Here are a few definitions of the word “Optimize” without any association or consideration for supply chain optimization or supply chain planning:
- To make as perfect or effective as possible
- Make the best or most effective use of (a situation, opportunity, or resource)
- To find the best compromise among several often conflicting requirements
All of these definitions still make sense when you associate them with supply chain planning and optimization. However, does one need a PhD in Mathematics or Supply Chain Operations to make sense of an optimized supply plan? There is no right answer to this question but I would argue that you do not need an advanced degree to make sense of optimized results if you can clearly state your definition of an optimized plan. I contend that this is always the bigger challenge and is an essential prerequisite before you consider going the optimizer route.
The proper definition of this optimized plan requires a macro level understanding of supply chain objectives, along with overall business goals of the company. It also means that you need a senior leader of the supply chain organization involved in the definition of an optimized plan and the associated cost model. Some of you may have heard of the phrase that goes something like – a camel is a horse designed by a committee. It is claiming that the horse is superior in some way but the thought is that several compromises are made if you have too many people involved in the design of anything. You could potentially apply this to the definition of an optimized plan too in that it requires a top down approach. The right resources will not only understand the bigger picture but should also be able to translate that into an optimum that is achievable and understandable.
One of the first things you need to do is define some boundaries around what needs to be optimized. Here are some of the questions that can be asked as a part of this exercise:
- Do we need to optimize the entire supply chain or can it just be a part of the network?
- Should we optimize the distribution network independently of the manufacturing sites?
- What are the major costs within your supply chain that have an opportunity for reduction?
- What is the product mix that will maximize profits?
- What kind of time horizon and buckets do I need to consider for optimization?
- Do I need to consider lot sizes for the entire planning horizon?
- Does it make more sense to optimize at an aggregate product group level? And, if this is feasible, can you define disaggregation logic that will be acceptable?
- Do I need to consider priorities along with cost as a part of optimization?
Reconciling Conflicting Objectives
The need for optimization arises from the fact that conflicting objectives exist and need to be resolved or balanced. Most companies have the objective of providing high customer service levels along with the goal of minimizing inventory carrying costs. This is just one of several competing objectives generally. The scenarios are too complex for a planner to determine manually or with heuristics and require automated mathematical algorithms. This complexity and the need to automate is exactly the reason why it is also difficult to comprehend the results. However, if there is an understanding of the variables, constraints, relative costs and the associated master data, the results are fairly consistent and as accurate as the mathematics behind the algorithms.
If you are still reading this, I sense a genuine interest in complex topics. You do have to reconcile to the fact that a lot of time is needed to interpret and evaluate optimization results, at least initially, as the parameters are fine-tuned. It definitely becomes easier and faster as planners and architects familiarize themselves with how all the parameters, variables and constraints influence the results. There is no such thing as optimization nirvana but one can attempt to get close to it by putting in a significant amount of time digging into details of the optimizer logs. SAP does provide explanation logs to help interpret and summarize results. However, it would help to have additional visual graphical tools that summarize the results. Some companies have either built these tools in-house or have exported to excel to evaluate the results.
I will leave you with this thought. I’m not sure who can claim this as their own, but I stole it from an SAP Optimization presentation by Heinrich Braun and it does a good job of explaining the approaches you can take to defining and understanding the world of optimization:
Idealist – Searching for the global optimum
Realist – Improving the first feasible solution until time out
Sisyphus – Knows that the problem has changed during run time
Pragmatic – Solves unsolvable problems
For an in-depth look at the optimizer, watch our on-demand webinar recording: SNP Optimizer: How to Approach, Manage and Maximize the Benefits, presented by myself, Claudio Gonzalez and Mike Raftery.
You can also check out Mike Raftery’s recent blog: Why Bother With an Optimizer.