©eakgrunge and ©manonallard via Canva.com
Where Can Artificial Intelligence Help Minimize Returns?
Emerging artificial intelligence is promising to reduce the elevated rate of returns online, with the primary focus on diminishing the likelihood of returns before the purchase happens.
According to a study published in the Journal of Retailing, 30% of e-commerce orders are returned by consumers versus 9% for brick-and-mortar stores.
Much of the technology focuses on improving recommendations, supported by a host of virtual try-on features available for products such as apparel, eyewear, footwear, and cosmetics. These features allow consumers to envision how a product will look on them before making a purchase, reducing the risk of disappointment and returns.
A McKinsey study found that a staggering 70% of all online returns across fashion categories are due to size, style, and fit issues.
With the benefit of AI, MySizeID — an omnichannel e-commerce platform — informs online consumers what size will most likely fit their body type.
“If you select large, for example, and we think you need a medium, we alert you that you need a medium,” MySizeID CEO Ronen Luzon recently told Fox Business. “And if you still order the large, we understand that the return is because of size related [issues]. And then we recommend you again next time, and we can alert the retailer as well of the habits of their consumers.”
Similarly, Volumental and ShoeAI tap AI to solve fit issues around footwear. The firms use data on the shoes the consumer already wears as well as insights from what’s worked for others. Volumental’s mobile app lets consumers scan their feet with their cell phones.
“With machine learning, you’re training the database to understand the relationship between bits of data done at scale,” Brent Hollowell, Volumental’s chief marketing officer and general manager, told WWD. “We use that information to produce very high-quality recommendations. So if you don’t know what kind of shoe or brand or size you want, we can say, ‘well based on your feet, this is what would fit your foot the best.’”
Improving the accuracy of product descriptions, including better addressing shoppers’ questions and concerns, is another focus to reduce returns via AI. Stitch Fix’s AI-driven algorithm analyzes customer data such as style preferences, body type, and size to create customized product descriptions that are tailored to each individual customer. Amazon and Shopify also recently introduced AI tools for sellers to help improve product descriptions to drive conversions and reduce returns.
Robert Tekiela, VP of Amazon Selection and Catalog Systems, shared, “With our new generative AI models, we can infer, improve, and enrich product knowledge at an unprecedented scale and with dramatic improvement in quality, performance, and efficiency. Our models learn to infer product information through the diverse sources of information, latent knowledge, and logical reasoning that they learn.”
AI is further being used to steer search ads away from consumers most likely to make returns. James Poll, chief technology officer at Acorn-i, an e-commerce agency, told the Wall Street Journal, “What we can’t do is stop Amazon letting somebody purchase. But what we can do is boost the targeting for the audiences that we think are less likely to return.”
Discussion Questions
Where will artificial intelligence likely be most beneficial in reducing product returns? Do you see AI playing a major role in reducing retail’s return rates?
Poll
Loading …BrainTrust
Georganne Bender
Principal, KIZER & BENDER Speaking
Jeff Sward
Founding Partner, Merchandising Metrics
AI can certainly help reduce return rates, especially if they allow consumers to make more informed decisions around size. I also think AI can analyze reasons returns are made and help retailers remedy those problems. However, even AI can’t entirely eliminate elevated online returns because without physically seeing products people can never have perfect information around things like materials, feel, exact fit, and so forth.
I’m not overly optimistic: consider this statement A McKinsey study found that a staggering 70% of all online returns across fashion categories are due to size, style, and fit issues. Would AI help here?? It might if problems are systematic – i.e. the “sizes run large” at this company – but not so if the sizes vary by supplier, and are all over the place. Something like “improving the description” strikes me as naive, or perhaps I might say they’re missing the point: human beings are quite capable of describing things, me thinks the problem is the descriptions are designed to maximize sales…that is to say, this is a governance/marketing issue, not really a “descriptive” one.
AI will not change the habits of consumers that consistently overorder with the intent of returning some number of products but AI can assist in more accurately identifying the chronic returners and AI can assist in recommending the products more appropriate for the individual customer. The real cure for reducing returns is to make it less desirable for the customer to buy with the intent of returning some portion of the purchase.
The return rates are off the charts for online fashion in particular, and AI is clearly helping. Another tool that’s catching on is using user generated content (UGC) to show shoppers others wearing the exact SKU they’re about to put in their basket. It leverages AI to collect social media—photos and videos—and places them on the product detail page (PDP). And if there’s a fit, the retailer’s site can present curated options that the system now knows will fit. So, it’s doing double or triple or even quadruple duty: improving size and fit, reducing returns, and increasing basket size and conversion rates.
By the way, I really like the use of AI to essentially hide items from shoppers known to have a “returns habit.” Out of sight, out of mind. That’s out-of-the-box thinking (pun intended) that retailers should use as a model for finding other uses for AI in retail. This AI is extremely powerful and versatile stuff.
