Can good decisions have bad outcomes?

March 22, 2012

in Advanced Analytics, Decision Management


There was an interesting discussion on LinkedIn recently – A good decision is the same as a good outcome. True, False, or “I don’t know? This illustrates one of the key challenges in Decision Management Systems – because good decisions can have bad outcomes.

This can seem counter-intuitive. After all if a company makes a strategic decision, to buy another company say, then a good decision will be widely regarded as one that has a good outcome. When we consider repeatable decisions, however, it is less clear. In this case a good decision should probably be defined as “a decision made using a good decision-making approach” – the right rules, the right analytics. Many such decisions will be made and it is almost inconceivable that all will have good outcomes. For instance a well designed and thoughtful cross-sell decision might result in a positive result (a successful cross-sell) 5% of the time – the other 95% will fail. Nevertheless this should be considered a good decision if 5% is a good result overall. In other words the fact that you have many bad outcomes does not mean you have a bad decision, only that a good decision has a low probability of success. Similarly a good risk decision will not result in no loans that go had, only in an expected number that is considered acceptable for that price and so on.

Good decisions, at least good decision-making approaches, can have bad outcomes.


Previous post:

Next post: