Friday 20 November 2009

Subjective Probabilities are an unavoidable part of decision making

Subjective probabilities are an unavoidable part of business decision making. 

You often have to make an opinion on strategic issues facing your business.  For example, you may be setting the five-year plan for your business.  You would have to assess all the factors which could have a big impact of the industry in which you operate in.

The situation is very complex. Your partners have different views and may not reach agreement.  On top of that, other industry leaders are making their views and this may have an impact. 

All these complexity doesn't prevent you and your partner from forming a view - maybe nothing more than an instinct or a hunch - as to what is going to happen.  Perhaps, you both agree that it is "quite likely" that a certain factor will impact the industry in the next two years.  Since this is of strategic significance to the business, you will need to accomodate this in the planning.

As you and your partner put your thoughts down on paper, what exactly does "quite likely" mean?  You may think it means "almost certain", while your partner considers it means "fifty-fifty".  In other words, you think "quite likely" equals a probability of (say) around 95%, while your partner assumes it denotes a probability of around 50%.

How can these two views be brought closer together.  Perhaps, they could use a probability that is objectively knowable - such as the throw of a dice - for comparison.  Do you think that such and such a factor is more or less likely to occur than throwing a six?  If less, the probability is lower than 1 in 6 (0.166).  If more, the probability is higher.  By discussing the issue in these terms, you and your partner can move closer to a picture of probability that you both share - and one that you can communicate with some degree of confidence.  You can both use this information to help pin down this probability - combined with your own opinions, experience and intuition. 

Let's assume you and your partner agree on a probability of 75% that a certain factor will impact on the business within the next two years.  It is important to note that just because two people have agreed a figure, the probability hasn't become any less subjective.  Using numbers adds clarity and precision but does not necessary indicate accuracy.  In your written report, you and your partner will need to explain the facts and reasoning behind your probability calculations, and stress the fact that the probability remains subjective even though it has been expressed numerically.  (You might use a range, such as '70-80%')

Some decision makers may regard this as pointless - how can that help you make a decision?  If you can't know probability objectively, why waste time trying to quantify it?  The answer is that it doesn't help you make the decision, but it does focus attention on the objective basis (if any) for assessments of probability.  It forces you to bring your information, reasoning and judgements into the open, so that others can see them. 

In the above example, you and your partner are forced to reach a shared understanding of probability so that you can communicate it and also, to others in your report.  While this doesn't necessarily makes it easier for you to make strategic decisions, it does mean that whatever decsion you take will be based on the facts that are available - or draw attention to the need for more facts.  Expressing probability numerically is also likely to focus everyone's minds on the urgency of the issue, rather than letting them adopt whatever interpretation of "quite likely" suits their own values and priorities.

Another benefit is the potential for sensitivity analysis:  to assess how the impact of a particular risk changes with respect to changes in probability of a particular factor.  Bigger changes mean higher sensitivity.

No comments: