Why Warehouse Networks Don’t Work

By Terry Harris, Managing Partner, Chicago Consulting

Most warehouse networks have built-in barriers. Overcoming these creates needless costs and diminishes the service networks provide their customers. The following lists five obstacles that distribution executives can gauge against their own warehouse networks. How well does your network work?

Inventory Isolation Adds Cost
The very idea of a warehouse network is to have separate warehouses in separate locations. Normally, each individual warehouse forms its own territory with its own set of customers or “ship-to” locations. Orders to destinations in the territory are then fulfilled from that warehouse.

This approach creates “inventory isolation” where the warehouse’s inventory is meant for its customers only and not others. Other warehouses’ inventory is for their customers as well. Operating this way increases inventory costs.

The higher costs stem from the need for more inventory because it’s isolated. This is due to the fact that creating more warehouses necessarily creates more slow moving items. Slow moving items require proportionately more inventory.

Dividing an area (country, region, etc.) means dividing the demand as well. While the total demand remains the same, the demand at the individual warehouse is lower. If you cut the country in half with two warehouses, for instance, then what was an item that sold 10 per month becomes an item that sells 5 per month in two separate regions – from two separate warehouses. The demand of every item decreases. This creates more and more slow movers.

Slow moving items have higher variability – their demand is less predictable. This is precisely the reason for inventory – to provide for the unpredictability and random nature of demand. Therefore more total inventory is needed when it’s isolated into separate warehouses.

Inventory Isolation Degrades Service
Lower service – lower fill rates and lower lead times, specifically – stems from holding or back ordering out-of-stock items (and whole orders) until the assigned warehouse is back in stock when some other warehouse could have provided the needed items. Not isolating inventory or allowing the inventory of the whole network to serve any customer’s order combats this problem. Many order processing systems prevent this from happening. Few do it well.

Some organizations aggressively isolate their warehouses and the demand among them. Some warehouse managers actively horde inventory for “their own” customers. This often happens with regional customer service departments that are linked to specific warehouses.

We’ve even heard some managers say, “we don’t want to fill orders from other warehouses because the sales figures get into our demand history and cause us to carry more inventory.”

Such nonsense! That a company should intentionally withhold products that customers want, just because it cannot handle its own data properly, borders on the Orwellian. As a matter of fact, if the order occurred once, it can occur again, auguring for its inclusion in the sales history for that warehouse anyway.

In addition, if the forecast in the assigned warehouse cannot recognize its own local demand, bigger obstacles ensue – perfectly good products will be “thought” to have less and less demand (because they get fulfilled elsewhere) so they become non-stock items in the original warehouse! How’s that for irony?

These difficulties customarily occur when shipments are used as the base data to forecast future demand. In fact, most companies use shipments as the base data for their forecasts setting themselves up for these kinds of problems!

Faulty Stocking Strategies Add Cost And Lower Service
All inventory systems have “stocking strategies”. These strategies may be hidden or subtle or hard to see, but they’re there nonetheless. Stocking strategies are incorporated in the safety stock (or re-order point) and order quantity calculations performed each month (or so) based on fresh usage or forecast data.

First, too few companies appreciate the importance of these calculations. Yet they are as significant as any in their entire supply chain. These calculations and the logic behind them specifies where you’d like the on-hand of every item in every warehouse to be. Inventory management systems strive to keep the on-hand at the stocking strategy. As a matter of fact, if everything else worked properly (forecasts were stable, no supply problems, and so on) the inventory position, the actual on-hand of every item, would be exactly as specified by the stocking strategy. It’s that relevant!

As such, these calculations have crucial consequences. They practically determine inventory performance – the service the inventory provides and the capital investment it requires. What could be more important? Few single issues in logistics have the power or leverage that the stocking strategy has.

Second, many of these calculations are performed in most primitive ways. Customarily the logic of the safety stock is buried in the computational bowels of a computer program that few question or understand. Out of sight, out of mind. What distribution executive knows how their safety stocks are determined? Who knows how order quantities are calculated? Who has questioned these methods or asked if there’s not a better way? Yet they are powerful influences on warehouse network performance.

