Supply chain optimization is seeing unprecedented adoption. Are you behind the eight-ball?
The current state of supply chain optimization (SCO) in manufacturing is well illustrated by the old story of a driver and his assistant delivering a truckload of goods. “Where are we going?” says the assistant. “I don’t know,” replies the driver, “but we’re making good time.”
Today, most manufacturers are faced with daunting, complex supply chains that are compounded by product proliferation, channel differentiations and the geographical stretch that further lengthens supply chains. Most manufacturers have realized the need for sophisticated tools to model these complexities and offer solutions based on optimization, however, they are not typically sure how to reach this end goal.
Certain traditional problems faced by manufacturers, such as capacity and material problems as well as transportation issues, often render themselves very well to linear modeling and there are quite a few existing solutions in this area to help manufacturers. However, the increasing complexity of supply chains as described above has given rise to a new class of problems, such as risk management, multi-echelon cost optimization and managing variability that cannot be addressed by simple, linear modeling solutions and instead need more sophisticated SCO solutions.
Fortunately, there is a new generation of SCO solutions to address these issues. These solutions are based on new non-deterministic, non-linear algorithms — most of which have emerged from academia over the last 20 years. Industry leaders who are adopting these new solutions are already reaping significant benefits.
However, despite the clear benefits these solutions impart and the momentum provided by “early adopters,” other manufacturers have been slow to adopt these cutting-edge SCO techniques. There are three major issues that create barriers to adoption:
1. Organizational dynamics. SCO solutions invariably optimize across multiple departments and multiple nodes of the supply chain. However, most departments are run according to their local metrics and therefore attempt to achieve “local optima.” The problem gets compounded if one of the departments is a completely separate entity, such as production outsourcing. Factories still tend to be measured on capacity utilization and productivity rather than if the right amount of the right item is being made. Sooner rather than later, specific job functions will emerge that will focus on these issues and push adoption of new techniques; similar to how the emergence of the Vice President of Supply Chain function in the last 5-to-6 years has driven faster adoption of Advanced Planning Solutions.
2. The rocket science myth. Many manufacturing managers look at SCO techniques as “rocket science.” They think it is hard to understand, is of little relevance to how things work on the manufacturing floor and is an academic creation not grounded in reality. In other words, “By the Ph.D., for the Ph.D., of the Ph.D!” Many leading-edge companies such as HP, P&G, Black and Decker, and Microsoft have adopted SCO solutions aggressively and have proven beyond doubt that the ‘rocket science myth’ perception is highly inaccurate, based on the results being achieved.3. Complex change management. SCO solutions can be "a very powerful flashlight if you know where to shine it.” Unfortunately this flashlight can often expose the pointlessness of existing practices and procedures, which in turn, creates a reasonable amount of opposition from non-believers. As an example, a high tech manufacturers discovered that a better manufacturing strategy would involve moving production from one location to another. However, there were several different levels of change management in different organizations required for this shift that resulted in an adoption barrier.
To reap the benefits of SCO, manufacturers must start breaking down these barriers. Companies should build upon their experience with existing tools to move up to state-of-the-art SCO technology.
If you are considering implementing complex SCO tools, first consider the following “checklist:”• Create an awareness first
• Try before you buy
• Open up the “black box”
• Understand the importance of user intervention
• Compare best practices with similar industry leaders Effective change management begins with creating awareness and building internal champions. Encourage your teams to use SCO diagnostics tools to measure how current processes are performing. Bring in experts/practitioners from outside to explain current conditions with early adopters. They can answer a lot of initial questions that can remove uncertainty and doubt from the team.
Nothing gives you and your team more confidence than a “try before you buy” model. For example, take a small data set, run it through the solution you are thinking of implementing and compare the results with your current information. It is a simple exercise that can be a real eye-opener — one that is easily shared with the entire supply chain management team.
One of the challenges of optimization tools is that you cannot easily take a pen and paper and quickly decide your expected result. Your success in having the tool adopted is a big function of your ability to break through the black box stigma. The tool will provide you with the needed information to explain why the results are what they are.
Coupled with the initial disbelief of the results as described above comes the reality that no optimization tool could have possibly modeled everything that a user had in his or her mind. Thus the tool should be able to take user inputs and overrides, and then come back with a new solution. Without this ability to create ‘what if’ situations, almost all optimization is useless to the user.
Finally, collect best practices from similar companies and compare notes on the results — many of which will relate directly to your own complex business problems. Solution providers, consultants and analysts should be able to provide you pointers. You will likely find that most large companies are already reaping benefits from these.
Rewards from risk management
Non-linear optimization tools are particularly good at solving complicated risk management problems. There are multiple factors in risk management, any of which can happen at any time with different levels of severity. In the post-Hurricane Katrina environment, there has been a renewed sense of urgency to be ready for any kind of risk.
The most useful optimization algorithms focus on managing risks to customer service, manufacturing operations, and individual plants. The biggest risk to customer service is managing demand and supply, and that is precisely where non-deterministic, non-linear optimization techniques are successfully solving key problems.
$130 million savings
For example, one of our global multi-billion dollar clients — which produces hardware and software — is using non-linear SCO algorithms to deal with supply chain design optimization and risk management. The client had to guarantee certain customer service levels for a new product, a challenge compounded by complex supply chain issues. By implementing non-linear SCO techniques, the customer was able to determine when to terminate old products as well as when to introduce new products without cannibalizing the existing product line.
These techniques helped the client identify optimum inventory lev