Seeing around the bend at BNSF

Today's business intelligence and tomorrow's opportunities meet at the crossroads.

by Karen D. Schwartz

On the surface, a railroad responsible for getting goods from point A to point B doesn't seem a likely candidate for state-of-the-art technology. But in this fast-paced world, where moving products quickly and efficiently is crucial to success, technology is critical to allowing the Burlington Northern and Santa Fe Railway Company (BNSF) to meet its goals of superior service, agility, efficiency and profitability.

Headquartered in Fort Worth, Texas, BNSF was created from the merger of Burlington Northern and Santa Fe Pacific in 1995. BNSF operates one of the largest railroad networks in North America, with 32,500 route miles covering 28 states and two Canadian provinces. The company is one of the largest transporters of intermodal traffic in the world. BNSF also is the largest grain-hauling railroad and hauls enough coal to generate more than 10% of the electricity produced in the United States. All of the company's Dash 9 locomotives include an integrated microprocessor control system that provides onboard diagnostics, as well as other systems that improve fuel economy, reduce emissions and increase engine life.

Teradata Magazine spoke with Bonnie Henn-Pritchard, who leads BNSF's technology services application team, to find out how the railroad powerhouse is using technology to further its goals.

Q: How have changes in the railroad industry fueled the demand for more sophisticated technology?
A: In the past decade, the railroad industry has changed significantly as our customers are pressing for visibility and exchange of data across the supply chain. They want to know where things are at every point in time to coordinate with their resources and ensure efficient and effective shipments. Today, customers don't want to treat each link of the supply chain as individual entities; they want an integrated approach to the information so they do not have to deal separately with consolidators, brokers, steamships and truck lines. This means each link in the chain—one of which we represent—has to work together with the others to form a cohesive unit.

Q: Given those changes, what are your business goals today?
A: Our challenges are to strive to find ways to better understand our customers, increase asset utilization efficiencies, grow revenues and reduce costs. Given the dynamic environment in which we operate, we need to continue to find ways to become more proactive and nimble in our decision-making and actions—and that's no mean feat for a 38,000-person, 32,500-mile community.

From a technology standpoint, we need integrated, sophisticated applications to support the operation in achieving these goals. For example, we have nearly 100,000 active cars at any one time on our routes across the western half of the United States. If we experience an issue—a flood or blizzard, for example—on one of our main lines, we have to be able to re-route and notify all customers as quickly as possible so we can get back on schedule. That takes integrated business processes supported by strong technology. Also, we have a mission to become easier to do business with. We need tools, processes and methodologies that allow our customers to interface with us more easily. That has driven us to develop BNSF.com, a suite of Web-based tools used by our customers, and even stronger integration with their systems.

Q: How has business intelligence helped address the business challenges you face?
A: We have developed several applications to assist in these challenges. As we discuss a few, I think it is useful to keep in mind the common business principles that we abide by as we expand the use of business intelligence: 1) provide a single version of the truth, 2) store the details, 3) provide increased value through productivity and 4) enable improved analysis. As we develop applications, we use these principles as guideposts to ensure we are delivering the greatest business value.

Q: What has been your greatest challenge in terms of forecasting and budgeting applications?
A: Our business customers had been using Microsoft Excel for years for forecasting and budgeting, and similar to other companies of our size, we had developed a plethora of spreadsheets. We knew we needed to develop a world-class, integrated planning and forecasting system through the use of centralized, driver-based business methodologies and best-of-breed technologies. The goal was to produce a more automated planning and forecasting system that decreased the amount of time that it takes to develop a forecast and provided flexibility in quickly doing 'what-if' scenarios for different situations. We also wanted to be able to use standardized business rules and definitions while simplifying the process for the business users overall. Previously, we were accomplishing this goal with a variety of spreadsheets, which was a tedious and time-consuming process.

Q: How did you address the issue?
A: In 2002, we designed OnTrack, our forecasting and planning system that we consider state-of-the-industry. By 2004, BNSF was named a Premier Technology Leader by Computerworld magazine, in part for our forecasting and planning tools. That same year, BNSF received awards from both Target and Toyota Logistics Services for customer service and from Wal-Mart for Carrier of the Year.

Our system uses the Teradata Warehouse to house the data, frontend technology is built around Hyperion's Essbase, and we also have some custom Java-developed screens. Reporting functionality is provided through use of the IBM DB2 Alphablox Web-reporting tool and Hyperion's Excel add-in.

Q: What can you do with OnTrack that you couldn't do as easily before?
A: Its flexibility is its greatest asset. Our budget managers use the system to create their forecasts, but they can also plan and forecast in a what-if mode. For example, they can change one factor, like adding more crew starts. By changing that one variable, they can create an alternate forecast that can be used to compare and contrast in order to arrive at the best forecast. The new system is much more automated and standardized, making it simpler and faster to use. In the past users had to key numbers into a spreadsheet; today that data is fed automatically, decreasing overall cycle time. Standard and ad hoc reporting capabilities are greatly increased for the 80 power users of OnTrack, with reports going out to 200-300 people.

