Analytics: Answering Key Questions

My subject here is the "analytics initiative" that has earned a high priority with INFORMS and CPMS. Actions resulting so far from the initiative include publishing the online Analytics magazine, renaming and reprogramming the spring practice conference, organizing the new analytics section, and, when it comes to fruition, offering professional certification in analytics. 

As the topic of analytics rose to prominence, we heard various anxiety provoking questions that still may trouble some of us. Let me therefore offer one person's perspectives on a sample of those questions. 

In the spirit of full disclosure, I should say that I emphatically support the directions regarding analytics being taken by the INFORMS board and CPMS. 

I'll organize my discussion in Q&A format. 

Q1: Why should we care? 
A1: From when Dr. A. P. Rowe, then head of the Bawdsey Research Station of the British Royal Air Force, in 1938 instructed investigators to undertake "operational researches" to make air defense more effective, the field of operations research/management science (ORMS) has existed to facilitate improvements in organizational performance - in other words, to deliver successful applications in practice. 

For years our biggest practice problem seemed to be that all too often the clients for our services either aren't aware of ORMS or else know about it but don't see it to be a source of valuable help.

After the merger of ORSA and TIMS to create INFORMS in 1995, INFORMS made two tries at addressing this apparent difficulty. First there was a publicity initiative called "branding," an initiative that got stuck in the board. Then there was another initiative, this time called "marketing the profession," perhaps better recognized as "the science of better" project, which also got stuck after a couple of years of activity. The analytics initiative is our third try. 

What's different this time? In a nutshell, the name "analytics" seems to have already achieved high visibility and popularity among many prospective clients. The board's job on behalf of our profession therefore is mainly to take rightful ownership of analytics.

Q2: Is analytics significantly different from ORMS?
A2: My short answer is that in practice the difference is negligible. 

So what is analytics? As is true of OR and MS, reasonable definitions vary. In my opinion the gist is that analytics combines the traditional decision-support role of ORMS with associated information technology (IT). It also gives more attention than some of us have given to analytical assistance from the field of artificial intelligence. 

Much of the overall view now called analytics was previously called "business intelligence." 

My main point is that success in practice does now, and has in the past, required being on top of virtually the whole scene now termed analytics. As Pogo might have observed, "We have met the analytics professionals and they are us!" 

Indeed, the basic idea of analytics is familiar. Franz Edelman, after whom our Edelman Award was named, promoted the tighter integration of OR and IT; he became head of both OR and IT at RCA. 

Note that the IT of interest in analytics is not general IT but instead is the IT associated with implementing decision-support applications. 

Q3: Should INFORMS focus on advanced analytics or on all analytics?
A3: In general ORMS practice it's routine to do quick, simple things as well as advanced things. A general physician will not recommend brain surgery when taking a couple of aspirin will suffice. But the ideal ORMS practitioner, like the ideal physician, recommends the appropriate treatment in light of all treatment possibilities; this requires the highest level of knowledge. 

The INFORMS initiative certainly will encompass the advanced end of the analytics-skill spectrum. To what extent the initiative reaches down to the simplest level is a judgment those immersed in formulating and implementing related projects, especially certification, must make. 

Q4: Why should INFORMS offer analytics certification?
A4: Offering certification is no minor undertaking. It entails cranking up a set of exams and related courses. 

The primary, strong reason for doing this is that it meets a real need. The proposal came to the board several times in the past. This time the need perhaps is more clear. 

Within many practice groups - whether the group is titled analytics, operations research, management science, or something else - some group members want certification to supplement their degrees. Usually, but not always, those group members interested in certification hold bachelors or masters degrees, not PhDs. 

Certification supplies a recognized credential. It typically can be pursued while the candidate stays on the job. Practice-group members often have specialized job roles; certification offers a chance to broaden their knowledge. 

Certification programs are commonplace in many professional fields. For example, the "professional engineer" program is well known in traditional ORMS groups. 

Q5: Will the name "analytics" replace the established names "OR" and "MS"?
A5: As the professional names "medicine" and "health care" illustrate, wide use of more than one name is not necessarily a problem. Our central goal related to the name, as perhaps became more evident in the INFORMS publicity effort that preceded the latest analytics effort, should be to "market the profession" - that is, to make the profession more visible to prospective clients and to encourage clients to recognize its practical value. 

Sometimes a profession might revise its public name to improve the tone; an example is the deliberate change from "personnel" to "human resources." This is not a consideration in our case. 

For us, the marketing-the-profession advantages of the name "analytics" emerge mainly from these features as perceived by prospective clients: (1) it sounds familiar and self-explanatory; (2) it sounds like something everyone should be doing, while operations research and management science sound impractical; and (3) it sounds catchy. 

Note that the name analytics could be, and is being, applied to non-practice pursuits (e.g., research) within our profession. 

Niels Bohr famously observed: "prediction is very difficult, especially about the future." It does seem likely that the names operations research, management science, and analytics, among others, will coexist for many years into the future. Which name dominates at a particular future time probably will depend on multiple circumstances only some of which INFORMS can influence. In any event, our primary concern should be the state of the profession rather than its most-popular name.

Q6: Will the new analytics section replace CPMS?
A6: CPMS, the practice section of INFORMS, is dedicated to advancing practice. The analytics section is dedicated to serving, in any appropriate respect, those in our profession who identify especially with the name analytics. 

The concepts behind these two missions differ significantly. It seems unlikely therefore that one will absorb the other as long as the missions stay the same. 

Q7: Aren't the analytics projects a lesser use of INFORMS resources when we could be doing more valuable things like starting new journals?
A7: Any member of the profession who believes that lack of visibility and appreciation among prospective practice clients isn't a top priority should reconsider. Art Geoffrion, the third president of INFORMS, often pointed out that we all are in the same boat; as practice goes, so goes the whole profession including academic jobs and our ability to serve the public good. 

Fortunately, we can afford what the board plans. The board is moving ahead carefully. For instance, while investing in the certification startup is not without risk, the certification program overall should at least break even soon and possibly make money later. 

Closing: I believe that the analytics initiative of INFORMS takes advantage of a golden opportunity. It deserves our support. 
I should mention again that the foregoing are my personal opinions, not official pronouncements. 

What's your view? Are you troubled by any other question about analytics? I welcome comments at randy.robinson@mac.com.