Imagine a world in which apparel retailers and manufacturers truly know what customers will buy. One in which they can anticipate which products will be returned because of poor fit. One in which they can bring each store team’s attention to the products in the highest demand to ensure availability. And one in which they can drive sales by offering the right pairing to increase basket size—and complete the look.
It may sound like the latest fantasy film, given the relatively high rate of markdowns that fashion retailers face. In 2018 alone, markdowns cost non-grocery retailers $300 billion—or about 12 percent of sales. Meanwhile, retailers lose $1.75 trillion a year due to out-of-stocks and overstocks combined.
Such missteps are clearly not from a lack of data. As Harvard professors Marshall Fisher and Ananth Raman write in The New Science of Retailing, retailers are “awash in data, but starving for insight.”
With so much information coming at any given time, it can be a challenge to keep up—for a human, that is. Machine learning (ML) and artificial intelligence (AI) can help to sort through the information and deliver actionable intelligence.
For apparel retailers in particular, the insights could have a tremendous impact on the bottom line—and set them up for success as shopping undergoes a transformation. According to EY’s FutureConsumer.Now research, in the near future, customers will leave commodity purchases to AI-enabled search assistants. With the “chore” shopping for basics out of the way, customers will spend their time shopping for purchases that will “help us express and shape our identities and give us an amazing experience.”
Read on to learn the areas where AI is making the biggest splash in apparel retail.
Fewer Unhappy Returns
Nothing removes the “thrill of the hunt” faster than slipping on a pair of jeans that fits poorly. British retailer ASOS has softened the blow of rampant returns by using machine learning to analyze which items and sizes customers keep and compares that to the items and sizes that are most frequently returned. It can then make appropriate size recommendations regardless of brand or fit.
ASOS is not alone. Levi’s addresses the same issue with its Indigo chatbot. The chatbot helps shoppers select the right item from its 20,000-unit catalog by guiding a shopper to the specific size, color, or style. Indigo can also guide with recommendations, or direct the shopper to promotions, a physical store, or a live agent.
Brady Stewart, Levi’s SVP Americas Digital, told WWD that customers who interacted with the chatbot were 50 to 80 percent more likely to convert—and that returns were reduced because Indigo provided better guidance on fit.
Anything that reduces returns can make an impact on the bottom line, especially in apparel where 13% of all purchases are returned. AI that reduces returns can quickly provide ROI—and start impacting the bottom line.
Machine learning also is the engine behind smart recommendations, significant drivers of additional sales. Amazon, for instance, has turned them into an art form. Still, shoppers only click on the recommendation in 7% of all visits, according to a Salesforce report. But what a return: those clicks represent 24% of orders—and 26% of revenue.
AI tools can make product recommendations based on items the customer has purchased in the past, or on their shopping history, including price data. Smart product recommendations also can help overcome the biggest hurdle in online shopping: out-of-stocks in a size or color. Recommendations that include similar items can save the sale.
Having the right product, size and color in the right quantities in the right location is even more important in physical stores. Lucky Brands took this challenge head on by striving to maximize product add-ons in their stores. The retailer uses ML developed at MIT to determine specific assortments to be allocated at store level. Before, each product was allocated individually; now, it is done based on groupings of styles that ML predicts will appeal to the same shoppers. This helps Lucky Brands reduce waste and upsell more effectively at each location.
Right Here, Right Now
H&M is another retailer using technology to manage store inventory at a hyperlocal level—something that was do-or-die after the retailer announced it had $4 billion in excess inventory in 2018. The retailer turned to data and AI to provide store-specific assortments. One of the first stores to adopt the technology—based in Stockholm—discovered it had mostly female customers, who were drawn to items like floral skirts and luxury items.
To better serve the customers shopping there, H&M cut the number of SKUs by 40 percent, removed most menswear, and increased the number of high-priced SKUs like bags and cashmere sweaters. The company said sales at that location improved significantly, but did not disclose figures to the Wall Street Journal.
This type of technology also works in reverse to reveal when products are underselling and failing to meet demand. When applied to a retailer’s basic POS data (what sold, where, and when) CB4’s advanced AI algorithms can recognize once indistinguishable patterns. Anomaly detection then reveals when SKUs are failing to meet demand. Store managers respond and act on the data quickly to request replenishment or investigate whether floor issues that are preventing sales. AI technologies like this can “tell a retailer how to align product drops to match demand, and even how to display products in a store to sell as many as possible,” according to Business of Fashion.
The Big Picture on Apparel AI
There is little doubt that machine learning and artificial intelligence have major potential in apparel retail. And as retailers see successes in predicting demand, pricing, and store-specific assortments, technology investments will continue. Juniper Research predicts retailers will spend $7.3 billion on AI by 2022—up from $2 billion spent in 2018. It’s a worthwhile investment; AI allows brick-and-mortar retailers to access the same level of data that e-commerce has enjoyed.
But it is an investment not to be made lightly. Throwing money at the problem without understanding the potential impact on the bottom line or customer experience won’t deliver the same results. But given the right situation and the right solution, AI and ML can go from a high-tech gadget into selling machines.
Learn how CB4’s AI and Machine Learning tool leads to clear, to-the-dollar revenue gains, an engaged store workforce, and happier shoppers.