October 2001

Five Steps to Better Supply Chain Performance

There are clear signs that the industry is beginning to appreciate performance concerns.

"Which comes first, the chicken (on-time performance), or the egg (shippers' trust)? Do the railroads understand that ‘on time’ means when it's needed, and that early can be as bad as late?" – Ed Rastatter, NIT League, in a September 2001 e-mail

Two months ago I wrote "Being on time means when the customer expects it." (RA, August, 2001). In that article I laid out three benchmarks for measuring railroad performance against customer expectations: Ease of Doing Business, Consistency, and Delivered Cost. There are clear signs that the industry is beginning to appreciate the performance concerns of Rastatter and others and is taking steps to fit performance to expectations. Here’s why and how.

Railroad owners have to approach each measure as a goal and then examine every activity and expense affecting the goal. Take Ease of Doing Business. Say the Fallen Flag & Eastern (FF&E) is having trouble supplying empties. On closer examination, it turns out the FF&E has a standing rule to minimize foreign car hire expense. Could this be part of the problem?

The first step is taking the right internal measurements. Here, one could define the problem as a perception of paying too much foreign car hire. How much is too much? The FF&E has to set a measure, say more than four days each way (interchange to empty release or load release to interchange). Once measured, FF&E managers must analyze what they find and drill down to root causes where the plan is exceeded. The next steps are to improve processes to eliminate root causes and to control future processes to stay on plan.

A real live example is a terminal railroad with a 19-hour goal (a preset measure) to receive a train, classify it, and have each car released to its departure track. It wasn’t happening. We defined the problem when we found the average dwell time over a three-week period was 26.7 hours. Our analysis showed individual car times ranged from 19 to 43 hours with one standard deviation equal to 5.6 hours. This told us that two thirds of the events fell in a range of 21.1 and 32.3 hours, all outside our goal.

Worse, the remaining third of all events is split – a sixth more than 32.3 hours and a sixth under 21.1 hours, meaning that less than a sixth of all events were within plan. To improve the process, we first had to take apart the process. Late trains caused delays by having to be reslotted amidst the days’ work. A sixth of the allotted time was consumed inspecting the inbound trains. More time was lost when too many tracks were used to make up the outbounds or sorting out no-bills. And so on.

Today, the average dwell time is 14 hours. Inbounds arrive on schedule and crews set upon them immediately. All class tracks are pre-assigned for the outbound blocks and track lists are more accurate. No-bills go to one place, shop cars to another – same places, all three shifts. And we watch dwell performance constantly so the controls are in. The most elegant part of the story is there was virtually no added cost and so the savings went right to the operating ratio. Better yet, better service begets more business further improving profitability, morale, and job satisfaction.

Happily, the small roads don’t have to do all this in a vacuum. Norfolk Southern and BNSF for example have developed performance measurements that are robust, improve customer service, and take points out of the operating ratio. NS has built its program around its Thoroughbred Yard Enterprise System (TYES) and the Algorithmic Blocking and Classification System (ABC).

The result is a tool that automatically finds the path of least resistance through the NS rail network. It assigns numeric penalties for activities that don’t keep the cars moving – humps, flat switching, etc. If certain routes get too congested, the system finds least cost alternatives, always keeping the original trip plan foremost. Moreover, TYES can look beyond the NS system at specific car movements, enabling the construction of a customer activity database that can then be used to fine tune the process to predict customer order and release behavior. Then you can start managing the pipeline for maximum throughput.

Central to the BNSF system is the "Transportation Service Plan" that creates a trip plan for each car as it’s released to the railroad (you can see the trip plan algorithm results at the www.bnsf.com Transit Time page). Being able to tell a customer exactly when the car will arrive and making it easy to for customers to access relevant BNSF movement info is a powerful marketing tool.

But whether we call it TYES or TSP, the common thread is the ability to define, measure, analyze, improve, and control every activity from car order to empty release. The benefit to shippers is measurable supply chain performance. The benefit to railroads of all sizes is measurable financial performance. So what are we waiting for?

Railway Age Columns Home
Send comments to roy@rblanchard.com