Brian Amend, the Senior Director of Supply Chain Systems at Walgreens
Walgreens
You hear the term “digital transformation” everywhere. That means different things to different people. But one thing many would agree it means is putting in new applications where none exist, or perhaps more often, tearing out old legacy applications and replacing them with more modern software platforms. But rip and replace is expensive, time consuming, and risky. Isn’t there a better way? Walgreens
WBA
Walgreens is a large drug store chain. Their large footprint stores sell a wide variety of goods besides just healthcare items. This is a large public company with 8,900 stores in the US, Puerto Rico, and US Virgin Islands. They generated $132.5 billion in sales in their last fiscal year.
Walgreens, like many other retailers, has embraced omnichannel. But the scale of the company does add complications. They serve 9 million customers daily. They have over 200 million item/store combinations. The sheer scale makes accurate order promising to online customers difficult.
Walgreens and COVID
Prior to COVID the company had a three-year plan to add omnichannel order flows by adding buy online/pick up in store and curbside pickup. They wanted to support a better, more seamless experience across in store and online shopping. And ecommerce and omnichannel fulfillment costs more. They also wanted to improve their cost to serve in these areas. The giant retailer was contemplating a full rip and replace of their legacy distributed order management solution – the key solution that supports omnichannel order fulfillment.
“Then COVID hit,” Brian Amend said. Mr. Amend is the senior director of supply chain systems at Walgreens. “Overnight customer expectations changed.” The company needed to accelerate their omnichannel roadmap. “We actually had curbside pickups seven years ago. We discontinued it because demand for the service was not there.” But now it was full speed ahead.
Mr. Amend explained Walgreens’ project in a speech at the Blue Yonder ICON
ICX
Walgreens Embraced a Microservices Solution
To get the capabilities they needed quickly, they could not rip out the legacy order management system and spend three years implementing a new one. Walgreens thought twice. They decided the existing solution had good workflow and pick and pack capabilities. What was needed was better available to promise capabilities. Instead of rip and replace, why not just augment the existing solution? But if augmentation meant adding custom code to the solution, that would also take too long and be too risky.
But technology has progressed. Today we have solutions built on microservices. A microservice architecture arranges an application as a collection of “loosely-coupled” services. “Loosely coupled” means changes in one component don’t affect the performance of another component. This means a microservices component can be brought to life independent of others. Loose coupling reduces all types of dependencies and the complexities around adding new functionality to an existing application or integrating that solution with other solutions from other vendors.
The ability to install a microservice to improve a legacy application does require deep domain knowledge of that application. But Blue Yonder, based on their acquisition of Yantriks in 2020, has the know-how.
Walgreens decided to implement a microservices solution from the Blue Yonder called Luminate Commerce to augment their existing solution. Blue Yonder told Walgreens they could get the advanced functionality needed implemented in 6 months. This would be a very fast implementation for this type of solution. So fast in fact, that Eric Orlosky, a senior supply chain manager at Walgreens, recalled he laughed when he heard the timeline. The pilot was implemented in 5 months. At the end of 7 months, they had implemented the solution across the chain of stores. Mr. Amend has high praise for the Blue Yonder team that supported them in their implementation.
Walgreens is ahead of most of the rest of retail in the service commitments they make to online customers. If a customer agrees to pick up the order in the store, Walgreens promises to have the items available 30 minutes after the customer hits “buy.” For home deliveries, Walgreens is delivering goods in as little as 1 hour for orders placed during a store’s business hours.
The Blue Yonder solution had improved algorithms on how to best fulfill an order based on the type of order – curbside pickup, pick up in store, ship from store, ship from ecommerce distribution center, or drop shipping – where the inventory to fulfill the order was and whether all the inventory for the entire order was present, and whether there was enough time to fulfill the order on time from a fulfillment location.
These new capabilities helped Walgreens grow their digital sales by 116% over the past two years. While fears of being infected have eased and people do not fear visiting stores, Mr. Amend does not expect digital sales growth to slow. “We have gotten used to the convenience.” Further, on products in short supply, like infant formula, customers want to know the product is in stock. Customers visit the internet site to make sure the store has the product so they don’t have to make a wasted trip to the store. Once online, they tend to just go ahead and buy it.
What Next?
Order management systems are real time systems. But the inventory planning systems that forecast where inventory will be needed are not. No forecast is perfect. This means that orders are often not fulfilled by the ideal location. The problem with this is that forecasts are based on history. If the order is fulfilled by a suboptimal fulfillment location, it looks to the demand planning application like the demand originated from that location. Walgreen is intrigued by functionality that better integrates inventory planning with order fulfillment. Over time, this functionality will improve inventory placement to support omnicommerce.
Secondly, inventory accuracy at stores is never as accurate as inventory accuracy at warehouses. That is because inventory at a store might be in customer’s cart and because of theft. Shrinkage is much greater in stores, and recent articles suggest this problem has gotten worse during COVID. The result is that a store does not promise all the inventory in the store. The order management system will not promise the last two items it believes the inventory system shows because it assumes that inventory may not really be available. But this buffer stock number is often somewhat arbitrary. Walgreens is interested in using machine learning from Blue Yonder to calculate an optimal buffer stock number by item and by store that better reflects the realities on the ground.
