Do fashion buyers want help from AI?

Sign up to receive the Vogue Business newsletter for the latest luxury news and insights, plus exclusive membership discounts. Buyers are the original trend forecasters. Tasked with going to shows and ‘re-sees’ to pick out the pieces they expect to be hits with consumers, part of the job is to anticipate what people will want to buy months before shoppers are able to hit purchase. (Or, at least weeks before pre-order.) Buyers, at their best, curate a compelling selection customers can’t find anywhere else. So what role should data play? This week, trend forecasting agency WGSN launched the Fashion Buying platform, a data-driven hub that collates intelligence and forecasting, combined with WGSN’s TrendCurve AI predictive analytics (which it has been developing for three years now). The platform aims to assist buyers across the pre-planning, development and in-season hindsight phases. WGSN’s artificial intelligence prediction tool pulls e-commerce data, which it’s found to be the best predictor of trends, alongside catwalk data and a search and social index. The output of the model is the percentage of the assortment that item or category should represent, forecasted up to two years ahead, explains Francesca Muston, VP of fashion at WGSN. Aside from general trends, colour and materials tracking, the platform also offers a ‘TikTok Trading’ section, devised to help buyers identify opportunities for remerchandising existing inventory to tap emerging TikTok trends quickly. For retailers who have been grappling with how to navigate such trends, the offering bodes well. The Fashion Buying platform is the latest in a line of tech that looks to refine trend prediction in the new age. Stitch Fix, for example, uses AI to predict trends, which its internal merchandising team uses to inform which items to send which clients. Tapestry Group’s consumer insights team is experimenting with AI to supplement its qualitative trend forecasting research (using ChatGPT and similar tools). LVMH is also experimenting with AI forecasting. Digital wholesale platform Joor provides brands with the data analytics to inform product assortment and market trend decisions — but its use of AI is so far less geared towards buyer decision-making. On Tuesday, Pantone launched Pantone Color Insider, a colour trend predictions and intelligence service for those in creative industries. And on the wholesale side, AI-powered “digital warehousing” company Stork makes suppliers’ available inventory visible and sellable for digital wholesalers using AI-driven insights to predict trends and curate product recommendations. It’s worked for Farfetch and Rue La La. “We very often hear buyers say they wish they had a ‘crystal ball’, as they never have access to future-facing information, and we’re able to provide them with just that,” says Monisha Klar, director of fashion intelligence at WGSN. Was AI the crystal ball buyers had anticipated? It’s hard to say. As evidenced by the impact of the algorithmic recommendations that now fully dictate our social media feeds, overreliance on AI technology can go badly for differentiation in fashion. Per an individual’s algorithm, the same trends and products are regurgitated over and over. If buyers are shown the same trends to anticipate, are we at risk of repetitive inventory across retailers? How much data is too much? All about the blend Buyers want balance. By and large, they’re open to using AI to enhance their work on a case-by-case basis. “In luxury fashion, it’s a 50-50 scenario,” Mytheresa chief commercial and sustainability officer Richard Johnson says. “Many big-name designers have built substantial businesses with iconic items that consistently sell well over the years. However, our customers also value innovation and especially the new and unexpected.” Buyers have long relied on data to inform their selects. But to date, this has largely been internal. Laura Baker, co-founder of New York boutique Essx, would be keen to see more robust data beyond Essx’s own inventory. “Data is everything,” she says. “I would be curious to gauge bestsellers on a commercial level, items that retailers and the market know will always sell.” Johnson is aware of the limitations of Mytheresa’s internal stats. “While all the analysis we produce — and we produce a lot — is important, it will only get you so far,” he says. For this reason, he’s keen on AI assistance for trend, style and colour prediction as an assist to the retailer’s existing strategies. “However, the real skill of our buying team lies in knowing when to lean on historical analysis and when to trust their intuition to place educated bets on untested new ideas,” Johnson adds. WGSN knows this. “We are not here to replace the role of the buyer. Each buyer knows their customer and their market, so they will interpret the information in a way that will work specifically for their requirements,” Muston says. “Knowing the projection of boho blouses is vital in ensuring you don’t overestimate or underestimate a trend. How you execute that boho blouse is still at the discretion of the buyer and the designer. A buyer will also know if their business will over or under-index the market in a particular category, for example party dresses, so can adjust accordingly.” “Integrating AI into our buying process should be seen as an enhancement rather than a replacement,” says Mytheresa’s Johnson. “It’s about using technology to inform and support our decisions while preserving the human touch that defines our brand.” Baker agrees: “Human curation from the buyer is essential as it protects our store and our brand’s point of view.” Keeping up post-buy Many of the WGSN platform’s features existed before this moment; its TrendCurve AI tool was already up and running, and trend forecasts have long been available to buyers. But the ‘TikTok Trading’ feature offers a fresh answer to an increasing industry pain point: how to navigate the ever-faster-churning trend cycle. Instead of focusing on how retailers can catch up to TikTok trends, the goal is to help them move alongside them, says Muston. “There is a lot of misconception around the role of TikTok in trend forecasting. From a purely chronological point of view, TikTok tends to hit after retail, not before it, which makes sense since the product is already out there for influencers to feature,” she says. Where TikTok is most effective is the naming and marketing of trends, Muston adds. “Not all TikTok trends are fast and viral, and not all of them translate to commercial opportunity,” she continues. “Many of our clients end up chasing after viral TikTok moments, airfreighting new product when they have existing product in their inventory that speaks to that trend.” The goal of this feature is to offer retailers insights into the aesthetics and items that will fall into a trend, before the TikTok trend moniker is coined. Essx’s Baker would be keen to use that knowledge to develop an effective merchandising strategy. “There are many psychological factors in understanding how consumers behave and make decisions when shopping,” she says. “Any help in keeping the attention span of the TikTok audience is beneficial in my book.” While this is sure to alleviate certain pressures, this should still be taken with a grain of salt, buyers agree. There can be such thing as too much data, according to Johnson, in that an overreliance can flatten a retailer’s offering — they may not always want to run with the trend of the moment. “While data is invaluable, fashion also thrives on creativity and intuition. Over-relying on data can potentially stifle the creativity and risk-taking that is essential in luxury fashion,” he says. Baker agrees: “There needs to be an individual POV to push boundaries.” Comments, questions or feedback? Email us at feedback@voguebusiness.com. More from this author: Nara Smith and the new era of influencer marketing America’s election chaos spells uncertainty for fashion Mall brands’ comeback partner is the cult fashion label

Make better business decisions

Make better business decisions