SUPPLY CHAIN SERIES: Copping a DC attitude
Glen Ferguson is an inventory wiz. When Salt Lake City-based Associated Food Stores consolidated its warehouse operations, Ferguson, the director of inventory management, was given the charge of transferring operations from four disjointed warehouses to the retailer-owned wholesale cooperative's 1 million-square-foot Farr West facility in Salt Lake City.
By the time he reached the last DC, in Billings, Mont., he had to be extra cautious, maintaining enough inventory at the old location to service the retailer customers in the territory, while continuing to stock up the new location before it went "live."
When it came time to flip the switch, the ending inventory at the Billings location was just $85,000 off Ferguson's estimate -- not bad for a DC that formally held an average of $60 million worth of product.
"It gave me chills," he recalls. "It scared me that we were so close on our estimate of where our inventory would be."
The secret to Ferguson's success in this venture lies at least in part in the Advanced Warehouse Replenishment tool from Scottsdale, Ariz.- based JDA Software, Inc. But he didn't use it in the traditional manner. He cheated a bit -- improvised, if you will.
"There are some triggers in the system, called shelf-life, which is used for perishable items and that helps prevent holding them too long," he says. "I used the shelf-life trigger to manage the inventory levels at the old facility, configuring the system to send me alerts in decreasing time increments, because we had to still replenish the old facility, but I didn't want to replenish it to the maximum that the system would say I should, if we were not moving."
It's no surprise that Ferguson is so adept at using the replenishment tool and at tweaking it to meet his needs. He began using it in 1996, when it was first installed, and is currently in the process of upgrading to a GUI-interface client-server system.
Putting it in place wasn't easy. Before the system's adoption, some of Associated's buyers, who had been using a pencil and calculator for 30 years, were resistant to change. Meanwhile, however, their warehouse buying was fragmented and disorganized, and as part of the company's re-engineering efforts, the buying function needed to be centralized and automated.
"There were lots of inefficiencies," says Ferguson. "We had a vendor community that was [also] calling on the major chains. We couldn't really perform, because of the fragmentation of customers and ad groups, and those types of things. Our c.e.o., being the visionary that he is, understands the necessity for us to combine and consolidate not only Associated Food Stores of Salt Lake City, but even the whole independent cooperative segment itself."
What it does
The warehouse replenishment system -- originally created by E3 before the company was bought by JDA -- now helps buyers create accurate forecasted demand and order projections, as well as stable order patterns. By eliminating tasks like entering and calculating orders, it frees buyers to focus on more profit-generating activities, such as buying in the most economic order cycle, and planning and executing promotions while maximizing forward-buying opportunities.
Using Advanced Replenishment's Order Policy Analysis, Associated's buyers can determine how often you should purchase from a vendor for maximum profit, rather than for the lowest possible cost or highest possible revenue. Instead of purchasing items on a fixed cycle -- say, weekly -- it allows the buyers to determine the order cycle that will maximize profitability, by considering such factors as carrying costs, acquisition costs, service-level goals, forecasts, bracket discounts, and purchase and selling price.
"In simple terms, the system looks at every vendor and every item, every night," says Ferguson. "Through the night batch it will determine whether a vendor is due to buy, based on the parameters of lead time, safety stock, and the cycle [on which] you should buy that vendor. When the buyers come in the next morning and log onto the system, it presents them with a hierarchy of tasks that they are to perform."
Buyers must first deal with Short Ships, or orders placed by AFS that are due in, but that the supplier can't deliver. These have to be deleted from the system so the buyer can recalculate and essentially rebuy that quantity from another source or for another date.
Following this task, buyers address late purchase orders. Due Orders are those orders that are scheduled to be purchased.
"Essentially, the system is saying, 'If we don't buy these orders today, write these purchase orders today, we will jeopardize the service levels for those vendors,'" says Ferguson. "It's the buyer's job to not only accept those, but also to validate whether they really are due. In other words, to make sure the maintenance is performed and make sure the information that the system is using to calculate that order is correct. Once they have done their due diligence in that, then they would be required to accept that order."
The system also relies on a series of checks and balances to alert the buyer and DC personnel to other events. For example, a vendor might not be due, but some of its items might be. In this case the service level isn't in jeopardy, but the buyer may have to act on it anyway.
"Say the vendor marks me out of refried beans," says Ferguson. "I go and minus that out of that vendor, and then recalculate it, but it still doesn't come up as due, because the vendor isn't jeopardized. But since that item will be jeopardized, I'll get what's called an order point check -- if the buyer can't afford to be out of that item, he can accept an order from that vendor, which could, in a sense, overstock them on other items. That's a judgment call made by the buyer and [based on] the criticalness of being out of stock on that item. The system does a great job in identifying those exceptions."
