Let’s jump on the ‘analytics’ bandwagon

“Analytics” is the latest hot topic – or buzzword, perhaps – in management. Tom Davenport’s 2007 book, “Competing on Analytics,” made it fashionable to seek quantitative methods to support better decisions. SAS Institute repositioned itself, aiming to be
seen as “the world leader in business analytics” rather than as a major provider of quantitative, especially statistical, software.About
a year and a half ago, IBM established a “Business Analytics and Optimization”practice area, which it has been marketing vigorously.
INFORMS has enjoyed considerable response to its new on-line news magazine, titled – what else? – Analytics.

So what?

I’ve heard a number ofmy colleagues in OR/MS say, dismissively, “That’s just O.R. by another name, maybe watered down so that more people can claim they do it.” What follows is often one or another proposal about how to protect the OR/MS “brand”from being diminished by competition from lightweight but heavily marketed competition. Other OR/MS analysts say, “Don’t knock it.At last the value of quantitative support for decision-making is getting recognized.”

Of course these conversations all take place within the continuing issue of how to market our profession, stimulated by the now largely abandoned “Science of Better” campaign. Then there’s the recent discussion, especially in OR/MS Today, about “soft OR” and, again, whether re-emphasizing client focus necessarily means diminishing analytical quality. A few months ago, a reader sent me a private e-mail admonishing me about the “Unsung Hero”ORacle column.

“Math is essential to what we do, and you’re implying it isn’t,” the reader wrote. He was right about what I was implying, but he still was missing the point.Mathematical ability is indeed essential to a good O.R. analyst, in my view.That doesn’t mean math is an essential component of any problem we will regard as within our field. Turning up our noses at problems that can and should be solved simply,without math, drastically limits how useful we look to our prospective clients.

“Analytics”is intended, I think, exactly to capture both the quantitative rigor that is sometimes required and the recognitionthat all problems come embedded within  organizational, cultural and interpersonal settings. Solving the technical problem in a way that violates cultural or social preferences is likely to produce resistance rather than appreciation. Solving the decisionmaking problem without math is often possible and, in those cases, best. Even when math is essential, a pretty good answer quickly is often much more valuable than a superb answer after a long, deep study. (When to terminate a search and conclude that the best answer observed to date is probably good enough would, in fact, be a most interesting O.R. research topic.)

The point is that we have to make a decision about what we think our profession comprises, and therefore who is truly qualified to practice it. If O.R. is mostly the set of techniques we teach in graduate programs called “operations research” or something similar, then we are headed toward being an increasingly elite but isolated group of specialists. If O.R. is a uniquely productive way of looking at the world, emphasizing multidisciplinary teams and a premium on innovation and creativity, we might be able to take over the decision-making world – but that “we”will refer to a much more diverse group, in capabilities and interests, than some of us would find comfortable.

I vote for inclusiveness.People in the profession express much frustration these days about why the world doesn’t value us more. It seems to me that most of the frustrations flow from the increasing codification of what O.R. means and what you have to know to be an O.R. analyst.We’re even moving away from the “MS” part of the “OR/MS” name, both in how we refer to ourselves and in what we imply about who we are. As I understand it, The Institute ofManagement Sciences (TIMS) started in large part as a practitioners’ rebellion against what they saw as an increasingly narrow, technical, inwardfocused definition of the profession. It was no accident that the Edelman Prize competition originated from within this part of the community: again, a way to redirect attention back to solving important problems rather than concentrating mostly on the cleverness and difficulty of the techniques used. In short, the debate about client focus versus technical brilliance isn’t new, and the implications shouldn’t surprise us.

My own experience in the business world taught me the supreme importance of meeting the clients’needs. It helped if, at least some of the time, my solution required technical expertise that few others could match; but first and foremost it had to work.Rejecting an offered problem-solving opportunity as “too simple” wasn’t an option. Staying close to our market is what concerns me now about INFORMS, as well. We need to make sure our society’s new initiatives keep that focus and advance us toward that objective.

Where our market is going right now is clear: big, powerful, influential firms are spending huge resources to promote “analytics,” whose definition is to be determined in progress. Whatever it turns out to mean, “analytics” is a bandwagon.We can jump on, get out of the way, or get run over.

To capitalize on the “analytics” movement in the right way, what we should do is what any good consultant does: offer helpful tidbits so that we’ll be welcome to stay in the client’s vicinity Then we can learn more about the problems and opportunities and then offer more substantive and substantial help.The annual SAS Global Forum, usually held over three days each spring, is a good model to emulate. Typically, the Forum begins on a Sunday with a one-day executive track that presents an overview of the most promising new analytical capabilities and how they are being applied. Subsequent technical tracks are aimed at teaching technical people how to use those capabilities. Scattered throughout are a few plenary sessions and a couple of receptions aimed to appeal to everybody. The intent is that teams of people from client and prospect companies will attend and, when they get back home, have a lively discussion about what new topics they heard about, why those topics are interesting, and how to make what they learned work for them.

INFORMS is not about to spend what I suspect SAS spends on Global Forum, but the idea is applicable enough.We can also learn from the perennial success the American Statistical Association (ASA) and other societies have enjoyed with continuing education programs. ASA is a particularly good example: what they do is identify authors of books that are about to appear and invite one to three of those authors each to do a one- or two-day short course keyed to the book, just before ASA’a annual national meeting. (ASA charges an additional fee, roughly equal to the conference registration fee, per day. It’s enough to be lucrative but low enough to look like a bargain to most employers’ training coordinators.) We could do short courses at this year’s annual meeting in Austin, Texas, for example,with Tom Davenport, who has a new book just coming out, and Doug Hubbard, who has just published a revised second edition of his 2007 book,“How to Measure Anything.”

We could then market these short courses and our “overview track” or “executives track” (better names than “spouses’ track,” I’d say, but same level of technical difficulty) to attract our usual attendees’ bosses’ bosses. Let’s get them all signed up to receive Analytics, which would then become one of our major vehicles to get the word out, at virtually no cost.

Of course, this and other things we could do might require some changes in INFORMS’ own decision-making structure and processes. We tend to look more and more like the academic world, where all new proposals get nibbled to death by committees in seemingly interminable deliberations. As with some of our modeling projects for clients, we let the decision drag on while we try to be sure our solution is optimal, and the perfect becomes the enemy of the good. Then something we’ve either over-deliberated or ignored  ecomes a crisis, and we decide by instinct, just like those MBA-trained competitors we like to criticize. And if they know us well enough to know what we do, they might well respond, “Why should we believe that your methods are better when you do not, in fact, employ them for your own decisions? How many meals do I want to buy in a restaurant where the cook won’t eat the food?”

We know what to do and how. We helped to invent the idea of agile decisionmaking, supported by good analysis but not enslaved to it, conducted by multi-disciplinary teams to reduce the blind spots, focused on doing whatever would work to solve real problems of some importance. Let’s apply that approach to making the most of the business world’s current interest in learning how to use quantitative thinking to make better decisions.Why not start today?