Getting Operations Management Some Respect

I recently attended a conference on manufacturing and services operations management, and it seems that quite a few researchers are grappling with the lack of recognition quantitative operations management suffers from in business schools and industry. More precisely, business students tend to care primarily about finance, and anecdotal evidence suggests that many companies don't make enough use of inventory control techniques, preferring ad hoc techniques without any optimization. Part of the issue is that researchers have to make simplifying assumptions to model real-life problems, while practitioners feel stylized models fail to capture all the complexities of their business, thus discarding the strategy suggested by the computer. While implementing a strategy blindly doesn't quite come across as a good plan, it should be possible to achieve a trade-off between providing quantitative guidelines and leaving the manager enough flexibility to apply his favorite rule-of-the-thumb within these guidelines.

In recent years, journal editors in the (very much overlapping) fields of operations research, management science, operations management (OM for short - pronounced O-M not "om"!) have encouraged authors to write non-technical summaries of their papers, keep industry executives in mind, ground their problems in real-life applications, but much remains to be done. There was some talk at the conference about the "closed" nature of the field, with OM academics reviewing papers by others OM academics, giving them prizes, deciding who has made a valuable contribution and who has not, with too little input from OM practitioners. While we'll have to wait for a real debate on how to get the field more visible and better noticed, some researchers (typically with tenure and full professorships) have advocated a more aggressive approach to consulting, especially at the higher-executive level, where the impact of OM techniques can be much greater because top management has the authority to implement far-reaching strategies.

One thing that I have not heard discussed enough is the line between academia and consulting. Again, some academics complement their income with consulting gigs. But no one ever talks about the industry practitioners or consultants who have a PhD in operations research or operations management and are in the trench lines every day. In some sense, and yes I know I'm exaggerating, but it almost feels like the default view of industry is a bunch of cavemen running around without any clue of what they're doing, and academics will swoop in and save the world in pure Superman fashion, before going back to teaching their courses with a modest look on their face. Many people in industry don't have PhDs and know nothing about demand distributions or queueing models (instead, they know a lot about the specificities of their own business and the relations with their suppliers and customers), but some companies, from Amazon.com to American Airlines, and many consulting firms (McKinsey, for instance) do hire PhDs. Those hires, because they were awarded a doctorate just like the academics, are supposed to be able to devise creative, technical solutions to complex problems - maybe their solutions will be less innovative, but is this going to prevent them from realizing cost benefits? (They might not find the optimal strategy, but sometimes suboptimal is good enough.) I am all for helping companies that do not hire PhDs and hire MBAs instead, but the emphasis on consulting could be interpreted to mean that we don't even trust our own former students to do well without our help - which is not even what we mean, because what we mean is that we're desperate for money to fund our current students and our summer salaries. And then you wonder why people resist the implementation of consultants' work? One of the plenary speakers did mention the importance of a good working relationship with the people you'll interact with on a daily basis when you're on-site, in five seconds out of a forty-minute talk. At some point you have to do the work, and in industry, much more than academia, you need other people to help you succeed - not all of them Chief Operating Officers.

One thing that I would love to see implemented to help OM techniques gain more recognition in industry without resorting to consulting projects would be the creation of an online repository (somewhat similar in spirit, but a lot more user-friendly, to the NETLIB library for linear programming problems) where industry members would upload sanitized data - sanitized means that the data is realistic but not real, to avoid giving any information to competitors on the state of their business - with their current performance statistics (total cost, percentage of time out of stock, etc) to beat, and academics could take that data, make whatever assumptions they want to make (demand is Poisson, customers who don't find the product they want leave without buying anything), solve the problem using their fancy math techniques, and upload the results on the site. Then a company could get in touch with the academic team who authored the most promising approach, and the industry PhDs could talk with the academic PhDs to gauge feasibility of implementation in their own business. That would also help judge numerical experiments in papers, since everybody could use the same test cases.

At least we are all becoming more aware of the necessity to publicize our work in a more systematic manner. I for one don't want to spend forty years in an office solving models to help decision-makers and realize afterwards no one but me ever knew about them.