Improving supply chain performance is critical to businesses because of the strong link between supply chain effectiveness and profitability. For example, there is a need to make best use of limited resources, target new investments wisely and deliver the business strategy cost-effectively whilst minimizing business risk. However, this is extremely difficult to do well because supply chains are complex and the business environment is uncertain in nature. Because traditional decision-making takes into account neither the full complexity of the situation nor the many inter-related factors, it is often too simplistic and static to make the right supply chain decisions, with the result that expensive mistakes are often made.
Many mistakes can be avoided and multi-million dollar benefits realized through the use of the right decision support and modeling methods. Modeling that incorporates supply chain complexity can create innovative options and identify much better solutions than when only one part of the chain is considered; the modeling approach can be tailored to meet the specific business situation; and the principle of ‘fitness for purpose’ ensures the delivery of business benefit.
Modeling can incorporate supply chain complexity in order to create innovative options and better solutions. Working in silos is easier to manage, but risks mistakes being made. For example, producing locally ‘optimal’ solutions that do not work when put together because they ignore supply chain dynamics and interactions or perhaps, missing the opportunity to have a better overall result by being deliberately less efficient in one area for the ‘greater good’ (eg minor amendments to production schedules).
Complexity and the benefits
By contrast, modeling that takes account of all relevant elements of the supply chain is much more complex, but the benefits can be enormous, for example:
The first ever end-to-end scheduling model of a car plant resulted in Nissan reducing work-in-progress stock holding by over 40 per cent (including engines awaiting installation being reduced from 250 to 20) because for the first time the schedule was fully achievable.
Also, a global media organisation was able to identify benefits of several million dollars per year through making more effective use of program-making staff and associated facilities, together with better planning of program making activity.
And a major shipping company saved several hundred million Euros by adopting a global optimisation model – the previous method that regarded each ship as a profit center performed far worse for the company as a whole because it resulted in very high numbers of empty container movements.
In all of these examples the complexity involved meant delivering the benefit would not have been possible without the right decision support approach.
There are three steps in ensuring that the chosen solution delivers the required benefits.
Ensuring fitness for purpose: The first step is to define the decisions the model will be used to help influence, for example:
Understanding the decisions that are to be made on the basis of the model; determining which parts of the business to represent, how and at what detail; defining the parameters, rules and data to be included; describing the users and their needs; and identifying what would constitute ‘success’.
Producing the solution: The second stage is the technical design to meet the business requirements, covering: choosing and applying the mathematical techniques (eg simulation, optimisation) that are most appropriate for the problem; initial development of the modeling solution; data collection, cleaning and validation; and reviewing the model and refining as necessary.
Delivering the value: The final stage is then system and user testing against pre-agreed criteria, including: Initial calibration of the parameters set to reflect reality; installation and final testing on user PCs; and training, both in basic use of the model and in how to get the best from it.
The vast majority of these models are then handed over for ongoing use by people who,in general, are not modeling or systems specialists. Consequently, it is very important that any such tools are flexible by being parameter-driven, robust and above all easy to use.
There are many different modeling approaches and techniques that can be used. Similarly, the solution is likely to differ whether the focus of attention relates to strategic, tactical or operational areas. Benefit delivery is best assured by ensuring the solution is tailored to the specific problem. The principle of ‘fitness for purpose’ ensures the delivery of business benefit.
The figure below [see PDF file of original article by clicking this link ] illustrates examples of the different areas of interest and focus. The central ‘arrow’ relates to time, running from strategic planning potentially years ahead, right down to the day of operation or even beyond.
The six boxes at the bottom of the figure indicate the types of question being asked at each point.
Towards the left of the figure, decisions involving an entire operation or major element of a business are being taken, whereas towards the right, the focus is on an increasingly small part of the business right down to the activity of a particular individual or piece of equipment. In general, an increasing level of detail in the data is needed as the time moves towards the left.
The three boxes at the top of the figure give examples of different types of objective for the modeling exercise.
The following case studies give practical examples of benefits delivered at the different points on the time horizon and for different objectives.
Strategic supply chain planning for a major healthcare products provider
Due to complex strategies and copious amounts of information involved, this healthcare company had been unable to develop a reliable and consistent methodology for calculating the long-term economic impacts. A supply chain relationship model was created which allowed the company to evaluate the economic impacts of complex supply chain reorganisations in a matter of minutes.
Benefits realised included an increase in annual profit by 50 million Euros. The client is now able to assess economic impacts and the model provides a central consistent repository for company information. They continue to use the model to evaluate other economic drivers fundamental to developing their long-term business strategy.
Tactical planning for a food manufacturer
A major food company was planning for significant growth in its US peanut butter operation. Greater recipe variety and new products, more innovative packaging options and regional variations in products, plus further expansion into export markets are now driving growth. This resulted in a 14 per cent sales volume increase on the current product portfolio with an additional five per cent per year expected. To cope with the higher demand, the company increased peanut roasting capacity by 50 per cent. But it was unclear whether this would deliver the required throughput or just move the bottleneck elsewhere in the plant.
A capacity planning model simulating the plant was created to examine how it would cope with numerous different demand possibilities. This allowed the company to see when capacity at each part of the plant would be exceeded and examine options for removing bottlenecks that could emerge. The model replicates the layout of the plant. Production ‘flows’ through the model and bottlenecks can be seen, and the effect of capacity/layout changes or different demand assumptions can easily be tested.
Tactical and capacity planning for a global media organisation
This company is one of the largest programmaking organisations in the world. It needed to plan its program-making more effectively and lower its costs by making much more effective use of people and facilities (eg studios and cameras). However, program-making is highly complex because of the many operational constraints and different types of program, their production staff and facilities involved.
A production planning model that seeks to make best use of the available staff and facilities was devised that has shown realisable benefits of several million Euros per year. The company can now also balance the trade-offs between any areas of future excess and under-utilisation to make production as cost-effective as possible.
Advanced production scheduling for Nissan
Having committed to building the Almera in its UK plant alongside the Primera and Micra, Nissan discovered that the large number of constraints made scheduling impossible using their existing techniques.
PA delivered a revolutionary solution that successfully scheduled three models down two plant lines while adhering to all production constraints. The system, believed to be the first of its type in the world,achieved an extra 30 per cent in production capacity with no appreciable increase in plant. Scheduling accuracy rose from three to 90 per cent, allowing Nissan to reduce stock holding by 40 per cent and deliver benefits of up to 3 million Euros per week.
Resource allocation and reacting at the Port of Felixstowe
The Port of Felixstowe is the largest container port in the UK. It was losing business to its competitors as it was unable to increase the rate at which it processed ships. There was a need to make the equipment assets operate closer to their maximum efficiency otherwise they would have to embark on a multi-million euro expansion programme.
A real-time scheduling model was designed and developed which dynamically controlled the movement of all vehicles around the port, and optimised the loading and unloading of all containers. The model is capable of rescheduling the entire port every 10 seconds. The model was also able to demonstrate that additional investment was unnecessary whilst simultaneously demonstrating that operational efficiencies were possible and achievable. The port was able to increase the number of containers handled annually from two million to four million, without any increase in equipment or manpower. As a side effect the number of ‘lost’ containers was reduced to zero.
Please click here to see the article as published (PDF file)
