Making a Prediction Free Decision

7 Apr

In my last blog I talked about decision-making and prediction, concluding that when human behaviour acts across an unstable domain then prediction is impossible. The blog was based around comments made by IBM CEO, Ginni Rometty, interviewed in Forbes magazine. Rometty observed many decisions being made in business are still based on anchor biases, which basically means that when you’re faced with complexity you do what’s worked in the past.

Making decisions in a dynamic domain based on yesterday’s reasoning, or even the reasoning employed in the last hour, can be very costly. However, unlike Rometty, I think regardless of the theoretical merits of prediction analytics, where human behaviour and uncertainty are involved, prediction is impossible. But that’s not particularly helpful if you’d like to make more use of data, and this is what I’m going to discuss in this current blog- how to make a decision using data, but without using prediction or prediction analytics.

When faced with a plot of data there is a statistical tendency to display it as a bell curve or a Gaussian distribution. With the bell curve comes the corresponding tendency to look straight at the middle of the curve, the central tendency. The central tendency can be sense made as “the most likely to happen” and, as a result, it is the central tendency which becomes the basis for decision-making. The use of the bell curve is a major anchor bias in business decision-making and this is due to a focus on the central tendency shutting so much of the bigger picture out; the bell curve can turn probabilities which need urgent consideration into impossible looking scenarios which shouldn’t be wasting anyone’s time.

To mitigate this, when using data, stay away from the average and look at the range of probabilities the data is showing you, to quote NN Taleb “never cross a river which is four feet deep on average”. What hurts most, is the unexpected. The reason the unexpected or rare event causes so much damage is the lack of preparedness. This is what the average can do to your decision-making- make you very prepared for the statistically likely but completely unprepared for the least likely. By examining the range of probabilities you can plan for the central tendency and the rare event.

When you explore the range of probabilities you are in a position to prepare for eventualities. To assist you with this I’d advise borrowing a method from decision-making expert Gary Klein, the pre-mortem. The pre-mortem is a simple but highly effective method for testing any decision by first subjecting it to risk analysis. I’m a firm believer that the best decision-making begins with risk analysis and the pre-mortem delivers this effectively. To implement the pre-mortem method start by writing down your decision, then imagine it is 12 months into the future and the decision has proven to be a catastrophe; next, write down the fictional history, taking no more than 5 minutes, of this catastrophe. Use the range of probabilities to help you with this and ask colleagues independently (this will help de-correlate errors) to carry out the exercise to gain further insight.

The pre-mortem and probability ranges will enable you to test your decision hypothesis, allowing you to identify hidden pathogens, missing elements and opportunities. Once you have completed the pre-mortem, repeat it frequently, each time you carry out this exercise it will reveal missing data and information that you can seek out and keep integrating into your range of probabilities and the pre-mortem stages, generating improved hypotheses. This process can become a key part of your decision-making as you constantly update based on new data and risk analysis.

It could be argued that the above alternative is just another form of prediction. I would argue it is a form of constant preparation based on data informed updating. Focusing on preparation, keeping your options open to all probabilities, is central for any business to meet the requirements of the current key commercial skills – adaptive, innovative and agile.

One Response to “Making a Prediction Free Decision”

  1. marciano49 April 17, 2013 at 9:27 pm #

    Reblogged this on decisionsdotorg.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: