In the quest to build the dynamic and agile distribution center, analytics, artificial intelligence, and robotics help the drive towards optimization. It’s time to dispel the notion that data science is purely abstract and theoretical: In fact, it can and should be applied to make a practical difference in evolving warehouse performance.  

A study of warehousing workers conducted last year by Wakefield Research yielded some remarkable findings, at least for those of a certain age. Nearly three out of four floor workers said they would consider a pay cut to work at a company with better technology. Among younger workers, that proportion reached 81%. The workers perceived that a company’s investment in technology is in fact an investment in them — 90% believe new technology helps attract and retain floor workers.

In an industry facing labor shortages, that’s pretty good information to have. But, when it comes to the myriad benefits DCs can derive from exploiting the latest in data science and digital technologies, it’s the tip of the iceberg.  

Operators are striving to build dynamic and agile DCs, and data science can help them do that. The latest systems can help supervisors, managers, and executives with decision-making by providing timely information and help optimize processes such as product slotting. Data analysis and artificial intelligence orchestrate tasks among robots and human workers to make workers’ jobs easier and better. Data and technologies are also being used in the gamification of warehousing tasks, providing line workers with continual feedback, which drives productivity improvements and employee engagement. And technology is being used to keep tabs on performance across a range of facilities — all in the quest to drive process efficiencies and continuous operational optimization. 

The warehousing worker study demonstrated that employees, especially younger ones, possess the sophistication to understand that data technologies help them with their jobs and make their work lives better. Taking certain tasks away from human workers, such as poring over endless data steams to make decisions or traveling from one end of the warehouse to the other to fill an order, makes the jobs of human workers more efficient and rewarding — and they know it. 

That conclusion did not come as a surprise to Mark McCleary, chief technology officer at Lucas Systems, a provider of distribution center productivity and optimization technologies. “Technology is ubiquitous,” he says. “When workers get to the warehouse and put away their phones, having to work with pen and paper becomes a struggle. Warehouse workers can clearly see that technology can help them do their jobs more effectively and efficiently, and that brings better job satisfaction.” 

“It helps with the quality of their life,” adds Kyle Franklin, a Lucas senior solutions consultant, “just like workers who might take a couple of dollars less an hour to have an easier commute. People want to be empowered to do their jobs well and providing them with good technology helps them get through the day easier.” 

A growing number of DC operators are using machine learning and other data science tools to drive efficiency and productivity and to optimize processes from one end of the DC to the other. “People are becoming more aware of the importance of data,” says McCleary, “and specifically good data engineering. There are many disparate systems still being used in warehouses from conveyors to cloud-based WMS and web applications, and good data engineering is necessary to find usable intelligence. If the data is not accurate, if it's not cleansed, if it's not stored at all, it prevents any of those activities from being meaningful.” 

Today’s cloud technologies, says Franklin, allow DC operators “to bring in larger datasets and pull out more insights.”  

The product slotting process is one example of how data science can help optimize processes in the DC. Slotting has been traditionally performed with Excel spreadsheets, with planners trying to match fast-moving SKUs with advantageous warehouse locations. “It’s all very manual,” says Franklin, “which is why it takes a long time, and it’s the kind of thing that just might get left by the wayside.” 

Slotting decisions can be made on the basis of product flow, which takes into account the cubic dimensions of the products and the space it takes within certain locations in the warehouse — but it takes a great deal of data analysis to accomplish that task. “Using machine learning techniques, it’s possible to tease out trends and build better forecasts,” says Franklin. “It’s possible to drill down to get granular trends week to week and day to day to make better decisions around where to slot products in the warehouse.” 

One benefit of optimizing product slotting is to reduce travel time around the facility. “Once you’re able to do that,” says Franklin, “you’re going to see picking productivity start to skyrocket.” 

Managing the workforce in near real time also benefits from data science.  “An operations manager might need to make personnel swaps within the warehouse based on the flow of activity in certain areas,” explains McCleary. “Being able to see activity through a dashboard is important to eliminate bottlenecks in the warehouse. Being able to monitor performance and tie it to individuals allows the operations manager to go in with precision to fix those problems.”  

