Profiles in OR/MS: Douglas Gray
|
Vice President,
B.S. Mathematical Sciences |
M.S. Operations Research,
Contact Information: |
Questions & Answers
Q. How did you arrive at your current job with ECWerks Inc.?
A. My last position at SABRE, from 7/95-7/97, was directing the development of Travelocity, a Web-based consumer travel reservation system. I caught the e-commerce/Internet "bug" upon seeing its potentially enormous impact on distribution-oriented businesses and industries. I was looking to expand my professional horizons by applying e-commerce concepts and technologies outside of travel and transportation. A colleague of mine at SABRE referred me to the founder and, then, CEO of ECWerks, Larry Buckley. Larry was looking to build an executive team to create an e-commerce consulting and technology solutions company that would offer services to a wide variety of companies across several different industries. Our missions matched up well, so Larry offered me the position of Vice President and Chief Technology Officer, and I accepted. I had always wanted to work with a "start up" company, so I jumped at the opportunity to join ECWerks in July 1997. I was one of the first 10 people on the team.
Q. How large is ECWerks Inc.?
A. At present, ECWerks has about 130 people on staff, the large majority of which are consulting professionals that specialize in providing enterprise e-commerce and e-business solutions.
Q. What are the main products and services ECWerks provides?
A. We do not offer any products, but rather focus on providing a combination of management consulting and technology consulting services. We help companies assemble business strategies to enable them to compete in the Digital Era, where information about customers, suppliers and supply chain operations is a critical resource. These strategies may involve creating new distribution channels for their products and services, or streamlining and optimizing supply chain operations. Once the strategies are defined, we assist customers in the design, development and implementation of enterprise-scale information technology architecture and supporting infrastructure to support e-commerce and e-business. Our service offerings really boil down to IT architecture design and systems integration work, as well as extending legacy systems out to suppliers, the sales force and directly to customers to enable companies to leverage Internet-paradigm technologies to increase revenues, decrease operating and selling costs, increase profitability and market share and, most importantly, enhance customer service.
Q. What are your main job responsibilities at ECWerks?
A. As a relatively young "start up" company (founded in April 1997), I wear a lot of hats, including technologist, consultant and business manager. I am intimately involved in the "business-side" of operating an IT professional services consulting firm. I also do executive level consultation with CIOs, COOs, CFOs and CEOs of clients and prospective customers. My predominant role, at the present time, involves technical sales and marketing. Our consulting business involves employing a lot of leading edge technology to enable companies to implement their e-commerce and e-business strategies. My job is provide the initial "technical consultation" with customers to demonstrate our company's knowledge, experience, and expertise in the creation of enterprise scale e-business solutions. I work with the sales team and our professional services team to identify exactly what the customer wants and needs, and then facilitate creation of a proposal or "statement of work" that clearly states the services that ECWerks will provide to that customer. I do a lot of "consultative selling" that involves brainstorming and "thinking outside the box" with customers to help them establish their e-business goals and objectives, and then lay out a technology architecture and implementation plan to achieve them. It is a very challenging position in that it involves a lot of communication, careful listening and open-mindedness to help customers turn what is very nebulous into a tangible design and a plan to get where they want to go.
Q. What are some of the OR/MS problems that you have worked on while employed at ECWerks?
A. To be quite honest, I have not worked on, what most folks would consider, "real OR/MS problems", using any formal OR/MS techniques since having arrived at ECWerks. However, I have identified a need for and an application of some OR/MS tools to solve some problems in the area of supply chain management and optimization. In 1994, I published an article in OR/MS Today about the application of airline yield management modeling techniques in other industries, e.g., manufacturing. I see several industries evolving towards being able to use some of the market segmentation and product/service profitability analysis concepts of yield management. I see that the Internet is creating an environment in which a variety of industries could leverage "real-time on-line information" to drive trading of commodities on-line much the way the stock market works today. "Spot markets", "auctions", "reverse auctions" for everything from electric power, to industrial chemicals, to steel-- anything for which there is supply and demand and price fluctuations that depend on the relationship between those two variables could give rise to the use of, what I referred to in my article as "revenue-based capacity management." Our company will be researching and developing some of these models in the not-too-distant future. Such models involve a combination of integer programming-based optimization models and heuristic algorithms, combined with statistical forecasting techniques. As most people know, yield management is one of the most challenging applications of OR/MS as it involves solving a dynamic, stochastic integer program--that is, deciding how many of each product category to make and sell at which price, when, the demand for which is subject to variation.