Here are a few of the mistakes made in commonly used stocking strategies:

  • Ignoring Cost – The purpose of inventory is to provide service – to be available when customers want it, when they place orders. Because most customers have choices, high service levels are necessary to retain them, and ultimately, to stay in business. While it may cost a lot any service level, however high, can be achieved with enough investment – emphasis on “enough.” Given all this you’d think that costs would play a pivotal role in stocking strategies.
    Not so. Believe it or not, most stocking strategies don’t incorporate costs. The $10 item is stocked in the same way as the $1 item. For example, the strategy that stocks two weeks’ worth (or three, four,…) of each item is independent of cost. If your safety stocks are determined by forecast error, a common method, you ignore cost too.
    Interestingly a few strategies that do incorporate cost considerations (see “Wrong Emphasis” below) do so with unintended consequences – the opposite effect that’s sought. They actually lower service and increase cost.
    Practically speaking, when the inventory gets too high (too much investment), management cuts inventory across the board by reducing stock levels, order quantities or both. (Of course, the only effective way to cut inventory by this broad-brush approach is to cut the fast movers – the “good” stuff.) Sure enough, service goes down as a result. By operating in this way far too many companies establish a limit on the capital they’re willing to invest in inventory and accept the service consequences of that decision.
    Further, stocking strategies not based on cost cause expensive items to be widely dispersed “down” the supply chain toward customers. Expensive items (especially slow moving expensive items) should be kept upstream. Why should every Chevrolet dealer carry rebuilt engines?
  • Too Few Classes – The ABC approach suggests that items with different characteristics should be put into different classes and then each class treated differently. Yet most systems have only a handful of classes – three, four, five, or so – to divide among all their items, perhaps thousands. You end up with too many items in the same class. A thousand items divided between five classes leaves an average of 200 items per class – all treated the same way. Some companies have thousands of items in a class.
    If it makes sense to treat different items differently, then why not apply this idea thoroughly? Carry it to the ultimate – have a class for every item.
  • Wrong Emphasis – Often ABC classes are based on sales dollars with the “A” items, the higher sales dollar category, getting more emphasis and better service. In companies with relatively common gross margins (probably most), this emphasizes the more expensive items over the less expensive items of the same usage.
    For example, take two items with the same (or close to the same) unit sales. One costs $10 and sells for $20, has a higher classification, while the other costs $1 and sells for $2 carries a lower classification. This approach will stock more of the expensive items than the inexpensive ones, which, in turn, increases the investment the inventory requires but does nothing for the service it provides.
  • Dead Inventory – Every company has dead inventory (or at least sick and dying inventory). It’s a natural outgrowth of being in business for any length of time. Products just head south. What used to be a fast mover becomes a slow mover. Seen a Plymouth lately? As a matter of fact, any successful product will eventually deteriorate in unit volume because it’s successful – line extensions are introduced that divide up the market, competitors copy it, its earnings support technological advancements that supercede it, and so on.
    The sin of dead inventory is doing nothing about it. There are two things to do – preventing further buildup of inventory that will die in the future and dealing with the dead inventory you have now.
    The first – preventing more – is best handled with the stocking strategy. Here many companies stock the same items in the same way at several levels in the network – plants, central warehouses, regional warehouses, and so on. This is a misguided approach. Once inventory is dispersed throughout the supply chain it is subject to different (read lower) demand and deteriorates faster. This is why more expensive items should be held upstream in central locations.
    This is also why the addition of new items into regional locations should be done with great circumspection. Unless these items are proven movers, they can easily deteriorate in downstream warehouses, gather dust, get sick and “die”.
    Here’s a real example. Some companies have “a rule” to stock any item that had, say, three demands in the last year and not stock any item that didn’t. (Some of the seemingly most sophisticated companies use this rule.) This approach builds up slow movers too fast. We’ve seen new items represent as much as 10% of the items carried at a location as a result of applying this rule. If 10% of the items were added to a location each month, the item count would double in 8 months!
    The second step that needs to be done about dead inventory is to get rid of it. Most companies hate to write off inventory. It’s an automatic decrease in a balance sheet asset and in current earnings. Such things are done very seldom, then only after great soul searching. (Of course, this is another reason why many companies have too much dead inventory – they hate to write off the dead inventory they have. Not only is it dead, it’s been dead for a while!) The sale of a company is one of those rare occasions when inventory is re-valued and written off. Actually, enough dead inventory necessitates the sale of the company. More irony!
    Short of writing off inventory, the most effective way to get rid of it is to decrease its price, promote it and free up the cash – get what the market will bear for it. Of course, it’s easier if companies do this routinely keeping the dead stock to a small manageable amount than if they have to go through the trauma of recognizing great amounts of dead stock and dealing with it only in times of crisis.