A new DC attitude
The beauty of the system is that it has allowed AFS to move from an old-world "warehouse attitude" to a new-world "DC attitude," says Ferguson. "Warehouse attitude is more stocking up, where the DC attitude is just-in-time inventory. Ten years ago, before our re-engineering process, we warehoused product. If our warehouses weren't big enough, we'd rent outside storage facilities, and we'd buy the inventory. At the end of the year, a P&G would walk into here and say, 'I have two warehouses full of Crisco shortening in the Midwest, and we need to get rid of that inventory, so we'll give you a $12 off-invoice to buy it.' We would take it and run some events to try and sell that inventory."
Because of the variety of data now used by the system to generate its recommendations in terms of profitability, the results it renders might not seem intuitive to someone who has been relying on a pencil-and-calculator system.
"Because of this, a large challenge from the [point of] implementation was the buy-in from the staff to make the changeover, and getting them to trust that the system was replenishing correctly," says Ferguson." The old way, we would check inventory weekly. Every Monday we'd look at Ralston-Purina, for example, and always think in terms of weeks. 'I want to maintain three weeks of inventory for everything from RP. If I move 10 cases a week, in three weeks I would need 30 cases. If I have 25 on hand, then I buy five.' Then I went to the next item. That's how you bought, the old way. You had a factor of how much inventory you wanted on hand, and you multiplied your average movement by that and bought the difference. Simple."
Simple on the surface, maybe, but the algorithms inside the software today are anything but simple. The system analyzes the acquisitions costs and compares them with the handling costs, then recommends how often a vendor should be bought. The final decision, however, is still based on the buyer's expertise and the current needs of the business.
'A matter of balance'
"You may have an item that, if you buy it every 10 days, you could make $10,000 on that vendor this year, but because you have to buy it every 15 days, you will make only $9,500," says Ferguson. "You're losing $500 in potential profit, based on the economics of buying, but when you go back and look at the fact that you're writing a couple fewer POs, you're getting a better pricing bracket altogether; maybe the warehouse doesn't have to handle it as much, so maybe $500 isn't too much of a decision-maker. It's a matter of balance."
Adding the dynamics of promotions creates an entirely new set of challenges and opportunities in the realm of warehouse replenishment, and that's the next area Ferguson plans to attack.
"I'm trying to understand the movement of our promotional sales to our retailers, and trying to bring that product in just in time," he says. "Sure, you can fill the warehouse with 20 truckloads of tissue. That's convenient, but it's not best for the business, and that's what we're looking for."
By the time he reached the last DC, in Billings, Mont., he had to be extra cautious, maintaining enough inventory at the old location to service the retailer customers in the territory, while continuing to stock up the new location before it went "live."
When it came time to flip the switch, the ending inventory at the Billings location was just $85,000 off Ferguson's estimate -- not bad for a DC that formally held an average of $60 million worth of product.
"It gave me chills," he recalls. "It scared me that we were so close on our estimate of where our inventory would be."
The secret to Ferguson's success in this venture lies at least in part in the Advanced Warehouse Replenishment tool from Scottsdale, Ariz.- based JDA Software, Inc. But he didn't use it in the traditional manner. He cheated a bit -- improvised, if you will.
"There are some triggers in the system, called shelf-life, which is used for perishable items and that helps prevent holding them too long," he says. "I used the shelf-life trigger to manage the inventory levels at the old facility, configuring the system to send me alerts in decreasing time increments, because we had to still replenish the old facility, but I didn't want to replenish it to the maximum that the system would say I should, if we were not moving."
It's no surprise that Ferguson is so adept at using the replenishment tool and at tweaking it to meet his needs. He began using it in 1996, when it was first installed, and is currently in the process of upgrading to a GUI-interface client-server system.
Putting it in place wasn't easy. Before the system's adoption, some of Associated's buyers, who had been using a pencil and calculator for 30 years, were resistant to change. Meanwhile, however, their warehouse buying was fragmented and disorganized, and as part of the company's re-engineering efforts, the buying function needed to be centralized and automated.
"There were lots of inefficiencies," says Ferguson. "We had a vendor community that was [also] calling on the major chains. We couldn't really perform, because of the fragmentation of customers and ad groups, and those types of things. Our c.e.o., being the visionary that he is, understands the necessity for us to combine and consolidate not only Associated Food Stores of Salt Lake City, but even the whole independent cooperative segment itself."
What it does
The warehouse replenishment system -- originally created by E3 before the company was bought by JDA -- now helps buyers create accurate forecasted demand and order projections, as well as stable order patterns. By eliminating tasks like entering and calculating orders, it frees buyers to focus on more profit-generating activities, such as buying in the most economic order cycle, and planning and executing promotions while maximizing forward-buying opportunities.
