When H&M boosted its shares last month by reporting a rise in the sale of full-price garments, it wasn’t just a tribute to the fashion sense of its designers. It was a sign that backroom improvements are at last paying off.
The world’s second-largest fashion group is investing heavily in areas like artificial intelligence and customer loyalty as it looks to improve the way it spots trends and plans logistics, and ultimately reduce discounted sales and piles of unsold stock.
Arti Zeighami, H&M’s head of advanced analytics and artificial intelligence, told Reuters the strategy is starting to bear fruit as the company extends pilot projects that seek to use data to match supply and demand more closely.
“Allocating the right goods to the right stores in the right markets is one of the key projects we are working on,” Zeighami said. “For 2019 we have huge plans for growing that and hopefully, by the end of next year, covering globally.”
In the age of social media, fashion companies have less power to drive trends, which come and go much more quickly as influencers promote their “outfit of the day” on Instagram.
That poses a particular problem for H&M, which produces most of its garments in Asia, far from its major markets, making it less responsive than its rival Zara-owner Inditex which boasts it can get new designs to its stores within a week.
Sportswear brand Adidas admitted last month it had been caught flat-footed when its suppliers failed to keep up with strong U.S. demand for its mid-priced clothing ranges.
H&M had seen stocks of unsold goods pile up over the past three years.
In the quarter through February, inventories grew to 40 billion crowns ($4.3 bln), or 18.6 percent of sales, but H&M said they consisted of a higher share of clothes that are newer, thus less likely to be sold at marked-down prices.
It has said this is a sign its overhaul is working, and it expects a better offering and improvements in buying and logistics to help it reduce inventories to between 12 and 14 percent of sales by the end of 2022.
“Companies like ours once dictated fashion in a certain way. Today fashion is growing organically: you have influencers, you have communities,” Zeighami said. “It is hard sometimes to quantify. Is it orange or pineapple, tassel earrings or choker?”
H&M Chief Executive Karl-Johan Persson said last month investment in AI was already helping predict trends and allocate garments to stores: “Over time, this will mean a lot of improvements.”
Companies in a range of industries are touting AI as the answer to their most pressing business problems, but many experts caution it may not live up to the hype.
Zeighami said he prefers to call it “amplified intelligence” because he wants to mine data to help humans make better decisions.
He cited the example of a maths model which showed mass market demand peaks when an influencer trend is going down. One designer he showed it to said having such data would have helped her stand up to buyers who jumped on a trend too late.
“She had the gut feeling, so us amplifying that would help her to take the right decision and overrule the buyers,” he said.
At rival Zara, merchandising teams use data gleaned from stores, webpages and app to adapt their designs, in addition to insights from social media, returns and reviews, with the entire stock gradually refreshed every four to five weeks.
Unlike Zara, H&M also has a fast-growing customer loyalty scheme from which it is harvesting information, in addition to analysing data from social media.
The club, which doubled membership to 30 million in 2018, is in 16 of the H&M brand’s 71 markets and will add seven more including the United States by the end of the year.
Samuel Holst, head of the H&M Club, said another eight markets would be added in 2020 and he expected to keep up the membership growth rate.
Holst hopes new functions such as members gaining bonus points for reviewing garments and, later this year, for sustainable actions such as recycling clothes, will help H&M understand shoppers.
“Knowing our customers – having this insight, knowing where, how and when they shop, knowing what they like – that is an important piece in how we will be able to predict trends,” Holst told Reuters.
“The better we know the customer the better we can do this,” Holst said. “That is the foundation for being able to have very good inventories with a healthy rotation of the garments.”