PRESIDENT'S DESK

Tell them the story

Steve Robinson, INFORMS Presidentpresident@informs.org

Steve Robinson

What INFORMS members do in analytics, operations research and management science is effective when it helps people understand better the options they have and the consequences of decisions they make. In the rest of this column I’ll use analysis to cover these areas, analysts for those who work in them, and clients for the people who have options and must make decisions.

To be effective, the analyst must at a minimum do three things:

1.     Understand what issues are to be addressed;

2.     Study and analyze to identify options and to predict consequences of actions; and

3.     Explain the options and their predicted consequences to the clients.

The ultimate objective that we all keep in mind as we prepare new analysts, whether in the academic world, in government or in industry, ought to be to provide competence in those three areas. From what I’ve seen in more than 50 years of doing analysis, the preparation process does pretty well at (2), less well at (1) and not very well at all at (3). In fact, new analysts often seem not to appreciate the crucial importance of explaining clearly what they’ve found. This is a serious fault: If the client can’t understand what the analyst has found in terms of the client’s concerns, she or he is very likely to decide that using the analysis is just too risky.

The italicized qualification in the last sentence is important, but junior analysts often misunderstand it. In many situations it’s not necessary for the client to follow the technical work that the analyst has done, and in fact the client may be completely unable to do so. What is necessary is for the client to be confident enough that the analyst has done competent and relevant work to be willing to accept and act on the results of the analysis. With this confidence, the client can proceed with implementation of those results in terms of operational actions with which he or she is very familiar. So the key communication requirement for the analyst is to put together a representation, which I’ll call the story, of the situation, the options, and the consequences that the client finds both intelligible and acceptable. And there may not be just one story; if the analyst has to brief people at different levels of the organization, they’ll usually require different stories.

Putting together a story is often difficult, and particularly so in a project involving significant uncertainty – which means most nontrivial projects. The uncertainty will have to be addressed, sometimes by probabilistic methods and at other times by devices such as scenarios. If the client isn’t comfortable with probability, then the analyst will have to put extra effort into the story, because it has to carry the burden of explaining the effects of uncertainty without using the usual terms and tools.

Uncertainty can also tempt briefers to avoid the problem entirely by concentrating on other aspects of the presentation, such as graphics and operational detail, while obscuring the real underlying tradeoffs that the client needs to understand. This can happen when an inexperienced briefer is simply unable to formulate those tradeoffs, but it can also happen when there are weaknesses in the methodology or study design so that the analysis is actually inadequate. An example of this avoidance problem came up when I was part of a group reviewing the development of a military combat simulation tool. Most of the briefing consisted of a beautifully animated graphic presentation of a mounted infantry deployment (a single action). There was no serious coverage of the tool’s capabilities and limitations. The graphics, though very pretty, contributed nothing to any real understanding of the simulation project.

In the United States, many future analysts are not getting from their secondary schooling the proficiency in communication that previous generations have had. This shows up in limited vocabulary and defective grammar when they enter a college or university. Some colleges and universities provide good remedial work in English, but fewer also provide good training in technical communication, which calls on different skills and devices than does writing an essay about a literary work. Communicating with clients is hard work even for someone with good writing and speaking skills, but without these skills it’s very difficult indeed.

The ultimate responsibility for preparing entry-level staff who can become competent analysts is with the school from which they graduate with their terminal degree. Often this will be an operations research program, a systems engineering department or a business school. We could take a good step toward fixing this problem if the hiring process required every would-be analyst graduating from such a school to have carried out a real analysis project for a real client, involving the three steps I mentioned earlier: understanding the issues, identifying options and predicting consequences, and explaining the options and consequences to the client. To keep their graduates marketable, the schools would then have to make sure that their students were proficient in each of these three skills, rather than concentrating on the middle one and skimping on the other two.