Thomas C. Redman: Decision Criteria Through The Data And Information Lens
Introduction And Summary 
Perhaps no single management topic has been more closely studied, from so many angles, as decision-making.  This paper introduces the “data and information lens,” as a new viewpoint for examining decision-making. It is the most direct way of doing so because decisions involve nothing but data and information. Further, other approaches simply do not capture the many roles data and information play, call out the debilitating effects of bad inputs, identify the steps needed to develop trusted sources, or clarify the tough job of balancing uncertainty against the need to make and execute decisions promptly. Finally, the data and information lens complements other approaches.
Then, the paper uses that lens to look at the roles that the criteria employed by decision-makers (either individuals or groups) play in the shaping a decision. The topic is especially important for two reasons:
- Different criteria lead to different decisions. There are clearly situations where “playing it conservative” is more appropriate than “going for broke.” – and vice versa.
- Disagreements arise when individuals employ different criteria. The rising star and seasoned veteran complain that the other is “an old fogey” and a “cowboy” respectively, when the real issue is which criterion is most appropriate for making a sound, executable decision.
Decision-makers must spend some time thinking through the best criterion  for the circumstances at hand.
The Data and Information Lens
To study an organization through the data and information lens, one simply “follows the data and information.” Start at their points of creation and track them across the organization. Determine where they go, who touches them, how people and processes add value (or not), what they are used for, and how they get fouled up. Pay particular attention to management issues, especially who is responsible for what as they wind their way across the organization and be aware of the politics surrounding seemingly mundane issues like data sharing.
Decisions are the quintessential data and information products. So pay special attention to the interplays between decision makers’ baseline knowledge and the ways decision makers shape the issue and frame the decision space. Look at the ways they seek out more inputs and select the most important, how they deal with suspect data and sources, how they build support for the direction they wish to take, how they manage the uncertainty inherent in any important decision, the ways different “camps” emphasize different “facts” to support their opinions, and how the organization resolves conflicts. Study how decision makers treat those who will be affected by the decision and align the organization to execute. Finally, include anything else pertinent to the decision or decision-making process at hand, such as laws that must be followed or constraints stemming from the needs of important stakeholders.
The data and information lens naturally leads one to look at the criteria employed to make final decisions. Anyone who has ever performed a return on investment (ROI) calculation is familiar with “maximize expected return” as a decision criterion. Bayesian statistics contributes “minimize maximum risk” as another important criterion. There are thousands of possible criteria. It only stands to reason that the use of different decision criteria leads to different decisions.
To fully illustrate this point, consider Table 1, below. The top-half features three possible decisions (A, B, and C) and the probabilities of “success,” a “neutral” outcome,
Table 1: This example illustrates how the choice of decision criterion impacts the "best decision." (Reprinted with permission, T. C. Redman, Data Driven: Profiting from Your Most Important Business Asset, Harvard Business Press, 2008)
and “failure,” and the anticipated reward/penalty for each. The lower left quarter of the table summarizes results under 5 different decision criteria, and the lower right highlights the “best” decision. For example, B is the best decision if the goal is to maximize the “reward” when right, but it is sub-optimal under any other criterion. Table 1 is, of course, far more simplistic that any real decision. But it highlights two very important and often under-appreciated points. First, frequently a heated disagreement about “which facts are most pertinent” is really a disagreement about which criterion is most appropriate. Or the up-and-coming superstar may argue for the high-risk, high-reward decision, while the experienced veteran argues for the more conservative decision. Each has plenty of facts (e.g., ammunition) supporting his/her position and the discussion about whose data is best can become quite heated. The discussion can even become personal, with the up-and-coming superstar calling the veteran an “old fogey” (or worse) and the veteran calling the up-and-comer “a lone cowgirl.” But the root issue is that the veteran, perhaps implicitly, applies a “minimize the penalty if wrong” criterion and the up-and-comer uses a “maximize the reward if right” criterion. And this issue lies unexamined even though it contributes directly to the final decision.
The discussion points to an even more important issue: selecting the most appropriate criterion for the decision at hand. It appears to me that “maximizing expected return” is perfectly acceptable for most decisions. But it is not appropriate for decisions that could cause the company grave damage, nor in situations when even being right will not alter a desperate situation. There are no “criteria for selecting optimal decision criteria,” to guide decision makers in such situations. The best thrive (or at least seem to), perhaps because they have a sixth sense of “the best way to decide.”
For the rest of us, these points highlight the need to be explicit about decision criteria and to understand the effects of the choice of criterion on the decision-making process.
About the Author
Thomas C. Redman (“the Data Doc,” firstname.lastname@example.org) is president of Navesink Consulting Group in Little Silver, New Jersey. His latest book, Data Driven: Profiting from Your Most Important Business Asset, was just published by Harvard Business Press.
 This paper is abstracted from the author’s recently published book, Data Driven: Profiting from Your Most Important Business Asset, Harvard Business Press, Cambridge, MA, 2008.
 I strongly recommend the January, 2006 Special Issue of Harvard Business Review devoted to decision making.
 As a practical matter, many criteria may come into play for any decision. Here, for simplicity they are assumed to collapse into a single criterion.