Logistics Site Selection
30 Jun, 2003By: Dale Brubaker
A Weighty Decision
A recent headline in a national logistics magazine screamed the results of a study by a well known consulting group: “Bloomington, IN selected best warehouse location.” This was determined by calculating the population weighted center of the United States. In other words, in Bloomington, you have minimized the distance to every person in the United States if you had one distribution center. The author went on to show the “best” locations for warehouses if you have a network of one to ten warehouses, based on the weighted population center of the U.S.
What is wrong with this approach? This over-simplification of the site selection process grossly miscalculates the impact of transportation costs to the overall operation costs of a company. By only looking at minimizing the average weighted distance to the population, you make two grossly over simplified assumptions: one, that your logistics costs are only related to transportation; and two, that your transportation costs are directly correlated with distance. Site selection, done properly, takes into account many other factors that will definitely affect where your company decides to locate a warehouse or distribution center.
In this article we will show a more comprehensive model for site selection and discuss some of the do’s and don’ts of the process. We will discuss many cost factors that have to be evaluated relative to your company’s situation. We will also discuss some of the qualitative issues that are often overlooked in the site selection process and wind up costing companies a lot of time and money. Finally, we will show you how using an oversimplified approach can lead to significantly increased costs of operation that will leave your company at a competitive disadvantage.
The following six steps can be used to form the framework around site selection for a new facility:
1. Understand your customer base
2. Know where your company is going
3. Model the quantitative factors
4. Understand the qualitative factors
5. Build the business case and cost justification for moving forward
6. “Get Local” and evaluate specific sites
It cannot be overstated enough that the first step in a successful site selection process is to know your customer. The number, size and location of your distribution centers (DCs) are directly related to the service level your customers demand. Understanding how much additional business you can gain by exceeding them is paramount. For example, if you are supplying repair parts for aircrafts, your customer is not worried about the cost of transportation; they just want the part as quickly as possible. Therefore you need a single DC located next to the air hub of an overnight delivery service. However, if you are shipping to a national network of retail stores and are expected to deliver the product within five days of order receipt then you need multiple DCs that are located such that you minimize your cost structure while still meeting the customer’s expectations. The overarching rule of site selection is: “Thou shall not ever jeopardize customer satisfaction to minimize costs.” However, once you understand the customer requirements there are a lot of things you can do to minimize costs while still meeting their goals.
Another rule of site selection is to never design and build a distribution network for the past. Always design and build for the future. Before a company can do a good job with site selection they need to understand where the business is going. What is the best educated knowledge as to where the business will be in three, five or even ten years? Typically, when you locate a new distribution facility you will be there for a minimum of five years. Any mistakes made up front will cost you for a long time. Also, your company will be much better off if the design is done such that you will be in the location much longer than five years. This allows you to reduce operating costs with longer leases and eliminates the significant costs associated with moving an operation to a new location. You need to understand your company’s growth in total as well as the pattern. Many companies are not growing equally across the country. Many will have expansion plans for certain areas or will be growing faster in some areas of the country than others. These growth plans must be understood and incorporated into the site selection process.
Once you understand you customer and know where your company is going, you are ready to begin a distribution network modeling process. The modeling process allows you to look at the overall cost basis that will result from locating in different metropolitan areas. The modeling process itself is a rigorous, data intensive process that has its own steps for completing an accurate model:
1. Collect all operational data about your company over the last twelve months.
2. Summarize the operational data and make sure the company agrees on it’s accuracy.
3. Build a network model of your existing structure and validate it by modeling the last twelve months of operations and comparing results to actual.
4. Apply all growth expectations and/or expected customer service requirement changes to the model.
5. Model alternative distribution networks and compare costs.
One of the biggest challenges you will face in doing a site selection process is defending the location you have chosen at the end of the day. Going through a rigorous modeling process will practically eliminate this challenge as a factor. Models, however, are only as good as the data used to build them. One way to head off potential challenges to the end result is to get everyone who will be involved to sign off on the data used for the model as being accurate before the model is built. The second way to head off potential challenges is to make sure the model is validated. You must show that any network model being used to assist with site selection is accurate. The only way to do this is to model your company’s last twelve months of operation. You can model less than twelve months but a full year is best so that no one can challenge that you have missed issues related to seasonality of demand. Once your model is shown to accurately portray your company’s actual costs of operation for the last twelve months (with models in use today, accuracy should be within one percent) you are ready to adjust the model for future operations. This is where good strategic planning up-front is important. The modeled demand must be updated to reflect anticipated growth patterns (possibly by product line) as well as any regional shifts expected across the network.
