HOW TO DESIGN A WAREHOUSE NETWORK?

Six Steps to a Low Cost, High Service – Warehouse Network That Beats Competitors’

By Terry Harris, Managing Partner, Chicago Consulting
Whether you are adjusting your current warehouse network or designing a new one…whether you are closing warehouses or improving customer service…and especially if you are reducing distribution costs, these six Steps will simplify your effort and streamline the process.

STEP I – Determine your Objectives
These are several objectives you might seek. Three of the most important are:

• Provide a high service network
• Operate a low cost network
• Design a network that’s better than competitors’

Recognize that your warehouse network is there to serve customers, so you need to have customer service as part of your objective. Most networks are designed to minimize costs within a given service objective. For example, an objective might be to:

Design a network with the lowest possible total distribution cost while delivering 90% of all customer orders complete and within three days and delivering the balance within ten days.

Surprisingly, however, warehouse network designers normally ignore competitors. Few take specific note of where competitors’ warehouses are located, what their service levels are and then design a network that provides better service. Smart warehouse network designers who do, gain a sales and market share as a result.

Whatever your situation, try to determine your objectives as clearly as possible. Put them in writing. Discuss them with others and don’t be hesitant to refine them when you learn something new or your circumstances change.

When Amazon.com first designed their warehouse network to deliver books ordered on the Internet, they knew that their competition was every bookstore in town as well as other mail order distributors. Consequently, they knew they had to have an extensive network that was close to their customers – the general population. They also knew they had to process orders very quickly and deliver customers orders inexpensively and conveniently. It was with a clear view of their competition that the general parameters of their network evolved.

STEP 2 – Understand Customer Service
Remember warehouses serve customers. You need to understand what this means. Do you need high availability levels? Do you have to have low lead-times…Both? How high…How low? Do you need warehouses close to customers? Why? What will happen if you have high service? Will you sell more? What will happen if a competitor provides better service than you do? Will you lose sales?

Most warehouse network designers cannot answer these questions clearly. Some don’t even ask them! Lack of understanding of the role service plays in your business is dangerous when designing warehouse network. One danger is you’ll over-design and spend more than necessary. The other is you’ll miss the ability to sell more or, worse yet, you’ll help a competitor sell to your customers.

By examining their customer service systems, Ralston Foods, a large food manufacturer headquartered in St. Louis, learned that they could dramatically improve service through improved systems. By having the ability to ship any product to any customer from any warehouse and having a smart way to split orders accordingly, they were able to avoid additions to their network that would otherwise be needed.

STEP 3 – Understand Costs
Warehouse network control four primary costs and several secondary costs. The four most important are:

• Transportation costs – inbound and outbound
• Inventory costs
• Warehouse facility costs
• Supervision and management costs

These costs are dependent on the network design and should be accounted for carefully. Transportation costs decrease as the network grows with more warehouses. This occurs because most inbound transportation is more efficient than outbound – truckload into the warehouse and less-than-truckload out, for example. More warehouses in the network increase the role of inbound freight and decrease that of outbound.

Inventory, on the other hand, increases as the network grows. This is due to the additional inventory needed to overcome out-of-stocks occurring in each warehouse. Facility costs cover the space necessary to keep inventory. They are similar to inventory costs and depend on whether or not inventory is concentrated in one warehouse or dispersed in several. Supervision and management costs are incurred at each warehouse regardless of its size or throughput.

To complicate the network designer’s task, these costs are inter-dependent. For example, low inventory, causing out-of-stocks, can be compensated for with additional transportation – emergency shipments or the use of more remote warehouses. The network designer needs to understand these costs, their dependency on the warehouse network design and on each other.

We designed a warehouse network for the Square D Corporation, a manufacturer of electrical products. While their transportation costs were significant, their inventory costs were too. Many networks are designed without a clear understanding that inventory increases when it’s stored in more places. As a matter of fact, it’s inventory costs that often provide a limit to the number of warehouse an organization should have. Square D went from five warehouses to three. Because of the inventory concentration, those three provide better inventory availability than did the five.

