Best practices: Slot leadership

Some of the most rewarding investments in distribution operations target labor costs and productivity. Under the circumstances—SKU proliferation, shrinking lines per order, demanding service levels—it seems strange that slotting, one of the most effective tools for pick optimization, is widely underutilized.

“The odds are that anyone reading this story has tried to solve the slotting problem, either manually or with applications,” says Tom Kozenski, vice president of industry strategy for JDA Software. “We’re now at a tipping point. Slotting has never historically been expensive, but now the affordability and return on investment make it a no-brainer.”

Put simply, slotting is the science of ensuring fast-moving SKUs stay accessible and slow-movers stay out of the way. But the problem with slotting is how easy it is to ignore. Orders come in, and the No. 1 priority is shipping boxes. Since slotting is a planning function, it often quickly descends on the list of priorities as execution moves to the forefront. Soon enough, a forward pick location will run out of inventory and the panic begins.

“You see less chaos in a well-slotted warehouse because there aren’t all these stockouts or people yelling on a radio, calling Frank on the lift truck to drop a pallet,” says Keith La Londe, director of sales engineering for PathGuide Technologies. “They tend to be more professionally run warehouses. One of our potential customers toured someone else’s facility and felt that the operation wasn’t a good match for their own because the activity levels seemed so much lower. It was so much quieter. They later learned that the facility produced greater volumes than their own—with fewer people.”

Conceptually, it can be hard to justify moving product around if it doesn’t add immediate value. After all, the inventory manager has done their job, so why use operations resources to manage something that’s already been managed?

“There has to be a dedicated amount of time to do slot maintenance work prior to execution, and that’s where we’ve seen things fall down,” says Dan Basmajian, president and CEO at Optricity. “It requires senior management buy-in, and you have to see the value of handling the box not twice, since you will handle it twice anyway, but in advance. It’s still low-hanging fruit after 30 years.”

Slotting from scratch
Depending on the starting point, improved slotting can create tremendous gains. Bob Kennedy, vice president of business development for DMLogic, says companies that don’t use a slotting application could start using a spreadsheet, review it twice a year, and still see a huge improvement.

“Looking at it more often is better than less often, even if you’re a very sophisticated company and use a slotting solution, simply because perceptions of business and realities are often different,” Kennedy says. “Look at your slotting rules and configurations and see if they need change, at least on a quarterly basis.”

Consider an operation with 1,000 SKUs on one or more pick lines, perhaps with mezzanines and conveyors and employing 25 to 50 people working on high-volume picks. If that type of operation is not optimized, Kennedy says, the amount of walking is enormous, and a 30% travel reduction could be possible. “Just think,” he says, “of how many times a picker goes by a slower product to pick a faster one.”

La Londe says customers new to slotting might start with the top 100 fastest-moving items, the 100 slowest, and make gradual improvements. “It can be a huge undertaking,” he says, “but once you have the entire warehouse correctly slotted, it’s all about maintenance.”

For those working with a spreadsheet, Basmajian suggests starting with a velocity report that includes, if possible, the size of the slot opening. This will enable you to calculate how many days of inventory each location should hold. “That said, a spreadsheet can help you do it in two dimensions, but can’t account for all the other dimensions,” he says.

One of the most fundamental balancing acts is between picking and replenishment. A pick location with one week of inventory means replenishment only needs to drop one pallet a week, but pickers have to travel an extra 10 miles in that time. On the other hand, if there is only one hour of inventory in a slot, pickers walk inches, but replenishment personnel are swamped.

Optimal slotting might consider picking, replenishment, receiving, shipping, storage media, the SKU weight, dimensions, packaging and fragility, product family groupings, store-ready pick sequences, SKU velocity and volatility, marketing and promotional initiatives, and as many as 20 other criteria. That’s where software applications come in.

Keep Bs from creating honeycombs
Slotting strategies generally come in two fundamental flavors: fixed and dynamic. The fixed slotting strategy strives to optimize item locations in pick areas as each item’s velocity tapers or grows, or as SKUs come and go. According to Marc Wulfraat, founder and president of consulting firm MWPVL International, the greatest focus on optimizing slotting is among distributors and retailers handling 5,000 or more SKUs. Those whose SKU bases remain relatively consistent throughout the year will lean toward fixed slotting, and those with greater volatility or extreme seasonality will tend toward dynamic slotting.

In a fixed slotting system, a new item is assigned a fixed location as soon as it is received. Still, that marriage between SKU and slot is temporary for as long as it is appropriate, whether six months or a week. In a dynamic system, new and familiar SKUs are assigned to open pick locations as they become available. When integrated with the warehouse management system (WMS), dynamic slotting will inform replenishment to fill a new forward pick location before the previous one is depleted, and the SKU might never return to its original slot. The new location might be 5 feet to the left or five aisles over. The system might also make the old and new locations active simultaneously, or only temporarily activate the second location without permanently undoing the first.

