Wow, how does that old saying go again -- the one about being "careful what you wish for"?
This is what I have been thinking lately on the topic of supply chain network optimization. I did my first network optimization project 32 years ago when American Motors (AMC) asked us to model their aftermarket supply chain design and network. I really saw the value in determining the numbers and the locations of distribution facilities by evaluating the trade-offs between transportation costs, inventory costs and distribution center costs.
As such, I worked very hard to get companies to take a more rigorous view of their network design, but I was often rejected. I really wished that companies would take a more methodical approach to their network optimization goals.
Well, this is where I should have been more careful, because this has come true, and now I have mixed feelings. Why? Because I see many companies jumping on the network optimization bandwagon without really having their acts together.
Here are the five most significant problems I see with supply chain network optimization today:
1.Project vs. Process: Continuing the above AMC story, over the next 25 years, this aftermarket "project" was repeated many times as AMC merged automotive supply chains with Jeep, then Renault, then Chrysler, then Daimler and then away from Daimler. I then learned that supply chain network optimization is not a project, but a process. I strongly believe that supply chain network optimization needs to be an ongoing process as a company evolves. As sourcing is changed, M&A takes place, markets are changed, transportation cost are changed, etc., the network model should be rerun to fully comprehend the evolution of the network over time.
2.Strategy Before Network: We see companies pursuing network optimization without being able to articulate the strategy of the firm going forward. Please! The strategy is the starting place, and it is from this strategy that the network will evolve. If there is no strategy, then a valid network optimization process cannot be pursued.
3.Drowning in Data: The starting point for all network optimization processes is an understanding of the requirements of what flows occur in your supply chain. These flows need to be a composite of what has happened in the past, adjusted for the strategy going forward. Point #2 addressed the necessity of understanding the strategy, so no need to repeat that here. The point is: do we really understand what has happened in the past, and what should have happened in the past?
This is more then data manipulation; it is the true understanding of the historical requirements of your supply chain network. In addition, companies need to forecast what will happen in the future with supply chain demand. So, it is important to look behind and ahead of you.
4.Costs Are Critical: I have seen network modeling take place with: incorrect assumptions on transportation costs (they are not linear with distance); with distribution center costs (have you looked at the best manual operation, mechanized operation and automated operations?); inventory levels across the network (have we looked at inventory stratification, hub and spoke, etc?); and lastly, with inventory carrying costs (which can vary very significantly across the network). And let's not forget an often neglected part of supply chain costs -- inbound often starts at the ports-of-entry with inventory and service time costs incurred there.
5.Common Sense Not Common: I have seen some silly RFPs come across my desk in the last few months. All of these were requests for a supply chain network optimization "project," when in fact there was no real need to model the network. What was needed was to answer some basic questions with respect to adding a China Hub, adding a West Coast deconsolidation center, adding one or two LSPs to augment the current network, adding additional pool points to a transportation network, and so on.
In other words, what is needed is common sense. Are the ideal savings that could come from an "optimal" network sufficient to pay for leaving behind millions of dollars of investment in the current network, sufficient to pay for the costs of employee displacement, sufficient to pay for the fixed costs of owned facilities? Too often, I see companies drastically simplify network alternatives to the point of having only a handful left to consider.
Furthermore, there seems to be a total lack of understanding as to what modeling can do in the context of network optimization. Business leaders often think that they can gather some data and throw it into a model and the magic "optimal solution" falls out. Not so -- it requires a lot of work outside the model, along with an iterative process that builds on what is learned from each model run.
The good news is that companies are thinking about their supply chain network. The bad news they are making the above five mistakes, and so the results of these efforts are often of little value.
My wish for a greater focus on supply chain network optimization has been granted. But let's not throw the benefits of this work down the drain by failing to ask the right questions and consider the best alternatives.
What has your company been doing lately with networking optimization? Is it working or is it just wishful thinking?
Jim
Tompkins Associates
More Resources
Case Study: Distribution Network Analysis for Prestolite Electric
White Paper: Designing a Distribution Network to Address Today's Challenges
Distribution Network Design Services