation.
• In truth, too much control can lead people to rebel and create shadow systems, meaning organizations risk ending up with data anarchy rather than data culture.

Link to full article here.