REI was on a mission to improve its supply chain.
But in that process, what it got was much more, says Rick Bingle, senior vice president of supply chain at outdoor gear and apparel retailer REI.
About two years ago, REI began working with IBM’s Order Management system. Prior to this, REI had a consolidated view of its inventory, but the system would not allow shoppers to have full access to it. For example, if a shopper purchased a product online at REI.com and selected to pick it up in store, a distribution center would ship it to the store for the shopper to pick up in a few days. With the new system, REI is now able to use store inventory to fulfill an online order for buy online pick up in store using that store’s inventory for pickup within the same day.
REI’s sales increased $100 million in the first year of using the IBM Order Management tool by tapping into this previously unseen store inventory, Bingle says.
“We calculated that we were saying, ‘No’ 800,000 times every year because we weren’t accessing the inventory in the stores itself,” says Bingle about not being able to fulfill online orders from stores.
This Order Management system also allows REI to put rules in place to determine the best location, either store or distribution center, to fulfill an online order. For example, if a shopper in a certain location orders a product, the system’s rules will first check a certain distribution center based on that shopper’s location, and if that center doesn’t have it, the rule will then tell the system to check the store nearest to the shopper. And so on.
But REI wanted to go beyond this rules-based approach to fulfillment and factor in more considerations besides distance to the shopper or split shipments, which is when a retailer needs to fulfill one ecommerce order with several shipments from multiple locations. Bingle knew that REI’s goal is not always centered on shipping costs. The retailer’s priorities change throughout the year depending on the season.
For this reason, REI decided to use IBM’s newly built Watson Order Optimizer tool that uses artificial intelligence to factor in and change with the different goals REI has during the year, such as product margin and shipping speed and preventing out-of-stock status in stores—in addition to reducing fulfillment costs—and match that to its inventory in its three distribution centers and 155 stores.
“This tool was designed for company outcomes not for supply chain outcomes,” Bingle says.
REI has three major time periods it focuses on—the two seasons for its products, summer and winter, plus the holiday season. Nearing the end of the summer or winter season, REI stores might start to discount its products to make room for new ones. The AI tool can see if a markdown date is coming up in a few weeks and put a cost to what that might do to the product’s profit margin and how popular that product is at that particular store with its typical sales volume. Then, if a West Coast shopper orders a product that is close to its markdown date in an East Coast store, REI may choose to pay the extra $2 in freight costs to ship it from an East Coast store, instead of having a closer fulfillment center or nearby store fulfill that order if the markdown date is further out, Bingle says.
This type of calculation would be hard to configure with a rules-based system, he says. Plus, Watson Order Optimizer allows REI to “feather in” that consideration or to turn it off if the retailer has a different goal in mind, such as speed of shipping to a shopper.
These markdown considerations also help if a shopper orders something online and returns it to a store that doesn’t typically carry that product. A store will likely start to discount the product until it sells. Now, REI can ensure that if an online shopper orders that product, that store will be prioritized to fulfill it, even if it is farther away from a shopper.
Order Optimizer also helps with what a retailer typically wants to reduce—split shipments. Split shipments can drag on costs because of having to pay freight costs for multiple boxes.
When REI shipped online orders solely from its distribution centers, its split shipments were about 1.11, or for every one order, REI used 1.11 boxes, Bingle says. Once it began using store inventory, split shipments surged to 1.4, he says. Now, with the Order Optimizer, its split shipment are about 1.25 or 1.30, he says.
“That’s a pretty dramatic improvement,” Bingle says. “It doesn’t sound like it—it sounds small, but when you ship millions of boxes, it starts to add up.”
While this 1.3 level is still higher than the original 1.11 level, that’s OK, he says. Previously, REI wasn’t taking advantage of all of its inventory and the advantage of shipping products from stores, such as reduced fulfillment costs because of proximity to the shopper.
About 35% of REI’s ecommerce orders are either picked up in store or shipped from a store, Bingle says. Within that, the breakdown is roughly 50-50 between ship from store and buy online pick up in store, he says. REI’s fulfillment centers are “fully omnichannel,” meaning they can replenish store inventory and ship directly to consumers, Bingle says. He says REI may soon open another fulfillment center.
The Watson Order Optimizer tool “is like a brain” that sits on top of IBM’s order management platform, says Jeanette Barlow, vice president, product management at the Watson Supply Chain team. REI began working with IBM on implementing the tool in spring 2018 and rolled it out in late October 2018, in time for the holiday season.
For the holiday season, REI’s supply chain focused on ensuring it could get packages to shoppers in time for Christmas. The tool, which could factor in localized inventory from every store, allowed REI to fulfill more orders later into the holiday season than previous years, he says. Plus, REI was able to push its “receive it in time by Christmas” date back by one day to Dec. 22, Bingle says.
Ecommerce sales during the holiday season (mid-November through December) grew year over year in the “high double-digit teens,” Bingle says.
REI did not disclose the cost of Watson Order Optimizer. After an implementation fee, IBM charges a software-as-a-service delivery fee staring at $50,000 annually, which varies based on complexity and the number of transactions it processes, Barlow says.
Recreational Equipment Inc. is No. 73 in the Internet Retailer 2018 Top 500.Favorite