Monkeying Around with the Supply Chain: The Chaos Monkey Strikes!

Do you have a supply chain risk management process in place? Have you evaluated the possible risks and built contingencies? How fast can you recover if a key component is no longer available? Maybe you have contingency plans in place, but how effective are they? Someone forwarded this article to me from Forbes.  It outlines an interesting approach to supply chain risk management.  It’s based on a process used at Netflix where they test their distributed systems by killing (temporarily) various sub-systems to ensure that the other sub-systems can compensate for the missing system.  They call this process the “Chaos Monkey,” and the Forbes article suggests that a form of this could be applied to supply chain. The article suggests having someone go down to the receiving dock or stores, pick a part at random, then work through an analysis to determine what would happen if the primary vendor couldn’t ship this part for three weeks. I like this idea, but I would like to see it used in conjunction with a more pragmatic risk analysis program that prioritizes risk assessment efforts on those parts that contribute most to corporate goals. A typical approach would be to evaluate each component to determine (by pegging up to the end item) how much revenue or margin would be lost if this component were not available anywhere. Components tied to high-value/high-volume end items would be at the top of the list as would common components used in multiple end items. The risk assessment/mitigation efforts should focus on these items first and be reviewed most often. But even if you do have a structured risk management approach like I've outlined above, testing your assumptions and your mitigation strategies (not to mention your ability to respond) makes a lot of sense. The Chaos Monkey approach seems to be an interesting way to do that. A twist to this would be to expand the exercise (especially given the events over the past year or so) to assess what would happen not just if the part weren't available, but if the vendor itself ceased to function. What if one of your key vendors was in Japan and was shut down by the earthquake? What if your supplier was in Thailand and was under several feet of water? An exercise like this will show that assessment and mitigation isn't necessarily going to be enough. Perhaps the part that was short wasn't one of the parts you have a mitigation strategy for. Perhaps the  mitigation strategy that you established several weeks or months ago is no longer valid. When this happens, your ability to meet customer demands depends on how quickly you respond. Your supply chain software must allow you to assess the impact of this problem, identify various alternatives and then quickly assess which alternative best meets the corporate goals. To do this, you need various capabilities:

1. You need to be able to quickly sense when an event has occurred. It is not enough to simply know that an order is late; you must understand the impact this order will have on customer demand (if the late order is safety stock replenishment, then perhaps dealing with it is less urgent than if the late order pegs to a large customer order).
2. Similar to my first point, you need to be able to assess impact of the event. If a supplier can't deliver any goods for 4 weeks, then you need to be able to make that change and instantly understand what that means to your key corporate metrics.
3. You need to be able to try several different resolution approaches. Can we get parts from another supplier? What are the cost implications? Can I ship parts from another location? Will they get there on time?
4. To do 2 and 3 effectively, you need to be able to model your supply chain in its entirety in order to capture the nuances required to make an informed decision. You need to have both the supply chain data from your ERP system and you need to be able to replicate your ERP analytics.
5. To do any of this fast enough to make a difference, you need to be able to run this analysis in minutes, not in days.

So, as the original article recommends, go down to the dock or storeroom and pull a part. Test your supply chain responsiveness. If you can't formulate a response you have confidence in within a few hours, maybe it's time to look at your supply chain risk and more importantly your ability to respond. Do you have a process similar to the Chaos Monkey? Do you think this is something worth looking at in your organization? Comment back and let us know!