“However, warehouse management systems with a single-bin limitation are probably the largest deal-breakers for good slotting,” La Londe says. “It’s not unusual to encounter someone who keeps 100% of an item’s stock in a primary bin, most often due to a limitation of technology.”

Even if software supports it, dynamic slotting has its downsides, Wulfraat warns. Forward reserve inventory placed above the pick slot will not necessarily move as the pick location does, which can potentially increase travel distance for replenishment labor. Dynamic systems might also call for a greater number of slots—as many as 1.25 times those in a fixed environment—to ensure availability for new and incoming SKUs. “We have actually helped customers undo dynamic slotting in situations where it had gone bad,” Wulfraat says.

Whether an end-user had a bad experience or not, there is a lingering perception of slotting as a large process requiring partial or complete shutdown. “That has really changed dramatically,” says Peter Schnorbach, senior director of product management for Manhattan Associates. He emphasizes consumers’ massive appetite for new products, which has driven retailers to introduce new items daily.

“Warehouses aren’t nearly as static as they used to be, and they need to get new products into active pick areas as fast as possible,” Schnorbach says. “There is much more tactical slotting, and daily adjustments are not uncommon. When an item comes in, it is slotted, delivered to a location and picked. It doesn’t happen the next day or two days later, it happens in seconds, right up front.”

In the past, a new item would need to spend some time in a random slot before enough sales data could be collected to make a decision about its optimal location. “That is no longer the case,” Schnorbach says. “There are all kinds of capabilities around forecasting, like looking at the activity of similar items or product families, and we can blend those properties to create reasonable forecasts.”

Where to slot the crystal ball
A slotting system with some amount of demand anticipation can have a dramatic impact on operational efficiency. Whenever possible, Wulfraat advises clients to do more than look over their shoulders to make slotting decisions based on a SKU’s velocity in the previous 12 weeks. “What you really want to do is get marketing data in advance,” he says. “If you’re expecting seven times the volume for two weeks, you can proactively re-slot the item ahead of the sales lift—even if it’s done with manual notification.”

However, as with single-bin limitations, many existing business systems are simply not set up to combine relevant data into valuable forecasting tools. Basmajian says that less than 10% of Optricity’s customers have forward-looking capabilities, and most of them are overseas.

“For the rest, there’s typically a disconnect between buyers, what they buy and outbound demand,” he says. “They’re motivated and incented differently than shipping people.”

Looking at order history is easy, because the WMS knows what it did last week or month. According to Kozenski, the hard part is an accurate forecast the WMS can digest. “It has to come from merchandising or the ordering system, and if there’s no interface for that the WMS has no idea what is coming and has to wait until orders come down,” he says. “Maybe you do some quick re-slotting at that time, but you must not interfere with pickers. Many customers are not chasing this aggressively enough, and those connections between systems are not being built.”

Even without those bridges, slotting can be a powerful tool for modeling future activity. What if a facility adds 1,000 items, or volume goes up 20%, or another company is acquired?

Jon Kuerschner, vice president of pre-sales for HighJump Software, has prior experience as an end-user. Having looked at slotting from the eyes of vendor and customer, he has come to value slotting as an essential planning tool.

“The general focus of slotting is to make sure the last leg of warehouse activity—picking, staging and loading—is efficient,” Kuerschner says. “To prepare those functions, we looked at it from a weekly perspective based on new promotional schedules and fast-mover predictions. Then we reviewed slotting quarterly based on how products moved, whether e-commerce, in-store or drop-ships.”

Kuerschner has witnessed the domino effect each relocation can create, the chain of item movements just to free up space to slot a fast-mover in the right place. That’s where an economic modeling function helps. “At pick location X with Y volume, it would cost this much over a month,” he explains. “If we move it, the costs change. If there’s a big enough delta between the two and the moves are not too onerous, you can make that call.”

Basmajian offers the example of a customer who used the modeling capabilities of a slotting tool to inform the wider supply chain. On the first simulation run, the customer focused on nothing but efficiencies inside the four walls. The goals and constraints minimized selector travel and built good, stable outbound pallets. They could see the savings before they committed and saw a six-figure benefit.

“Then, before he pulled the trigger, he put on his team-player hat and tried a store-friendly model. There was far less benefit for him, but big savings at retail stores,” Basmajian recalls. “In either instance you have to change the labor plan. If you don’t, first you have selectors finish an hour early one day, 45 minutes the next, and by the end of the week they have learned to manage the workload across the shift. The lesson is that if you improve slotting optimization, make sure you manage related resources.”

Companies mentioned in this article

• DMLogic
• HighJump Software
• JDA Software
• Manhattan Associates
• MWPVL International
• Optricity
• PathGuide Technologies