Tuesday, April 13, 2010

Too much information

@unorder In his post on Weak Signals, Shawn Callahan repeats four fallacies of sensemaking identified by Gary Klein. In this post, I want to look at the fourth one he identifies - that more data leads to better sensemaking.

In an earlier post, Shawn talked about when there is too much information (May 2005), which produced some interesting comments from Denham, Andrew and Clive Wilson.

Several presenting problems were identified in this discussion.
  • conflicting and voluminous information
  • complexity, uncertainty, unclear path
In such a situation, many managers will seek more information. Is this always incorrect?

Firstly, we may note that an automatic (knee-jerk) demand for more information may be motivated by a number of factors.
  • Ease, immediacy and relatively low cost of certain kinds of information retrieval (especially what is freely available and easily found using internet search engines);
  • Prior investment in information technology and analytic capability (skilled professionals);
  • Institutionalized risk aversion and "due diligence";
  • Requests for more information (or inquiries or studies) being a strategy to defer or avoid making a decision. (This strategy is commonly found in dysfunctional organizations as depicted in the Dilbert cartoons.)
But even if people are driven by ignorance, fear or laziness, they might still be doing the right thing by gathering more information. (Things may be done for the wrong reasons, but still produce good outcomes.) Organizations that encourage managers to think twice before acting may not be as bad as organizations that encourage managers to act without thinking at all.

Instead of gathering more information, Shawn advocates focusing on principles, values and preferences to help you make a choice, and basing your decisions on plausibility instead of accuracy. For his part, Clive disagrees; he thinks the problem is often not a lack of information as such, but a lack of relevant and useful information. "In such circumstances, you would expect to get better quality decisions by obtaining the additional information first, rather than making a decision in its absence."

I can see that both positions might be reasonable in different contexts. The real issue for me however is achieving a good balance between three things.
  1. The quantity and quality of information.
  2. The capacity of the people (working collectively) to use the information effectively.
  3. The demands of the situation.
In other words, we need to have just enough good information to support the collective intelligence of the management team in addressing the complexity of the situation. And in complex situations we need information not merely to support the decision itself but also to support the timing of the decision, which (after Lacan) we can divide into three phases: the instant of seeing, the time to comprehend, and the moment to conclude.

If the management team is already overloaded with information, then there seems little point merely trying to get more information. However, if we are overloaded with poor quality information, then it may be very useful to replace it with higher quality information. (This entails an ability to ask the right questions, and to frame any investigation intelligently.) At the same time, we may wish to increase the reasoning capacity (sense-making and decision-making) of the management team. If we can't manage to connect the dots, simply having more dots all over the place isn't necessarily going to help, but having dots closer together may help us to see the pattern. (By the way, I think I agree with Klein that sense-making isn't just about connecting the dots, but it is part of the story. I shall come back to that in another post.)

Gary Klein, Brian Moon, Robert R. Hoffman, "Making Sense of Sensemaking 1: Alternative Perspectives," IEEE Intelligent Systems, vol. 21, no. 4, pp. 70-73, July/Aug. 2006, doi:10.1109/MIS.2006.75

See also Sensemaking in a World of Shadows, a review by Stephen Few of Gary Klein's latest book.

No comments:

Post a Comment