Deep learning, an advanced form of artificial intelligence (A.I.), is all around us, and it’s growing increasingly ingrained in how we live and work. Deep learning tech is “the brains” behind the automated traffic control on our city streets, our sophisticated translation systems, and the fast, accurate facial recognition at our airports. It’s also an integral technology for any breed of self-driving car. Additionally, deep learning can also vastly improve both shopper experience and retailers’ sales, on and offline.
Technically speaking, deep learning uses artificial neural networks, software constructs that are inspired by the biological structure of the human brain. Neural networks are excellent at rapidly processing and understanding unstructured data such as video, images, audio or large amounts of text without requiring intervention from human engineers, who could not possibly keep pace with the A.I. Neural network tech particularly shines at visual perception, natural language understanding and the ability to predict behaviors such as online shoppers’ desires and purchase intent.
Some practical examples of the predictive power of deep learning in action include the ability to increase online ad response without increasing overall advertising spend, delivery of more personalized offers for shoppers based on browsing history or habits, or instant, deep insights about a particular consumer based on their recent purchases or shopping behaviors.
A recent global study co-sponsored by IBM and the National Retail Federation found that two in five retailers are already using intelligent automation (via deep learning or machine learning), and adoption is expected to double in the retail industry by 2021.
Retail innovators investing in deep learning today
Some of the early adopters of deep learning and other A.I. technologies include well-known retailers such as Macy’s, Target, Walmart and Wayfair. For instance, Macy’s is currently piloting chatbot-guided shopper assistance that uses natural language processing technology co-developed by IBM Watson and Satisfi Labs. Other retailers have rolled out more operational-focused uses such as how Amazon and Walmart use A.I. to optimize their supply chains, delivery routes or checkouts.
One of the first tasks assigned to Andrew Pole, a statistician and predictive analytics expert retained by Target, was a better understanding of new parents, who are invaluable customers for the big-box retail chain. Pole focused on studying pregnant women in their second trimester to capture their brand loyalty for years to come. By analyzing shopping habits, they were able to create a rather accurate “pregnancy prediction” score, optimize Target’s supply chain and boost the personalization of customer offers.
Since that early pilot that back in 2002, Target constantly analyzes a vast trove of data in order to make shopping more fun and relevant for its guests. Its tens of millions of shoppers lead to billions of transactions and interactions that cannot be put to use without the power of A.I.
More recently, Wayfair launched a visual search tool that lets customers take photos of products they like and find visually similar ones on its website or mobile app. Those search requests can be tricky when shoppers are asking to make a room more “bohemian” vs. “nautical” vs. “minimalist” via search terms they input. But advanced algorithms and computer vision make it possible for Wayfair to deliver just that.
Deep learning vs. machine learning
The phase “machine learning” was first coined in 1959 by Arthur Samuel, an A.I. pioneer who defined it as “the ability to learn without being explicitly programmed.”
One of the more recent examples that got a lot of attention is DeepMind, the research lab Google acquired in 2014 that introduced AlphaZero, an A.I. program that could defeat world champions at several board games. It used no instructions from humans on how to play the games. Instead, the program developed its winning behaviors through experience instead of explicit commands.
In contrast, deep learning is even more advanced than machine learning, and can handle much larger quantities of data very quickly for common retail applications, such as delivering a personalized offer to an online shopper in milliseconds that’s more likely to elicit a response. When it comes to fashion sections of ecommerce sites, for example, deep learning can review and classify garments into different colors, styles, textures and seasons, and deliver the most likely “best fit” quickly and accurately enough, based on customers’ preferences, almost like a personal shopper.
For a more in-depth, yet simple overview of the difference between machine learning and deep learning, this recent Medium article is a worthwhile five-minute read.
In brief, deep learning delivers even more mathematical calculation speed, as well as the “horsepower” necessary to recognize and analyze patterns, and find correlations between massive data sets. That makes it possible for retailers to achieve highly precise audience segmentation, dynamic or real-time pricing, and personalization. It can also help solve a widespread problem of ensuring brand safety online by avoiding ad placements next to dangerous or brand-inappropriate content.
As covered by Forbes this spring, “the greatest challenge of the 21st century so far for the retail industry has been adapting to the online world.” After consulting with more than 200 business leaders and experts, Kristina Rogers, global consumer leader for EY, the global professional services firm, predicts that A.I. will completely change the landscape for retailers and how consumers shop. Retailers who keep up and survive in the end will need to adopt to change and switch strategies to remain relevant.
RTB House is a global company that provides retargeting technology for brands worldwide. Based in Poland, the company set up its U.S. operation in April 2018.Favorite