right_division Green SCM Distribution
Bookmark us
SCDigest Logo

SCDigest Expert Insight: Supply Chain by Design

About the Author

Dr. Michael Watson, one of the industry’s foremost experts on supply chain network design and advanced analytics, is a columnist and subject matter expert (SME) for Supply Chain Digest.

Dr. Watson, of Northwestern University, was the lead author of the just released book Supply Chain Network Design, co-authored with Sara Lewis, Peter Cacioppi, and Jay Jayaraman, all of IBM. (See Supply Chain Network Design – the Book.)

Prior to his current role at Northwestern, Watson was a key manager in IBM's network optimization group. In addition to his roles at IBM and now at Northwestern, Watson is director of The Optimization and Analytics Group.

By Dr. Michael Watson

September 11, 2014

Just Because the Feature Exists, Doesn’t Mean You Should Use It

Putting too Much Into a Model can be Worse Than Having too Little

Dr. Watson Says:

...and these problems may cause you to get no results from the model...
What Do You Say?

Click Here to Send Us Your Comments
Click Here to See Reader Feedback


Over the years, I’ve noticed something interesting that happens to many people when they first use a commercial network design tool for supply chain modeling:  there is a strong tendency to want to use almost every feature in the tool.  Or, said another way, they become very afraid of not modeling some aspect of their supply chain.

What makes this fact interesting is that their previous alternative to using a commercial tool was a spreadsheet (or worse).  And, typically, this spreadsheet captured almost no aspect of the business and likely wasn’t even internally logical.  For example, the total demand in the spreadsheet might not even match the total supply, the spreadsheet may completely miss the fact that customers are in different geographical locations, and the transportation costs may not change if the warehouse location changes.

Then, once this person has a commercial tool, with all the features that entails, he wants to make sure that every aspect of the supply chain is captured in the model.  He wants to model every nuance in the manufacture of the product, every business rule in transportation, and every exception in customer deliveries.  He never quite seems satisfied with making an assumption and the tool gives him the false comfort that he doesn’t have to. 

So, what is the problem with modeling every nuance and rule?  Isn’t this a better approach?  There are two main problems with this—and these problems may cause you to get no results from the model.

Previous Columns by Dr. Watson

The Three Use Cases for Data Scientists

Learn Python, PuLP, Jupyter Notebooks, and Network Design

EOQ Model and the Hidden Costs of Fixed Costs

CSCMP Edge - Nike Quote: "It is All an Art Project Until you Get it on Someone's Feet"

Supply Chain by Design: Why Business Leaders should think of AI as an Umbrella Term


The first problem is that that 80/20 rule applies to supply chain modeling.  A few key parameters (the 20%) will drive the results of the model (the 80%).  So, if you spend time modeling everything, you are going to be wasting a lot of time on factors that don’t matter.  By collecting data that doesn’t matter and spending time putting it in the model, you risk getting nothing delivered within the timeline and budget.

Second, by adding complexity to the model, you are going to make the model very difficult to understand and difficult to change.  That is, the results returned by the model may not make sense and it may be difficult to determine if the results are actually valid.   And, inevitably, you will want to test different ideas and scenarios.  If the model is too complex, it may be impossible to make adjustments without breaking the entire model.  If the model is too complex, you again may not get any results.

In general, your supply chain model should include only a handful of different parameters and factors.  You want to keep your model as simple as possible.  I’ve found that it is better to start with a very simple model and slowly add complexity rather than think about a complex model and take things away.  When you do the former, it forces you to add in the things that matter the most.  When you do the latter, people tend to get nervous that they are removing something important.

Final Thoughts:

Creating simple models is often easier said than done.  We devoted a good part of our Supply Chain Network Design book to this topic.  In the early chapters, we show how you can slowly add features to a model and then stop when the answers are good enough.  In the later chapters, we talk about the Art of Modeling and how to think about aggregating data.  We felt that this could help people build better models.

Recent Feedback

This is reminiscient of "In defense of model simplicity" by Laura Mclay

Distinguished Engineer
Sep, 21 2014