Using Advanced Replenishment's Order Policy Analysis, Associated's buyers can determine how often you should purchase from a vendor for maximum profit, rather than for the lowest possible cost or highest possible revenue. Instead of purchasing items on a fixed cycle -- say, weekly -- it allows the buyers to determine the order cycle that will maximize profitability, by considering such factors as carrying costs, acquisition costs, service-level goals, forecasts, bracket discounts, and purchase and selling price.
"In simple terms, the system looks at every vendor and every item, every night," says Ferguson. "Through the night batch it will determine whether a vendor is due to buy, based on the parameters of lead time, safety stock, and the cycle [on which] you should buy that vendor. When the buyers come in the next morning and log onto the system, it presents them with a hierarchy of tasks that they are to perform."
Buyers must first deal with Short Ships, or orders placed by AFS that are due in, but that the supplier can't deliver. These have to be deleted from the system so the buyer can recalculate and essentially rebuy that quantity from another source or for another date.
Following this task, buyers address late purchase orders. Due Orders are those orders that are scheduled to be purchased.
"Essentially, the system is saying, 'If we don't buy these orders today, write these purchase orders today, we will jeopardize the service levels for those vendors,'" says Ferguson. "It's the buyer's job to not only accept those, but also to validate whether they really are due. In other words, to make sure the maintenance is performed and make sure the information that the system is using to calculate that order is correct. Once they have done their due diligence in that, then they would be required to accept that order."
The system also relies on a series of checks and balances to alert the buyer and DC personnel to other events. For example, a vendor might not be due, but some of its items might be. In this case the service level isn't in jeopardy, but the buyer may have to act on it anyway.
"Say the vendor marks me out of refried beans," says Ferguson. "I go and minus that out of that vendor, and then recalculate it, but it still doesn't come up as due, because the vendor isn't jeopardized. But since that item will be jeopardized, I'll get what's called an order point check -- if the buyer can't afford to be out of that item, he can accept an order from that vendor, which could, in a sense, overstock them on other items. That's a judgment call made by the buyer and [based on] the criticalness of being out of stock on that item. The system does a great job in identifying those exceptions."
A new DC attitude
The beauty of the system is that it has allowed AFS to move from an old-world "warehouse attitude" to a new-world "DC attitude," says Ferguson. "Warehouse attitude is more stocking up, where the DC attitude is just-in-time inventory. Ten years ago, before our re-engineering process, we warehoused product. If our warehouses weren't big enough, we'd rent outside storage facilities, and we'd buy the inventory. At the end of the year, a P&G would walk into here and say, 'I have two warehouses full of Crisco shortening in the Midwest, and we need to get rid of that inventory, so we'll give you a $12 off-invoice to buy it.' We would take it and run some events to try and sell that inventory."
Because of the variety of data now used by the system to generate its recommendations in terms of profitability, the results it renders might not seem intuitive to someone who has been relying on a pencil-and-calculator system.
"Because of this, a large challenge from the [point of] implementation was the buy-in from the staff to make the changeover, and getting them to trust that the system was replenishing correctly," says Ferguson." The old way, we would check inventory weekly. Every Monday we'd look at Ralston-Purina, for example, and always think in terms of weeks. 'I want to maintain three weeks of inventory for everything from RP. If I move 10 cases a week, in three weeks I would need 30 cases. If I have 25 on hand, then I buy five.' Then I went to the next item. That's how you bought, the old way. You had a factor of how much inventory you wanted on hand, and you multiplied your average movement by that and bought the difference. Simple."
Simple on the surface, maybe, but the algorithms inside the software today are anything but simple. The system analyzes the acquisitions costs and compares them with the handling costs, then recommends how often a vendor should be bought. The final decision, however, is still based on the buyer's expertise and the current needs of the business.
'A matter of balance'
"You may have an item that, if you buy it every 10 days, you could make $10,000 on that vendor this year, but because you have to buy it every 15 days, you will make only $9,500," says Ferguson. "You're losing $500 in potential profit, based on the economics of buying, but when you go back and look at the fact that you're writing a couple fewer POs, you're getting a better pricing bracket altogether; maybe the warehouse doesn't have to handle it as much, so maybe $500 isn't too much of a decision-maker. It's a matter of balance."
Adding the dynamics of promotions creates an entirely new set of challenges and opportunities in the realm of warehouse replenishment, and that's the next area Ferguson plans to attack.
"I'm trying to understand the movement of our promotional sales to our retailers, and trying to bring that product in just in time," he says. "Sure, you can fill the warehouse with 20 truckloads of tissue. That's convenient, but it's not best for the business, and that's what we're looking for."