What costs should you include in your model? Today’s modeling tools are fairly sophisticated and can model most any cost of operation specific to your company. You want to capture all costs incurred to get your product from your suppliers through your distribution network and finally to the end customer. Cost factors typically include such factors as labor, operational (including utilities and taxes), facilities (lease or amortized build cost) and transportation. Transportation, while typically the largest part of these cost factors is not necessarily correlated directly with distance. Depending on your shipment profiles, your transportation costs may be more zone-based and not directly related to distance. In these cases, the labor and operational costs become deciding factors within certain regions. After all this, you are finally ready to evaluate alternative sites and use the model as a tool to estimate the cost basis of each alternative metropolitan area.
But it is not as easy as determining the metropolitan area with the lowest costs. Every company says they want to minimize costs but I can assure you that every company also has other, non-cost related factors that will affect the location chosen. Oftentimes these factors do not drive a company to a certain location but they will be used to veto locations from consideration. Non-cost factors are more difficult to deal with but are extremely important to identify. Prior to doing any network modeling study, every company should be facilitated through a strategic visioning session to determine what non-cost factors will be specific to their company, what the company hopes to achieve with these factors and the relative weighting in the decision making process. Typical non-cost factors include quality of life (you need to determine what this means to your company for your employees), employee educational level, ability to find readily employable workers (unemployment level), presence of union organizing activities (does the state have right to work laws), presence of a readily available part-time work force (often college age students, mothers with children in school, or military spouses), and access to rail or air services. Without going through this step of identifying non-cost factors, you take the real risk of choosing a location that you will never get consensus about within your company.
So how do you determine what metropolitan areas to consider for your distribution centers? First, determine where the weighted average distance to your customer base is located. This is similar to the simplified weighted average distance to the population but it is only the starting point. If you have current locations that will not change then they can be fixed in the model. When these location(s) are determined, look around a 500-mile radius and evaluate all metropolitan areas against your non-cost factors. (There are methods that can be used to try and quantify these factors in relation to each other and rank order the metropolitan areas.) At this point, determine those metropolitan areas where you would allow your company to locate. These are the locations you want to model. In this way you make sure that you don’t waste a lot of time and money modeling locations or investigating real estate in metropolitan areas where your company will not allow you to locate. Once you’ve agreed on the areas that meet your non-cost factors you can begin modeling the relevant costs for your distribution network knowing that the end result will be acceptable. Each metropolitan area will have its own cost structure for labor, operations, and facilities. A good site selection company can provide these factors for you or you can try to get them from government supplied data. The overall low cost alternative that meets the qualitative requirements is then the location to begin looking at for a site. At this point, you probably need to engage a good site selection company that can put “feet on the ground” and evaluate the area on a county by county basis and also evaluate alternative sites (if building) or existing buildings (if looking to lease).
So what about the Bloomington example? In a recently completed study by another national consulting firm, the inadequacies of choosing locations SOLELY upon the advice of minimizing distance to population were exposed. This firm performed the same study that showed the minimized distance to the population. However, they then performed a study using the same analysis, but included costs for labor, facilities, LTL, and small parcel shipping. When comparing the results of the two models, they found that by including the above costs, the better location is in Western North Carolina along Interstate 40 where labor and facilities are less expensive. In fact, despite being 83 miles further away from the population, the distribution center operation costs were 23% lower than Bloomington. This is a significant number when you apply it to your total logistics budget. Those companies who used the simple approach and did not do a thorough study of their network could find themselves paying over 20% more to get their goods to market and could suffer competitive disadvantages. The chart below shows the results of a study comparing the choice of metropolitan areas for distribution centers based on simple distance to a costed model that takes into account labor, lease costs, and transportation costs.