Apparel is definitely the biggest contributor to high return rates. Using AI to help customers make better size, style and color decisions based on more accurate and complete information will help reduce returns. AI can also help identify other causes for returns by sifting through enormous amounts of data and transactions. However, AI may not be able to break customers of bracketing apparel sizes with the intention of just returning those that don’t fit. One of the best strategies to reduce returns is to eliminate the practice of free returns, which several retailers are adopting. If there is a cost to return items, consumers will change their purchasing strategies.
If 70% of fashion returns are due to sizing then do something about sizing. Women’s sizing is all over the place; nothing is standard. Some vendors size medium fits a toddler, while others think anything over that size is beneath them so they stop there.
Consider the plus size mannequins in stores now. Each one has the same proportions as a size 8, just larger. Most women are not built like that. AI can be a big help with sizing provided it is fed the right information.
What the industry needs in order to solve the sizing problem is to create an international standard for sizing. Global business and scientific organizations created standards for many other areas in which they trade and/or collaborate and it is not unreasonable to create a sizing standard.
Sounds like AI will have the ability to profile products for various attributes, and then profile customers for attributes that either align…or don’t align. And then make recommendations. Great! But in the contest between AI and the 5 senses, the 5 senses will rule every time. And there are so many nuanced variables in shopping for apparel that while I’m sure AI can help make a dent in ecomm returns, I’m not sure how big that dent will be.
The other angle is the simple identification of those customers with high or abusive return rates. There is no reason retailers have to impose market wide solutions in order to deal with easily identifiable abusers. When a high rate returner is about to make a purchase, the checkout page can include a note. “We are happy to take your order. If you return it, there will be a processing fee of $X.XX.” Of course that will cause second thoughts. That’s the point. And of course it will cause some cart abandonment. That’s the point.
What Georganne said. Artificial intelligence is absolutely necessary to help reduce returns, but the elephant in the room is women’s sizes. Why, in the year 2023, do women have to buy two items and return one to figure out which size works for them?
In my view, the potential of artificial intelligence (AI) to significantly reduce product returns lies in its capacity to enhance sizing accuracy, address fit issues, and provide tailored product recommendations. By leveraging AI algorithms to analyze customer data, including body type and preferences, retailers can offer more personalized and accurate suggestions. This not only minimizes the risk of dissatisfaction due to sizing discrepancies but also contributes to a more streamlined and efficient shopping experience.
Additionally, AI’s ability to optimize product descriptions and search ad targeting can further play a pivotal role in reducing return rates by ensuring that customers make well-informed purchasing decisions based on their individual needs and preferences.
Returns can definitely be automated with artificial intelligence (AI).
In addition to reducing manual work, it will also speed up the refund process by reducing the time it takes to verify eligibility, generate return labels, and process refunds.
AI will help predict which customers are most likely to return a product, based on their past purchase behavior. How about that?
Utilizing this information, you can target customers with preventive measures, such as providing them with more accurate product information or offering them discounts.
Returns can be segmented by AI based on their condition, which can help you determine how to handle them.
Returns will be handled more efficiently with the help of artificial intelligence. Returns will become more efficient, accurate, and customer-friendly thanks to artificial intelligence.
In order to automate the returns process, Amazon uses artificial intelligence. Amazon’s AI system automatically verifies a customer’s return eligibility, generates a return label, and processes the refund. In just a few minutes, this process can be completed without human assistance.
I don’t speak for Walmart, but it’s possible they are using AI to predict which customers will return products most often.
With Al being a relatively new field for retailers, brands, and manufactures, I am happy to recommend businesses that develop eCommerce apps. One example is Appic Softwares. Rest assured I have no connections with them. I can make other recommendations, as well.
No matter what, make sure the company has a highly experienced app development team. Db
There are 2 sides to impacting the issues associated with returns – first is to reduced the amount over-purchase / mis-purchase which results in returns. The second is to streamline the overall process of managing the return. Both need to be addressed together.
AI can help to reduce over-purchase with technology like clothing size-estimation tools for consumers but also using predictive modelling to determine when a consumer might be over-buying (eg buying multiple sizes) & then offering incentives – for example discount on future purchases – for not returning product.
Based on the product category, it could also be possible to predict the likelihood of return – this could be used to influence the value at which the item is priced.
AI is not magic. Most AI use cases cases drive towards finding a hidden pattern, making a better forecast/prediction, or a combination of both. With returns, the goal is to avoid the return in the first place & AI can be deployed to find a hidden pattern or make a better prediction. Identification of shoppers with high frequency returns & products with high frequency returns – use that to manage the offer & terms of offer for specific shoppers & products. For sizing – use AI models that help determine fit & recommendations. In customer service understand the root cause of the return & look to educate consumers with personalized content, expertise, & customer support. In the end the reliability of the AI model will come down to the reliability & cleanliness of the data pumped into the model/algorithms.