Retailers collect plenty of data, but do they make use of it?
All of it?
These were typical of the questions inventory technology firm Nextail began asking when it was called in to look at Italian retailer Stefanel’s inventory management and controls.
Stefanel is a 50-year-old textile and apparel business which over-extended itself in the 1990s. By 2016 it was in bankruptcy, emerging a year later under the control of British asset management firm Attestor.
As a retailer, Stefanel operated as many as 1,500 stores around the world, including the United States and Canada. Today the chain has been reduced to just over 400 stores throughout Europe, concentrated primarily in Italy and Germany, more than half of them franchised. Though Stefanel no longer operates stores in the United States, it does maintain an online business.
Heading the day-to-day activities at Stefanel for just about a year is Cristiano Portas, the former head of swimwear brand Arena, managing director of Lego Italy and a veteran of the Procter & Gamble organization. One of the operations Portas oversees is the transformation taking place in Stefanel’s approach to managing inventory and stocking stores.
Nextail has been working with Stefanel almost since the retailer emerged from bankruptcy. The first thing Nextail does on a job is look at “how the organization works,” says its co-founder and CEO Joaquin Villalba.
AI AND ANALYTICS
What Nextail found with Stefanel was an inventory management system based on a nearly manual Excel database. Villalba describes it as an “unstructured way” of following stores’ inventory as well as an unstructured method of determining when and what to transfer based on stock in stores and customer demand.
“As retailers grow, complexity is increased,” Villalba says. As a result, “We have to push them from the start to become more limber.” The preferred business model is one that works from the bottom up, rather than the top down. The latter, he says, is “too rigid because of its authoritarian style and structure. You need flexibility at the top levels.”
In setting up a new installation with a retailer, “We identify the things that need to change. Then we go to the data. We start with what they have but historically don’t use.”
Retailers can tell you how an item did last week, last month or last year, he says, but not any longer back than that. “Retailers don’t keep historical information, so we have to put their house in order.”
Information is the heart and soul of Nextail’s system; it uses artificial intelligence and predictive analytics to manage inventory allocation, replenishments and store transfers.
The next step is to determine where “sales were lost because they didn’t have the right stock in the right place,” Villalba says.
In the end, Villalba says Stefanel was able to lower stock levels in stores and yet still increase sales by having the right merchandise in the right stores.
Nextail’s machine learning system minimizes human intervention and human errors while forecasting actual demand across the whole chain. The system also functions in ecommerce and omnichannel environments, though Stefanel has not yet utilized that capability.
The size and scope of the retail operation does not make much difference to the effectiveness of Nextail’s approach.
“We work with retailers who have merchandising teams of five to 10 members and others who have hundreds working in merchandising,” Villalba says.
There is not much effort or input required of the retailers, either. Once Nextail installs its system, he says, “the retailer can start using it the next day.”
Working mostly with collection-based retailers, primarily selling fashion apparel and cosmetics, Nextail helps users right-size the amount of inventory in stores. In addition, ordering and stocking cycles can be cut by more than half.
Using the mantra “it’s not how much you stock, but how much you sell,” Villalba says the important thing is getting SKUs where they need to be. This contrasts with the old retail method of allocating — a year in advance — similar amounts of goods to each store, shipping them for the season and marking down whatever doesn’t sell.
With Nextail, “You can move stock between stores if need be and improve customer experience,” Villalba says, or, rather than emptying the warehouse and sending everything to the stores where it might sell in three months, it might be better to keep items in the warehouse where they could be shipped to stores and sell the following week.
Taking a data-based approach, Villalba says Nextail isn’t interested in why something is selling, just in what is selling and where.
It took only a couple of weeks after installation for Stefanel to see the effect on replenishments, allocations and purchasing.
“We were more flexible in buying and in replenishment or for the moment of the season we are in — the weekly calendar, the weekly scenario,” says Vera Bortolato, global sales director for Stefanel. The system “is quite user friendly. It’s really easy to see which is the best scenario.”
Stefanel produces two major collections a year, one for fall/winter and another for spring/summer. Within that structure, there are five sub-collections as well as three events, each with different stock. Bortolato says Stefanel now tries to maintain a never-out-of-stock position.
Citing examples of the new flexibility Stefanel has, Bortolato says the predictive analytics are used “from predict-preview to the buy module of what we are going to do by size, color, style, etc. Now we can check what’s in stores very easily, if there need to be transfers.”
Bortolato also says that in comparing the old and new ways of doing business, “the percentage of errors was much higher” the old way.
Even attitudes within the company have changed with the new inventory management system. “It’s changed the mood of the company, it’s changed the culture,” she says. “Old habits were broken, and we became positive quite quickly.”
Just as importantly, there has been a change for the better for customers, too. Under the previous system, good service and support service were able to be maintained only in high-volume stores. “Now we are able to provide that in all our stores,” Bortolato says.
Artificial intelligence and predictive analytics may seem like buzzwords because of their widespread applications in things like facial recognition and computer-vision technology in a variety of businesses. “But in retailing, this is just getting started,” Villalba says. “It’s $2 billion now and will be $7 billion by 2022.”
David P. Schulz has been writing for STORES since 1982 and is the author of several non-fiction books.