Calculating Risk

Any logistics director moving goods through North American ports can hardly be blamed for being risk-averse. Whether it’s a trucker strike in Vancouver, British Columbia, hours-long backups at the Port of New York-New Jersey, or marine terminals buckling under the weight of thousands of containers disgorged by single mega-ships, there’s no shortage of reasons for shippers to build extra time into their supply chains in the chaotic environment of 2014.

And shippers are doing just that as they attempt to deal with any possible flare-ups at ports and because they know they may need to build a few extra days into lead times to account for ships that arrive on weekends.

The problem is, it’s easy to get carried away in building in lead time. Fear of disappointing customers just once can lead to overly cautious transit time assumptions. Days can be piled on simply by using the transit time of the slowest carrier with which the shipper has a contract.

“Some logisticians go to great lengths to factor in many items that could make a shipment not meet its intended delivery so that they can be assured the goods will arrive within the window specified,” said John Motley, CEO of logistics software and services company Log-Net. “Some international supply chains can have 10 to 20 days of additional duration to avoid any of the risks that can happen in a supply chain.”

With studies showing growing risks across the supply chain — one reason often cited for the growth in near-shoring — it’s understandable that risk would be prominent in lead time calculations. But while supply chains based on excessive caution ultimately may accomplish the goal of on-time reliability, what is the ultimate cost? Inventory-carrying costs, storage costs and added transportation expense may greatly exceed the benefits, leaving the shipper less profitable and less competitive.

Thus, the question posed by every logistics professional: how to build in extra lead time — the time from when an order is placed with the factory to when it’s delivered to the customer — in a way that ensures a consistent level of delivery reliability while minimizing unnecessary expense?

Not all shipments can avoid the fate of having to be sent by expensive air freight, not all can avoid unanticipated labor disruption, and not all will flow seamlessly in one door of a distribution center and out the other. But is there a way to factor in all elements of risk while achieving this elusive balance?

Motley says many shippers get themselves into trouble by determining lead times based on the average time their containers require to complete key milestones, such as time on the ocean, processing at the port or customs clearance. The problem with that approach is that actual experience can deviate widely from the average. It’s like saying a shipper had 100 containers that took anywhere from 0 to 100 hours to make it through a terminal. The average is 50, but how can you plan based on that figure when you don’t know if the next container might take two hours or 98?

This leads naturally into excessive lead time planning, Motley says. “As supply chain and purchasing professionals gain experience with the pain of dealing with an average that is exceeded 50 percent of the time, they will invariably begin to add additional lead time as contingency.”

The solution is instead to look at such historical data in terms of how variable the data points are statistically — in other words standard deviation, a measure of the dispersion of a set of measurements from a mean. A high standard deviation in this context means the transport lane has very unpredictable transits. The key is predictability, the confidence level in percentage terms that the logistician has in achieving a certain result, such as an on-time arrival.

Because the availability and quality of data pertaining to transit times, port dwell times and other supply chain junctures is improving, it’s possible to deeply analyze transit time variability and call out the problem areas. It might be that scheduling sailings from Shanghai on Mondays is rife with variability or that a particular carrier’s arrivals at Los Angeles are all over the map. It could be that variability is tied to specific terminals and using on-dock rail versus transloading.

In one example Motley cited, a carrier had a fortnightly barge service at the origin that caused an extra five days at origin coupled with a Friday arrival at the destination port, which caused a second four- to five-day delay.

“It represents an opportunity for tremendous savings if we can understand and eliminate the deterministic cause of the variability,” Motley said. Those savings take the form of inventory-carrying costs and at a basic level having to ship fewer containers if there is less inventory tied up as buffer stock.

The concept Motley is talking about is called predictive analytics. It may sound complex, but growing numbers of shippers are embracing this approach as the next big step in their supply chains.  

Contact Peter Tirschwell at ptirschwell@joc.com and follow him on Twitter: @petertirschwell.

For the full story: Log In, Register for Free or Subscribe