Q: Your company seems to be very focused on metrics. Do you have systems in place to quickly pinpoint and address problem areas that can crop up unexpectedly?
A: We have extensive reporting capabilities on our metrics. We're creating an online approach for focusing on the key indicators we look at each day so we can quickly dive into where the problems are and drill down to understand the root causes. We're currently in beta test mode with a tool we call the Corporate Dashboard, which will provide a single place for everyone from executives to analysts to go to find coordinated and integrated information and work on a common scoreboard. We're building the systems ourselves using data pulled from our Teradata Warehouse, DB2 and SAS systems, with Essbase on the OLAP side and Alphablox for Web delivery.

Q: How will the Corporate Dashboard work when fully operational?
A: There are three key views within the dashboard. Through the first view, which uses 'stop-light' color coding to quickly point out exception areas, a general manager might look at the status of trains or crews, check revenues and look at other key metrics around the railroad. Then, through a series of simple clicks, he will drill down to the details of the exception areas to delineate the exact place where the problem is occurring.

If he needs more details or more correlation with other data points, he can pop out of the visualization view into detailed data analysis where there are four quadrants. The first quadrant has the exact data associated with the problem; a second compares that metric with another key indicator that is a determinant of why that might be a problem; a third is for trending; and the last provides area-specific information, such as standards of performance, for that metric. So a manager can quickly determine where the problems are, drill down into those problems and get several different views correlated with other types of metrics.

Finally, for deeper analysis, he can slice and dice data elements from the data warehouse and other sources to get a really comprehensive perspective.

Q: What business benefits do you expect the Corporate Dashboard to give BNSF?
A: Faster reaction time. Right now we do have a good amount of analytical information that is used extensively. It requires a bit of time to review and determine the root causes.

We believe a tool that focuses on key indicators, quickly pinpoints problem areas, simplifies navigation through those areas and summarizes critical data factors will speed up analysis time—giving more time to evaluate and make decisions. But just as important, we're establishing a foundation for predictive analytics and providing insight into the business.

Q: When do you expect the Corporate Dashboard to be in production?
A: During the first quarter of 2005, but it will be a work in progress. Each quarter, we plan to release a new version, with additional functionality like expanded metrics, more drill-downs, automatic alerts and notifications, and personalization.

Q: How are you using technology to automate your invoicing process?
A: We developed the Revenue Single Source system in 2002, which enhances our ability to make changes to waybills, eliminates redundancy and helps us process bills more quickly and accurately. The idea was to provide BNSF employees a single source from which to view information concerning a shipment from the time the shipment is ordered until the bill has been paid correctly—cradle-to-grave processing. This is all available through a Web system that uses rule-based engines, looks for patterns for processing and combines tolerance alerts with events notification while utilizing an active data warehouse.

Q: What technology was involved in developing Revenue Single Source?
A: We pull transactions from our IBM System 390 mainframes using DB2, Cobol, Natural and other languages into our Teradata Warehouse, where it is stored. On top of that, we have several applications written in Java. Similar to our other business intelligence systems, we also use Essbase and Alphablox.

Q: How have you enhanced Revenue Single Source to further utilize the information you have gathered?
A: We've recently added case-based reasoning, which not only allows us to track accuracy but also analyzes the quality of a decision. The system gains intelligence and recognizes not only real patterns, but it also incorporates fuzzy-driver logic, allowing us to choose to automate decisions that haven't yet been encountered. For example, the system might recommend a decision or fix based on a 90% accuracy rate from similar decisions humans have made in the past.

Q: What are your future plans regarding technology?
A: Our vision is that of an 'eSynchronous railroad' in which our operation, our customers and our employees are tied together with information, knowledge management, collaboration and integration among systems to reflect the status of the railroad in near real-time—or what we call 'right time.' The underpinnings of that vision are the quality and flow of information, which means we need to have core systems which are industrial-strength, stable and secure while building strong analytic systems to support decision-making. To that end, we are continuing to develop applications and find ways to get more information and analytics from our transactional systems into the Teradata Warehouse in near real-time. T

Behind the solution

Teradata Warehouse powered by: Teradata Database V2R5, 16 nodes across 5250/5350/5380 NCR Servers Storage: Total Disk: 23,904GB User Disk: 10,963GB Operating System: UNIX MP-RAS Teradata Utilities: MultiLoad, BTEQ Tools/Apps.: Products from IBM, Hummingbird Ltd., Siebel Systems, Inc. and Hyperion

Karen D. Schwartz is a Washington, D.C.- based business and technology writer. Her work has appeared in such publications as CIO, Information Week and Business 2.0.