MAIN NEWSLETTER SIGN UP

Amazon Sizes Up Trillion-Dollar Returns Challenge, Enlists AI for Perfect Fit Solution 

Amazon wants you to feel confident while shopping at home. They want you to make more confident purchasing decisions, and they want to lower the likelihood of you returning a product, which has become a $1 trillion problem

In fact, according to Karen Webster, returns pose significant logistical and financial challenges for retailers and have become an inconvenience for consumers. In response, retailers have implemented return fees and collaborated with innovators to offer alternatives such as immediate store credit, aiming to retain funds within their ecosystem. 

“But so far none of that is making a big dent, as more commerce moves online and returns just keep piling up and retailers aren’t solving the real problem: There is no sizing standard to help consumers pick the right size,” said Webster in her latest article.  

Read more: The Eight Pivotal Strategies for Payments and the Digital Economy in 2024 

In light of this, Amazon is leveraging artificial intelligence (AI) to gain insights. By addressing size and fit challenges through AI, the company aims to enhance your shopping experience and reduce returns. 

In an announcement on Monday (Jan. 8), the company revealed a few recent AI-enabled features that let customers confidently shop on Amazon for fashions and styles. 

AI Size Recommendations

“Not sure what size to order? Nobody wants to spend time cross-referencing reviews, size charts, product images, and detailed product descriptions,” said Apoorv Chaudhri, director of computer vision and machine learning at Amazon Fashion in a company blog post. That’s why Amazon employs AI for personalized size recommendations, delivering customized suggestions on each product’s detail page. 

The algorithm groups customers anonymously based on similar size and fit preferences, as well as products with comparable fits. This involves analyzing millions of product details, including style, size charts, and customer reviews, along with billions of anonymized customer purchases. 

Chaudhri also says the algorithm also adapts to changing size needs over time, considering sizes retained or purchased by customers with similar preferences. For example, if a customer buys child pants in a specific size this month, the algorithm anticipates potential size changes in the future.  

Amazon uses advanced AI, particularly large language models (LLMs), to enhance personalized size guidance for its customer base. The system generates review highlights for individual customers, simplifying the process of determining how a particular style will fit. It provides advice on sizing based on insights from customers who purchased the item in the same size, drawing from details like size accuracy, garment fit, and fabric stretch. The AI condenses this information into an easily understandable review highlight, so customers can quickly access the most relevant details without manually going through numerous reviews for each item. 

Reimagined Size Charts

To ensure accurate and user-friendly size charts on Amazon, LLMs are also used to simplify the extraction and refinement of product size chart data. This includes standardizing sizes, removing duplicates, and correcting errors for a more precise size chart.  

Chaudhri also noted that Amazon is looking to improve the visual appeal and understanding of size charts, experimenting with creative methods like grouping measurements for the recommended size, making it more convenient for customers to find the information they seek.