Friday, February 1, 2013

Full Speed Ahead: onward and upwards towards the impossible

The comment of seeking enlightenment in last Tuesday's class caught my attention. Supply chains - especially those of today - are super complicated. They are so many moving parts, relationships, and contextual constraints depending on political, geographical and social factors. The world is so complicated that even trying to balance the information related to personal life is overwhelming. And advances in information technology make it ubiquitous and convenient. This convenience allows rationality to get even messier. But there's assurance in the movement towards big data, analysis and modeling. Although some people posit that computers will make the decisions for tomorrow, the article: Clearing the Crystal Ball by Tim Laseter et al. reveals a perspective of balancing the chi of forecasting and being comfortable with uncertainty and scrutiny. Organizations should support a culture of diversity and openness so decisions are informed by broad perspectives. Computer models process lots of inputs, but each is dependent on a set of conditions that often rely on intuition. Rather than deferring to a model for an answer, we should use the model to answer what its simulations ask. The culture of uncertainty and scrutiny does not have to be cynical, but instead structured to get closer to perfection.

The example below shows how assumptions can mislead.

This National Geographic article compares the value of an average pound of earth to the average pound of asteroid by amount of precious metals found. And yes, there are definitely more precious metals found in an asteroid than your backyard. And although the average total value of 100 tons of asteroid is worth $12,844 while dirt is $85 - as professor Jonathan Caulkins put it: "At $13,000 per 100 tons, comes to about 0.06 cents per pound. A carrot at the store is worth  more than that." Despite asteroid's value over dirt, this depends on the assumption that precious metals hold value and that average dirt is comparable.

Although this example is not directly related to supply chains it shows how data and technology can be implemented for useless analysis. Models can be made with complex algorithms and flawless logic, yet if they are not understandable by decision makers - how useful can they be? Managers and executive decision makers must be able to understand how the models apply and ask the following question. Does this help me answer the right question?

The perspective of being comfortable with uncertainty and scrutiny supports imperfection. However, managers would like to be close - and much more often than not. So when we consider supply issues like the bullwhip effect. (Which is the aggregating surplus cushion developed by each supplier following the consumer's indication of demand. So at the end of the "telephone-game" chain the last one holding the can phone is left supplying with cushion multiplied by the surplus ordered by each before them.) If we know this happens, and can supply chain managers need to resist the temptation and stay as close as they can to the source.

Here's a fun strip showing how perfect information can be perfectly wrong:

With advances towards automation, simulation, un-head-wrappable data, and demand for guaranteed forecasting - I wonder: What is the future position of the supply chain manager? How will they balance and sift through mountains of data and be sure the data answers the right question and they don't become guilty of Type III error?

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.