For example, a highly efficient picking process might be feeding into a less-than-optimal packing station. That’s eventually going to cause a backup that might even result in an operations shutdown. “Adjusting personnel to balance that flow can help prevent those bottlenecks,” says McCleary. 

Applying artificial intelligence to analyze DC worker performance data allows systems to make recommendations on personnel changes that drive warehousing efficiencies. “The advantage,” says Dan Keller, a Lucas senior solutions consultant, “is to take as much of the manual decision-making as possible out of the hands of managers so that they can focus on exceptions.” 

A simple example: The system perceives that lower-than-expected performance levels on a certain shift are going to prevent the facility from hitting its order fulfillment numbers. The fix in that case might be to assign another worker to a certain area to speed things up. AI orchestration would automate the reassignment of users to that zone, without the need for manager involvement. “Supervisors would be made aware of these types of things so that they can intervene, if need be,” says Keller. 

Another example: The system is aware that Worker A, picking in Zone One, historically performs at a certain level. If that worker’s performance drops below that level, the system informs the supervisor to consider implementing a solution to get things back on track. “The system may look beyond that and say that it’s because there's a conveyor jam or somebody's not packing fast enough,” explains Keller. “With the new breed of data orchestration engines that are available, AI can provide much better and faster visibility to all of the things that are occurring in the warehouse and make recommendations on fixes.” 

Gamification is one way to use data technologies to engage warehouse employees to make their work lives more satisfying. “It’s about trying to meet people where they are by figuring out ways to bring some fun and competitive spirit to mundane tasks,” says Franklin. “It can be as simple as establishing picking leaderboards and handing out achievement badges. Those kinds of things provide a little extra incentive to perform well, and it helps workers move through the day.” Franklin also cautioned that any gamification initiatives should create systems that are scalable to different user groups, acknowledging that, “some users will enjoy it and others may be demotivated by it.” 

In addition, any discussion of personnel productivity improvements must include their automated co-workers. The warehouse worker study showed that over 40% had favorable views of working with robots, believing they would help reduce physical stress and help them achieve better speed and accuracy.  

“A system that's orchestrating robotics should also be orchestrating humans,” says McCleary. “The ‘why’ is twofold: Most warehouses that implement robotics still have humans working with them and the complexity of interaction between robots and humans is where things typically fall apart. It’s all about looking for opportunities to improve overall performance.” 

The big benefit to deploying robots in the DC, given the current state of the art, is to reduce human travel time. “A dramatic reduction of travel time means that the ratio between material handling time and travel time will go up,” says Keller. “Human workers should be spending much more time handling product and less time moving between locations. 

“Most of the industry agrees today that collaborative robots, or cobots, are the way to go,” Keller adds. “Collaborative environments need to be fostered by software that has visibility into where the people are and where the robots are and can manage the orchestration of labor and machine. AI is very good at doing that.” 

Getting beyond the individual DC, data analysis can be used to compare the performance of different facilities in an effort to understand where improvements need to be made. “If one DC is performing at a lower rate compared to others that have a similar product mix,” says McCleary, “it can be indicative of personnel or process issues.” 

The quest to get a consolidated view of operations is challenged in many cases when enterprises have implemented an array of disparate systems. “They may be integrated and working together,” says McCleary, “but each is providing its own actionable intelligence.” 

That’s why it’s necessary to deploy an analytics platform to bring data together and compare metrics to show how systems are inter-connecting. “Business intelligence tools can aggregate all of those systems’ data together to derive actionable intelligence and make more informed decisions,” says McCleary. “And that ranges from the operations manager on the warehouse floor all the way up to the executive level, where they want to know how their technologies are working and whether they’re getting the results that they expected.” 

For companies wondering about the right timing to make these investments, the answer is, sooner rather than later. “Most of these technologies offer a rapid return on investment,” says McCleary. “And once you've gotten that return, everything else is cost savings. Implementing these technologies now means getting those cost savings much earlier.  

“If you don't have advanced analytics and you don't have good insights at the enterprise level, you're losing money,” he adds. “The longer you delay and wait for the data to tell a story, the more that you're missing out.” 

Resource Link: www.lucasware.com