STEP 4 – Evaluate Alternatives
Now the network designer evaluates various alternatives – arrangement of warehouses. The evaluation calculates service and cost – the service established in Step 2 and the cost identified in Step 3. For example, given we have five warehouses in Philadelphia, Atlanta, Chicago, Dallas and Los Angeles, what’s the service level and what’s the cost.

In this Step it is very useful to perform evaluations on a computer – for two reasons. There are many alternatives to evaluate and merely identifying the alternatives requires a computer. For instance, if you wanted to test placing five warehouses in 100 possible cities, you’d have to make 75,287,520 evaluations – too many to do by hand, and you’d need a computer just to manage the process. (See USING A COMPUTER FOR THE EVALUATIONS)

You may want to determine and evaluate

• The best network(s)

Closest to customers
Highest service for a given cost
Lowest cost for a given service level

• The competitors’ networks

How close are they to their customers? …To your customers!
What are their service levels and costs?
Do their service levels justify their market shares?

• Options to your network

Different territories than you use now
Where warehouse should be added?
Should you consolidate…if so, which warehouses should be kept?
STEP 5 – Select an Alternative and Implement
In this Step you’ll select which alternative best meets your objectives and implements the course of action that improves your network. It’s time to use your judgement and consider all those issues you’ve been unable to quantify and formally evaluate. It’s time to involve others in organization with different responsibilities and perspectives.

Experience is very useful in this Step. Don’t be afraid to critique the evaluations. The computer calculations cannot be more accurate than the data that went into them. Don’t hesitate to override the computer results with your intuition and experience. Don’t mistake computer precision for business success – it’s better to be “approximately right than precisely wrong”.

When you decide on a direction – consolidation, better service, beat competitors – you have to gauge how fast to change from your current network to the new design. This will involve deciding which warehouse to close first, which to open next, how to inform customers, and so on.

STEP 6 – Count your Profits and Do it Again
Once you’ve implemented your new network its time to confirm the results of your decision. Calculate the actual service, costs and benefits of the new network. Confirm that what you planned really happened. If it didn’t, explain why not and make adjustments for the next time.

It’s also time to do it again. Things change – markets evolve, technology improves, and competitors catch on. Reevaluate the network again. You’ll be far faster and better at it the next time around.

Many leading companies continually reevaluate their warehouse networks. Some have entire logistics planning departments that do this work. Many mid-sized organizations, $1 million to $1 billion in sales, say, do not need this degree of attention, however. Normally an assessment of their networks every three to five years is adequate.

Sidebar to HOW TO DESIGN A WAREHOUSE NETWORK
USING A COMPUTER FOR THE EVALUATIONS
For all but the simplest evaluations a computer and the right software is necessary. A few tips in its use may save you mistakes made by others.

Don’t try to develop your own software. The techniques need to perform the evaluations can be very complex and there are lost of systems available off-the-shelf. If you don’t have software already, don’t buy the most advanced system. Try something simple, learn with it, then add more complexity and sophistication later when you have the need for it.

This approach applies to collecting data too. Don’t spend extensive resources on large data collection efforts at the outset. Rather start simply. Perform a few evaluations with approximations or even fictitious data. Then if the sensitivity of the results justifies it, invest added effort in collecting more accurate and extensive data.

There are two fundamental differences in systems that evaluate warehouse networks. Some are “cost calculators” and some are “optimizers”. Cost calculators merely report the results of an alternative that you input. Optimizers find the best alternative among many. These differences can be crucial. In the example of evaluating five-warehouse networks in 100 possible cities, you’d have to input over 75 million separate arrangements in a cost calculator. A good optimizer can find the best among the 75 million in seconds.

Another important difference is the flexibility of the system to represent different cost relationships – particularly transportation costs. Some systems are rigid in what they represent while others have far more flexibility or allow you to design any relationship you wish.

Finally, there’s the cost of these systems. You’ll find a wide variety. Some software costs tens of thousand while other are virtually free.

Whatever the cost, the benefits of good warehouse network design can be spectacular. An experienced network designer can provide a thousand percent return – well worth the investment.