Q. Prior to your employment at ECWerks, what are some of the past OR/MS projects that you found especially memorable and rewarding? And why?
A. I was very fortunate in my almost 10 years at American Airlines/SABRE to have been able to work on such a wide variety of projects in the transportation industry, using many different OR/MS tools and techniques.
Early in my career, I worked on two projects with Qantas Airways involving airport capacity analysis and expansion planning at the airports in Sydney and Melbourne, Australia. Qantas did not feel that these airports had sufficient runway/taxiway capacity to handle the expected growth in traffic, and that the Federal Airports Corporation (FAC) was not planning to expand airport infrastructure accordingly, i.e., more runways, to handle the traffic. Using discrete-event simulation analysis, we were able to demonstrate that the runway infrastructure was indeed insufficient to handle the growth in traffic within acceptable delay tolerances.
That project was very successful, and most rewarding, because our analysis had a direct impact on the FAC's decision making process, and influenced subsequent airport expansion that has occurred over the 10 years since our study in 1988--bottom line, they built more runways!
I also had an opportunity to design, develop and implement a decision support system for strategic aircraft maintenance planning. The system, built in C/C++ on a Macintosh, enabled the rapid generation and evaluation of 5-year fleet maintenance forecasts and schedules, at the individual aircraft level, and maintenance hangar capacity plans.
This was my most successful and most rewarding OR/MS project in my career at AA, for several reasons. First of all, the user group had been generating fleet maintenance schedules manually and using spreadsheets. The schedule scenarios took approximately three hours to generate, and were not of very good quality, i.e., aircraft utilization in between maintenance visits was 80% versus a goal of 95%. The lower aircraft utilization, when extrapolated over a 20-year fleet life span, meant that the airline was performing 1-2 more maintenance visits, per aircraft, than was required by FAA regulations. At a cost of over $1 million per check on each of 300 wide-body aircraft in the fleet, the cost avoidance opportunity over the fleet life span was between $300-600 million dollars.
The system employed an interactive, iterative optimization-based scheduling heuristic approach, combined with a graphical user interface, to generate aircraft maintenance schedules in a matter of minutes. On average, aircraft utilization increased to between 95-100% using the new system. The benefits attributed to the system were substantial, and included: millions of dollars in unnecessary maintenance cost avoidance, increased aircraft utilization keeping aircraft flying in revenue-generating service, increased efficiency and effectiveness of the maintenance planners, and even an opportunity for the airline to lease out excess hangar space and perform contract maintenance work for other airlines as a result of improved hangar utilization on AA's maintenance bases. The cost savings and additional revenue more than paid for the project, which cost about $500,000 to roll-out over a three year period, not to mention my salary for my entire tenure at AA!
In general, the projects that were most memorable, and most rewarding, were in those in which we achieved the desired result and the customer was satisfied that we met or exceeded their expectations. Fortunately for me, that was the case more often than not! The OR/MS methods actually wind up being just one of many parts of a successful project.
Q. What are the most valuable technical skills that you believe are needed to be successful in the OR/MS industry?
A. Mathematical modeling skills are, in my opinion, the most important technical skills that an OR/MS practitioner needs to be successful in industry. The art and science of translating some real-world business or operational problem into some type of mathematical model that can be analyzed and observed to approximate the behavior of the real-world system is crucial to doing successful OR/MS.
