The English term “boxing clever,” loosely translated, means behaving in a careful, clever way to gain advantage in a difficult situation. Today’s challenging logistics environment offers a new, more literal definition, as warehouses deploy advanced software to optimize their packing operations. 

A Rapidly Changing Environment

The COVID-19 crisis has only accelerated a broader two-decade market shift from traditional in-store retail to e-commerce. A war of attrition for market share among large online players like Amazon, Walmart and Best Buy is forcing change down to smaller online competitors and distressed brick-and-mortar mall stores. 

At relatively little incremental cost, retailers can provide distinct customer experiences remotely, display full inventories, adjust pricing and promotions in real time, and process payments instantly in the cloud, with free shipping and returns. But it’s the warehouse and last mile where all of that rubber hits the road. Managing customer expectations for free shipping, same-day or two-day delivery, with multiple items combined in a single package, isn’t easy. 

COVID-19 has taken e-commerce fulfillment to the next level, as shelter-in-place orders and store closures beginning in March had, by the end of June, driven an 18% year-on-year spike in non-store sales for the first half of 2020, to more than $456 billion versus $387 billion for the first half of 2019. Operational pressures to move freight quickly through strained fulfillment centers now coincide with an urgent need to wringe time and cost savings out of the fulfillment process to remain competitive. Omnichannel and safe delivery strategies add further complexity. 

As in football, it’s a game of inches in which small process improvements can yield outsized gains. A key area targeted for improvement is the distribution center packing station. 

Incremental but Compounding Improvements

Packaging optimization, or cartonization, has become a priority for volume parcel shippers of all kinds, but particularly in online retail. Time and materials are baseline considerations: packaging decisions are often made quickly, on the fly, by the number, size, weight and shape of items. Customers prefer receiving all items in an order together, in an appropriately sized box, but businesses aren’t always able to accommodate them. While automated “box-on-demand” systems can build cartons of any size from raw corrugated sheets, they’re often only cost-effective for larger, high-volume shippers.

Typical packing operations rely on specified sizes of corrugated boxes used individually or in combination to fit the most commonly shipped items. Warehouse workers under pressure to move orders out the door seldom have the training or experience to weigh operational and cost factors: Does nesting multiple items in a single box require a size that adds to corrugated and fill costs and heightens the risk of damage? Is the order too time-consuming or involve too many people in different parts of the warehouse to be put into a single box? Which products should or should not be combined in a carton because of their relative sizes, weights, shapes or fragility? Most packaging involves snap decisions — or plodding trial and error — to arrive at a “least bad” solution that can have dramatic impacts on profitability.  

AI Navigates Complex Packing Incentives

Dimensional weight (DIM) pricing, implemented by FedEx and United Parcel Service in 2014 has now become a standard rating method for other carriers seeking to “cube out their vehicles, confusing matters further. DIM pricing simplified decades-old truck and rail rate formulas by ending the rate tradeoff between weight and volume for smaller packages. It uses a “DIM factor” — a constant minimum weight per cubic foot of space — applied to package volume. In the process, DIM pricing achieves a second carrier objective; it offers a price incentive to optimize carton size to improve space utilization and stackability in the truck. 

Six years later, however, pricing is no simpler: As parcel carriers become more efficient, their rate and fee structures become increasingly complex—and it’s not just DIM fees anymore. Transportation cost and capacity vary widely depending on the mode and zone of a shipment, which translates into rate table incentives to pack in very specific ways. That, coupled with requirements to keep up with diverse consumer demands, makes packing decisions nearly impossible to eyeball. Complexity and speed is making tribal knowledge an inadequate basis for controlling cost-effective packing.

“Warehouse workers don't have the luxury of sitting down for an afternoon with their rate tables and a calculator before packing an order,” explains James Malley, co-founder of Paccurate, a Brooklyn, New York-based developer of cartonization AI software. “Their job is to try to keep up with the orders and get them out the door as fast as possible. Specialized AI doesn't need an afternoon; it needs less than a second.” 

Paccurate algorithms analyze the dimensions and weight of an item or items in an order, against negotiated rates and terms from one or multiple carriers based on client-established rules and exceptions, to determine the optimal pack for the item or order. Packing determinations can originate at the order entry, or in the wave picking, packing or shipping stage, and include a detailed 3-D visualization of the recommended pack. 

The program can be used in combination with box-on-demand systems to speed packing operations with fewer line workers, but more typically it optimizes packaging from a specific set of carton or pouch/envelope sizes and programmed shipping instructions. It operates on a broad principle of “items within boxes” that can optimize cartons as items on pallets, or cartons and pallets as items to assemble a truckload. Paccurate’s narrow AI — a capability to process large volumes of data and many variables to make complex decisions toward a specific goal — enables a degree of “predictive cartonization.” For example, the software currently underlies the order entry and shopping cart functions in numerous e-commerce sites, to calculate shipping cost as an order is constructed.

The company reports average reductions of 6% to 20% on total spend for multi-item shipments; 13% corrugated used; and 20% in package cubic volume. A choice of licensing and subscription options has brought prices down for smaller accounts, to as low as $25 a month for a basic version.  

