On Day 2 we talked about strategic sourcing decision optimization, the technology you need to make the right buy given the myriad of constraints you have to adhere to and the large number of costs and bids you need to take into account. Today we’re going to talk about supply chain optimization – the process of optimizing your supply chain, or distribution network, to minimize costs and maximize value.
Even though the only way to truly get the optimal buy every time is to use the optimal supply chain, the reality is that you can’t transform your supply chain overnight for every bid. The realities are that it takes time to acquire, lease, or dispose of distribution centers and warehouses, that you have contracts in place with suppliers and carriers for anywhere from three months to three years in a typical organization, and that changing global distribution patterns requires time to research the regulatory, documentation, and taxation requirements of different countries and trade zones. Thus, when it comes to strategic sourcing, the best you can do is optimize the buy within the supply chain you have available to you today. However, if you can improve the supply chain, then you can reduce your costs and save even more across all of your buys.
Supply chain optimization is something you should do on a regular basis. Whereas in the past, experts would say that it is something you should do only once every five, seven, or ten years – today it is something you should do every year! Today’s optimization solutions are a lot more powerful than they were ten years ago and allow you to build much more sophisticated models, which are now usually solved in hours compared to the weeks that was once required for models of this magnitude.
Even though it probably doesn’t make sense to buy and sell manufacturing and distribution center assets every year, there’s nothing stopping you from modeling the cost associated with such a sale, or modeling the cost of breaking or failing to renew a lease, of each asset you have if new options present themselves, such as alternative low-cost distribution centers or the possibility to sell a manufacturing center to an outsourced contract manufacturer who might be able to manage it more cost effectively. Today’s solutions can model all of the costs associated with acquiring, running, and disposing of an asset in your global distribution network, and can help you truly identify what the optimal network is for you at any given time for any given period of operation. (Thus, every year you can redo the analysis and assume that the network is only going to remain stable for the next year.) You can also tell a good supply chain network optimization solution that certain aspects of the network aren’t allowed to change and that certain aspects of the network must change and have it tell you whether or not your current network is optimal or if you should consider making some changes.
So how do you identify the right supply chain modeling and optimization solution? As with any other technology, you ask the right questions. The following questions should be enough to get you started and help you identify the real solutions with the power you need from the imposters.
1. Can the solution model your supply chain as is?
This is a question you need a resounding yes to. How do you know how much a potential network redesign is going to save you if you don’t even know how much your current network design is costing you? This brings us to …
2. Can it derive a cost baseline?
Once you’ve modeled your current network, the solution should be able to run the model and tell you how much your network should be costing you. (If your current network is actually costing you significantly more, than either you have some inefficiencies in your processes to work out or you have not accurately modeled your network and need to revise or expand your model.)
3. Can the solution support the construction of a model depicting a desired state?
If you have a solution in mind, you should be able to construct that solution and derive a cost baseline for that solution. Similarly, you should be able to define your own modification of a suggested network design and derive a cost baseline for that modification. After all, it’s not the lowest cost solution, it’s the highest value solution – and that’s not necessarily the solution with the lowest cost today, but the network design with the expected lowest cost, and highest value, over the expected lifetime of the network.
4. Can it derive an estimated cost of any model you specify under a projected range of activity?
The reality is that any given solution is only optimal for the specific (set of) demand value(s) and the specific (set of) cost(s) that the model is defined on. However, you’re optimizing your network for a future period of time, where demands are only forecasts that could change. Thus, you want a solution that also has simulation capabilities and the ability to run multiple models under multiple demand scenarios and cost differentials to allow you to come up with a network plan that is robust and most likely to save you money over the range of scenarios that are most likely to occur.
5. Can it allow you to drill down into the expected cost differential between two models and determine why?
It’s not enough to know that one network design is expected to cost 2M more than another, you also need to know why, especially if the more expensive network design is the one you’d prefer. If you know that most of the costs are associated with lease payments, then you know that if you could negotiate a lower lease price, you could end up with a network design that you like and that is only slightly more than the lowest cost solution. If such a design also has lower risks, then it has a higher value and you can choose it.
6. Can it help you optimize your supply chain improvement investments?
Converting from one network design to another will occur a lot of upfront costs associated with asset acquisition, lease, and disposal as well as penalties if you have contracts in place that you need to back out of early. These up front costs need to be covered somehow, and if you only have a fixed amount of capital available for supply chain improvements, you want the model to be able to take that into account and the solution to provide you with different, near-optimal, improvement possibilities that are within your budget today.
7. Can it model the impact of fixed asset disposal or cost reduction on projected service levels? inventories? greening?
When optimizing your network, it’s not just about cost and risk, it’s also about service optimization, inventory optimization, supply chain greening, and a slew of other initiatives. It’s important that such a solution not only allow you to specify all of your constraints, but allow you to calculate whether or not you’re trading service level or inventory risk or carbon credits for that cost reduction.
8. Can the solution support sensitivity analysis?
Building on the last question, if the system tells you a certain network design is likely to reduce your projected service levels by 1%, you want to know how much money is required to bring that down to any threshold between 0 and 1%. Maybe you only have to sacrifice 25% of your maximum savings opportunity to achieve a service level decrease of only 0.1%. That could be a good trade-off – a 0.1% decrease in projected service levels is much better than a 1% service level decrease, especially when it costs you only 25% of your maximum savings potential to achieve a projected service decrease that’s ten times better than the projected service decrease that you would be stuck with if you went with the greedy solution.