To that point, fundamentally understanding the assumptions that underlie the mathematical models is critical to ensuring the relevance and applicability of a given model in representing the real-world system under study. I believe that there are a lot of fundamental skills involved in building mathematical models, like being able to represent various systems phenomena both algebraically in the form of equations, and geometrically in the form of pictures, to better understand complex real-world systems. Obviously, the algebraic formulations lead to optimization-based representations of problems, be they linear or non-linear, and the geometry leads to tools like discrete-event simulation analysis to see how the various elements of the system are going to interact. I think breadth across the spectrum of both deterministic and stochastic OR/MS models is important since real-world systems often involve elements of both modeling approaches. While algorithm development is an important and valuable skill, I believe that learning how to find the underlying structure of a problem first is critical to model it correctly. If you understand the structure properly, the algorithm development should be an experimental "trial and error" process that falls out naturally.
I believe that good programming skills are critical as well, especially if the models are going to implemented in the form of enterprise decision support systems. I would strongly encourage OR types to take a class in software engineering or at least programming in a modern language, like C or Java.
All of the above skills, unfortunately, don't come from a Master's degree in OR, but rather through a career full of learning experiences, project work or research. Graduate work in OR laid the foundation in theoretical mathematics for my OR career, but the best class I had to prepare me for reality was an upper-level undergraduate class in Mathematical Models. The class drew on a wide variety of real-world problems in economics, biological ecosystems and operations management for which there were no standard model forms in existence! We literally had to create models from scratch with nothing given other than a statement of the problem, unlike a class in statistics, for example, where the problems at the end of the chapter on Poisson processes are going to be problems suited to modeling with a Poisson process! This is what the real-world is like—first, you have to figure out the nature of the underlying problem; before you start slamming equations and writing code you have to determine what makes sense on your own!
Q. What are some of the most valuable non-technical skills that you believe are needed to be successful in the OR/MS industry?
A. Business, and the world in general, operates in many dimensions. There is the mathematical dimension of the problem (OR types love these!), the organizational dimension, the financial dimension, the legal dimension, the political dimension, the technology dimension, and finally the human dimension. I think that OR types need to be able to think and act in all of these dimensions simultaneously if their work is going to have a positive impact on the organizations and customers they serve.
Communication skills, specifically writing and speaking, are critical, but most importantly LISTENING is most important. A colleague of mine who is in sales told me once that God gave us two ears and one mouth because we should listen twice as much as we talk! That is as true in OR/MS consulting as in sales! We really need to understand what problem it is that the customer is trying to solve prior to launching into rigorous mathematics. I think that skills in all of the areas I mentioned above are valuable.
A class in finance, accounting or economics is critical to cost justifying your ideas and convincing the "bean-counters" to write the check! I think that keeping up with the latest in management trends (e.g., Drucker, Peters, Covey) along with courses in technical writing, business law and organizational psychology are useful. Even a Dale Carnegie Sales & Marketing course would help you sell your ideas to management more effectively. The Prince by Machievelli is a good read for understanding politics. These "soft, liberal arts areas" are the ones in which OR/MS professionals, including myself on occasion, have most often failed to be effective, hence limiting, in my opinion, the widespread success of the profession.
Q. In what ways do you continue to expand your knowledge of new technologies and techniques in OR/MS?
A. I try to keep up to date by reading OR/MS Today and Interfaces, however I no longer have the time, tenacity or will to read the Journal of OR anymore--way too mathematical for me! I follow the advances in applied optimization technology through supply chain management software product companies like ILOG and i2.
The area in which I am most interested lately is search engine technology that utilizes a combination of neural networks and adaptive pattern recognition to efficiently and effectively search large volumes of data. This area is of particular interest to me given the e-commerce and Internet marketplace in which my company is engaged. I believe that our company will be hiring some "hard core OR types" to build some of the yield management models and systems for our customers that I spoke of earlier. I would expect those folks to go to INFORMS conferences, read the appropriate journals and keep in touch with universities, research labs and other practitioners on the latest and greatest emerging technologies and techniques in OR/MS.
Q. What do you find most rewarding about your career in OR/MS?
A. What I enjoy most about my career in OR/MS is that I work at the intersection of challenging, real-world business problems, bright people with great ideas, powerful mathematical concepts and leading edge technology—and get paid pretty well for it! I thoroughly enjoy what I do for living, and what is most rewarding is that I get to help people solve complex problems that they either can't or don't have the time to solve on their own, i.e., I provide a value-added service to business and the world at large.