Paccurate had its beginnings in 2008, when Malley and business partner Patrick Powers were designing transportation management software. “As we looked at cost saving opportunities, cartonization stood out like a sore thumb,” he recalls. “We saw a real need — people would keep coming to us and asking questions about how packaging could be made more efficient.” As DIM pricing took effect, the need became more widespread and urgent. By 2015, Paccurate had a prototype. 

Clients have typically included logistics software integrators, resellers and 3PLs offering packaging optimization in their larger service portfolios. The company currently serves a growing number of direct users, and Malley says they’ve recently seen a marked uptick in demand, as grocery and consumer goods retailers managing their own shipping grapple with the retail transition to e-commerce and adapt to COVID demand. “As carrier capacity has shrunk, the potential market has widened significantly,” Malley says. “For all of these essential goods like groceries, household items and cleaning supplies, it’s really important for them to pack efficiently, move through the warehouse, and get on the truck.”

Multiple Use Cases for Cartonization Tech

Three main types of customers benefit from packaging optimization software, according to Kyle Ous, senior analyst at Chainalytics, an Atlanta-based logistics and supply-chain consultancy that specializes in finding costs-effective, efficient and sustainable packaging strategies: 

• Businesses that are new to e-commerce, shifting some B2B sales to B2C to improve margins and stay close to customers, or experiencing sharp COVID-19-related or other surges in demand, and need to better manage materials and shipping costs;
• Companies with complex orders to a range of locations that need to make the most space-efficient packing decisions in less time; and
• High-volume customers looking to plug optimization software into their box-on-demand systems order to shorten packing times and reduce errors with a smaller workforce. 

Ous says that cartonization is getting more attention from businesses as shipping restrictions and volume demand raise costs, and as warehouses surge-hire less-experienced packers to keep pace with demand. “Whether it’s small, medium or large businesses, it might be hypothetical, but they can see the value and are trying to improve their decision-making about what packaging to select,” he says. Sustainability is a growing concern as well, whether from wasted packing materials in overboxing or from inefficient loading of trucks and extra trips that add to fuel consumption and CO2 emissions — a particular issue with less-than-truckload and last-mile expedited service, where demand has soared.

Pierbridge Inc., a Marlborough, Massachusetts-based shipping software provider acquired by Wisetech Global in 2018, incorporates the Paccurate API in its Transtream TMS multi-carrier management system, but also uses it to load full containers for domestic and cross-border shipping. “We’re seeing the technology in a lot of different scenarios,” says Pierbridge Vice President of Product Marketing Mike Graves. “The primary market is large retailers, for their retail supply chains, but also 3PLs providing logistics services fulfillment.” 

Big savings can be realized across the supply chain by first optimizing how items are packed at the SKU level, so that more boxes can be loaded into fewer containers. “Often, with single orders, to maximize loading of a container, shippers might do cartonization multiple times to get a quick estimate of where things stand,” Graves adds. “When the container gets to the warehouse, for example, they may re-optimize in case items are out of stock, discontinued, or can’t be located, and then consolidate again.” COVID-19, he says, provides another use case as traditional divisions in handling, delivery and pricing — sofas or home gyms, for example — become blurred amid store closures and more people sheltering in place and shopping online. “It used to be an easy question,” says Graves. “Is it freight or is it parcel? Today, local delivery can be either one.”

Charlotte, North Carolina model train manufacturer Lionel is a logical customer for automated package optimization. Lionel’s business is highly seasonal, tied to holiday shipments; it earns 60% of its revenue over the September-December period, its peak shipping period both inbound and outbound. Lionel manufactures overseas and receives trainsets, cars, modular track, transformers and accessories over July-August, for shipping to Walmart, Kohl’s, and Amazon and other retailers in September-October. At that point, says company chief information officer Rick Gemereth, “all of that business goes quiet and all of our direct-to-consumer business goes crazy — on the level of Black Friday and post-Black Friday.”

In 2019, manufacturing delays put the seasonal schedule back by a month. Retailers all have different packaging requirements in terms of how they sell, store and ship particular sets and individual products. As the holidays approached, seasonal warehouse workers were in short supply; inexperienced packing hires were taking too long to make efficient packing decisions by trial and error from nine different box sizes, special retailer requirements and complex carrier rate structures. 

Lionel first used Paccurate in late 2019 to generate packing instructions, based on standard product weights and dimensions, on the warehouse pick ticket, along with items in the order and their location in the warehouse. Pickers presented the items and the ticket at the packing station; the nesting of item and box sizes was pre-selected. Customization of the software allowed Lionel to easily bypass default settings to address special customer requirements, and send alerts when weights and dimensions needed for the DIM price were missing. 

Gemereth estimates that Lionel was able to handle a 20% increase in outbound B2C traffic in Q4 2019 without hiring additional workers, and save 10% on corrugated costs.

The next step in cartonization, experts agree, is machine learning to detect usage and demand patterns, measure performance, and suggest process and packaging improvements. But for now, the time and cost savings provided by specialized AI is a good start.

Download this article here.