The Crystal Ball

"Crystal Ball" is actually a software program that is quite useful at analyzing future events. Keep in mind that "useful" does not necessarily mean "accurate". It's actually a statistical and analytics package, using mathematical methods to predict the probability of certain outcomes based on prior system behavior. Getting back to yesterday's conversation - I can't say that I would use a computer program to assess the potential success or failure of a potential candidate, but the underlying methodology is the same. The big question is this: are you an interpolator or an extrapolator?

An Interpolator only predicts future data that is within the bounds of historical observances. For example, if an interpolator were predicting the exact hour of sunrise, he would need observations from the begining of a month and the end of a month to predict the time of sunrise on, say, the 15th of the month. An extrapolator, however, recognizes trends and patterns and makes a few assumptions. So he may observe the time of sunrise on the first few days of the month and, based on this, predict the time of sunrise later in the month. From an uncertainty or "risk" standpoint, the interpolator is more certain of the correct answer.

So, what does this have to do with predicting success of a new leader? The primary data available to assess the future success of a leader is past experiences. The assumption is that the individual will perform under a similar pattern (interpolators can't be bothered assuming improvements or learning on the part of the leader). So, if I can find someone who has performed a similar or identical job function in a similar or identical company with a level of success, it doesn't take much analytical power to predict a high potential for success in the new job. Interpolators look for the "been there, done that" candidate.

Extrapolating, however, takes a bit of work and gets decision-makers a lot more nervous. Just like in statistical modeling and prediction, extrapolating human performance requires a thorough understanding of the inner workings of human motivation and team dynamics. It takes as much intuition as analytics to come up with a reasonable estimate. Going back to yesterday's conversation, this is where the discussion of learning from mistakes versus never making mistakes comes into play. What does the extrapolator use as a modeling assumption? Do you assume that the person has learned a lot from their mistakes and will be effective at avoiding similar mistakes in the future? Do you assume that someone who has made mistakes in the past is more likely to make mistakes in the future? Do you prefer someone who has a seemingly unblemished record, assuming that they are smart enough to never make a major mistake? It's always a lively discussion.

Fill a room with a combination of interpolators and extrapolators, each with a different set of experiences. Have them all interview a candidate for a newly created position in a unique matrix organization (one that has few, if any, direct parallels in peer companies), and sit in the room and listen to the debate about who is the best candidate. You don't need a crystal ball to determine the outcome of that exercise....