Although I am no longer what I would call a "hard core OR/MS practitioner", I still attribute much of my success in business to being able to clearly think through, analyze and solve very complex operational business problems using leading edge information and decision support technology. Coming up with creative solutions to client's problems is what consulting, and for that matter business in general, is really all about, regardless of what you are actually selling!
The diversity of opportunities for OR/MS and information technology people in the world today is really incredible. We live in a time of great technological change and progress, where everyone is surrounded, even inundated, with data, but not a lot of information with which to make better business decisions! What a better time to be an OR/MS professional—computers are more powerful and cheaper than ever, data abounds—there is no reason not to be gainfully employed solving complex, challenging problems. I may sound like a Pollyanna, but that is the way I see it!
Q. What advice do you have for those just starting out in a career in OR/MS?
A. Hopefully, you have already finished at least a Master's degree in OR, MS, IE or related field, as I feel that this is the best way to establish a firm foundation in the mathematical and statistical theory, and computational methods, that underlie the real-world application of OR/MS. Unless you plan teach at a university or do "hard core" research, a PhD is not really necessary to be an effective OR/MS practitioner, in my opinion. However, additional education is rarely a bad thing, as long as you don't mind the 4-6 years of lost income.
Secondly, I would seek out a position with a company/organization in an industry/sector that has a strong history of employing OR/MS, or is aggressively moving towards using quantitative methods. Good examples of potential industries/sectors include, but are not limited to, those recently represented in the Edelman Prize Competition, such as transportation, e.g., the airlines, logistics, e.g., trucking and shipping companies, financial services, i.e., the Wall Street investment firms, telecommunications firms, consulting firms, e.g., the "Big 5", military/industrial complex, law enforcement, food processing industries. Software technology companies that apply OR/MS, such as ILOG and i2 Technologies, are leaders in building the tools that enable practitioners to apply OR/MS more effectively.
Once you find a position, seek out an experienced OR/MS professional to serve as a mentor. Hopefully, this will be someone with whom you can actually work on project assignments and use as a "sounding board" for ideas, suggestions and guidance. I recommend volunteering for a wide variety of projects involving different areas of the business, using different OR/MS methods, to avoid getting "pigeon-holed" as an "inventory specialist" or "simulation specialist." Be sure to work on your communication skills, and the other "soft skills" mentioned above, because without them you will be labeled a "techie" or worse. Don't be afraid of failure, but be sure to evaluate your performance, learn from your mistakes and don't make the same mistake twice! Managers REALLY hate that! Take "calculated risks." Hopefully, you have found a career that you are passionate about and can have fun working at almost everyday. That, to me anyway, is the real key to success!
Q. What do you predict the future has in store for the field of OR/MS and for OR/MS practitioners?
A. I recently provided my thoughts in a cover article in OR/MS Today (April 1998) on what the profession would be like in the year 2048. I refer the reader to that article for a more detailed answer to the question. In general, though, with the rapid advancement of information and computer technology, and the Internet that is overwhelming everyone with data and information, I can't imagine anything but a bright future for the profession and practitioners. Data is the lifeblood of OR/MS models, and we as practitioners have a unique opportunity, and a professional responsibility, to help decision-makers turn that data into information to enable better decision-making. The opportunities to apply OR/MS to add value to the world at large are essentially unlimited, in my opinion. As long as we stay focused on our respective customers, whoever they are, and concentrate on satisfying or exceeding their true needs, and don't get too caught up in our own mathematical rigor, I think the future is bright. It is only when OR/MS stops being a means to an end, and becomes an end in itself that we get into trouble.
Q. Is there anything you would like to add that was not asked?
A. While my career has been varied and sometimes harried, marked by successes and checkered by some failures, I have thoroughly enjoyed my time spent to date as an OR/MS practitioner. I feel that my impact on the companies that I have worked for, my customers, and the profession itself has been a "net positive." I would encourage anyone who enjoys working at the confluence of ideas, people and organizations, challenging problems, mathematics, technology and business to consider pursuing a career in OR/MS. I doubt that you would regret it!