Mistaken Territories Increase Costs
Warehouse territories are assembled for a variety of reasons – covering a sales region, closest to the customer, and so on. But if territories are not assigned on the basis of cost (the least cost, at that), then cost will be higher than it should be.

In determining territories based on costs, the inbound cost to the warehouse plays a role too, often a decisive role. Not including the inbound cost leaves them hidden and unable to properly influence territories.

When companies include inbound cost, their plant warehouses often increase in scope (because the inbound cost is minimal, sometimes zero) and their stand-alone regional warehouses diminish in scope. Many territorial assignments are not simple, but the inclusion of costs (total costs, at that) in the design of territories is necessary to insuring low costs.

Flawed Order Processing Lower Service And Add Costs
Warehouse networks and order processing systems should work together. One can defeat the other, however. Network problems created by poor order processing include the following:

  • Batch Processing – Yes, even today too many systems are still batch. More often, some aspect of them is a batch process. These include inventory allocation, order releasing (in both mainframe order processing and in WMSs), backorder purging, and so on. Of course, batch processes automatically build in unnecessary lead-time and prevent visibility to inventory availability.
  • Partial Orders – A problem related to inventory isolation (described above) is the way partial orders are processed. A multi-line order is only partially filled and the rest back ordered, when the order could have been filled completely from a nearby secondary location. By filling the order completely from another location, the customer usually gets their ordered items sooner and the transportation cost is less.
  • Shipment is cheaper because ground tariffs (LTL and Parcels) are mostly dependent on weight and only slightly dependent on distance. (Most air tariffs have zero dependency on distance.) So it’s typical to ship one complete order from a warehouse further away more cheaply than two shipments from a closer warehouse. Depending on the replenishment of the out-of-stock item(s) in the nearby warehouse, the customer frequently gets what she wants sooner too.
  • Think of it this way…as if you were a customer in, say Phoenix and were paying the freight. Having ordered five items, would you want the three in stock items shipped from Los Angeles today and the other two shipped “when Los Angeles is back in stock” or would you rather have all five shipped today from Dallas?
  • Early Reservation – Too many systems provide no visibility to when inventory will be re-supplied. To prevent future orders from not shipping when desired due to unavailability, many of these systems reserve inventory against future orders too soon. This ties up inventory that’s needed for other customers.
  • Too Little Consolidation – Lots of order processing systems routinely contain multiple unfilled orders for the same customer – often, by the way, created by poor inventory availability that necessitate backorders. Too many systems do not recognize and consolidate these orders into one shipment with lower costs, especially transportation costs.

TL, LTL and ground parcel tariffs are dramatically more cost efficient at higher weights. So it’s smart to consolidate what you can. Yet we routinely see companies make multiple shipments to the same customer on the same day. You can bet that there’s only one delivery a day, because that’s the way carriers operate, but two shipments are paid for. (Carriers love it!)
On the other hand, there’s a countervailing issue with some carriers – UPS is one of them. The UPS ground tariff takes a huge jump at 70 pounds. This motivates you to de-consolidate larger parcel orders into multiple smaller shipments. For example, the UPS tariff for a Zone 4 shipment of 120 pounds is about $51. Splitting this into two shipments of, say 50 and 70 pounds costs about $29 ($13 plus $16). That’s a 43% saving!
A difficulty in uncovering these barriers is that some are tough to see – almost invisible. In some organizations they are like the 7/8th of the ice burg that’s under water. Distribution executives, busy solving day-to-day and more apparent problems, have little time to think about seemingly minor or subtle ways their networks might not work.

Some of these problems “creep up” slowly, then suddenly, having passed some threshold, become major obstacles. (Sick inventory that dies over a period of time is a good example.) Others, while they may not build up over time, require constant hurtling.

Smart distribution executives, however, take time to scrutinize improvements they’ve not thought about before. Their networks work.