6 Modeling Tips from LLamasoft and JLL

Michael Watson Ph.D Partner
Read Time: 3 minutes apprx.
data cleansing data processing network design supply chain

Man interacting with virtual world map

 

(This article first appeared in SC Digest.)

At the same CSCMP talk as Benjamin Moore (covered here), Jason Brewer of LLamasoft (a provider of supply chain design software) and Kelly Gray of JLL (or also known as Jones Lang LaSalle, a full services commercial real estate firm) opened up the discussion with some of their unique supply chain design insights.

I know that CSCMP likes to focus on the educational content and tries to minimize vendor and service provider talks. But, when done right and done in the spirit of education, vendors and service providers can provide a unique perspective and have the advantage of deeply knowing their subject area. I think Mr. Brewer and Ms. Gray did a great job.

My three big take-aways from Mr. Brewer’s talk were:

1.   No one has clean data. All companies struggle with getting the data you need for a supply chain design strategy. LLamasoft has seen many companies struggle with this issue. This is also my experience and that of the consultants and companies I talk to. You can take comfort knowing that you are not alone. But, this shouldn’t stop you either. You should plan on having bad data and needing to do data clean-up. And, you should know there are proven strategies for handling this. Brewer pointed out two important ones: you should be willing to make assumptions and you should be willing to use the 80/20 rule to focus on getting and cleaning the 20% of the data that will drive 80% of the value.

2.  There is a big push to do modeling faster. Given the fact that data is not clean and that these projects take 3-4 months (see point #1 from the Benjamin Moore talk), vendors, consultants, and end companies are working hard to make this process much faster. Some of this work includes using older technologies—like staging the data in better data warehouses. And, some of this work includes using new technologies—using new tools for data munging and data blending (we’ll cover this more in future SC Digest articles) or using web interfaces to allow multiple people in the organization to own and maintain their piece of data (like having the transportation team maintain the accurate rates).

3. Risk has an upside too. Often, we hear supply chain risk discussed in the context of what to do when something goes wrong. Brewer pointed out that when doing supply chain design, you also need to worry about what happens if things go much better than expected. What happens if your new product is 10X more popular than you expected? Is your design flexible enough to allow you to scale and capture the potential upside?

My three big take-aways from Ms. Gray’s talk were:

4.  The big US Metro Areas are getting bigger. This is an interesting mega-trend that suggests that as you build your facilities, you should make sure you account for the fact that more and more of your demand will be concentrated fewer metro areas.

5.  Real Estate incentives shouldn’t drive network design. Many local and state governments offer generous incentives to locate in their area. The JLL team advised against using the incentives as the primary reason for picking a location. Instead, determine several alternatives and then let the incentives break the tie. This reminded me of what Patrick Haex of Buck Consultants International once said at a talk on taxes in Europe: Don’t let the taxes in Europe determine your design, instead determine a good set of alternatives that are good for the business and then figure out how to best manage the taxes. In both cases, the idea is the same—the design of your supply chain drives a lot of value, don’t minimize this value by chasing savings in just one aspect of the overall design.

6.  After the network design study, there is a lot of work in the selection of the site. This point is a good reminder that the mathematical modeling of the supply chain gets you down to the general area where you want a facility. After the modeling exercise, it is then important to determine the correct site based on local zoning rules, local taxes, the local workforce, access to infrastructure, and incentives. Gray brought up an interesting point—if you are locating an e-commerce site, you don’t necessarily want to locate close to an Amazon facility—they may have already soaked up the local labor market.