As part of my MBA work at DePaul University I recently took a class on organizational behavior with Dr. Charles Naquin. His class on decision making really got me thinking about why companies automate operational decisions. I will get to why this applies to IT in a minute but let’s talk about biased decision making first.
In class we learned about heuristics which is a complicated name for mental shortcuts. There are literally thousands of heuristics that have been discovered and studied over the years but the following list is so common that nearly everyone in the world will be victim to all of them sooner or later (including you!).
But just by being aware of these shortcuts you have a better chance of avoiding them as well as blocking other people who are trying to manipulate you. In a short blog post there is not enough room to cover all these in depth so I have provided links to more information if you are interested in finding out more.
- Confirmation Bias – You only focus on evidence that supports your position and ignore conflicting evidence. http://en.wikipedia.org/wiki/Confirmation_bias
- Overconfidence – If you know less about something you will probably be more confident in your decisions about it. http://www.rtnlblog.net/overconfidence-bias
- Selective Recall – You will rely on things that are easier to remember when making judgments. http://en.wikipedia.org/wiki/Availability_heuristic
- Prospect Theory – You will choose the risk-averse choice if options are presented positively or the risky choice if they are presented negatively. http://www.innovationmanagement.se/imtool-articles/prospect-theory-risk-and-innovation/
- Anchoring – When estimating something your mind will rely on whatever information is nearby regardless of whether it is applicable or not. http://youarenotsosmart.com/2010/07/27/anchoring-effect/
- Escalation of Commitment – Cutting losses is very difficult even if it is the right thing to do. http://artpetty.com/2008/09/17/the-dollar-auction-and-a-failure-of-rational-judgment/
Why the Irrationality?
It all goes back to Biology. Think about your friendly caveman–we’ll call him Ug–thousands of years ago. When Ug was hunting breakfast in the morning there was no shortage of ways he might die. He could eat something poisonous, he could fall in a canyon or he could end up being the one on the menu instead. So our friend Ug needed to quickly make decisions as a matter of survival. Those same survival mechanisms survive in us today.
No Free Lunch
And thank goodness they did survive! Try to imagine if you had to put a pros and cons list together just to pick out your breakfast cereal in the morning; you might never make it to work.
But as you can probably already tell, heuristics come with a downside just like most shortcuts: They often lead to biased decisions which can result in embarrassment, failed projects or, worse, loss of life. Most people most of the time do not even realize they are relying on heuristics; they “go with their gut” and move on to the next thing. This is why it is important to know about them. That way you can explicitly decide when to use a more rational decision making approach instead of relying on happenstance.
Cool, but Relevance to IT?
In working with business rules technology over the last 7+ years I have talked with hundreds of customers and prospects who are interested in automating decisions. Sure, sometimes organizations automate decisions to improve efficiency but in retrospect I think many of them are looking to take bias out of their really important operational decisions.
For evidence, look at the customer list of any BRMS vendor. The biggest users of the technology are organizations in healthcare, financial services and the public sector. These organizations are constantly trying to automate the thousands of operational decisions that are made every day. And now you know why that makes sense… There are huge sums of money and human lives riding on many of those decisions. For example:
- Healthcare – It is pretty common knowledge that medical doctors get to spend less time caring for patients than in the past. They are expected to make diagnoses (decisions) quickly. No doubt heuristics are in play in the ER or exam room. More and more we are seeing rules used to make clinical recommendations to doctors based on a pre-defined protocol and patient data. If the doctor’s diagnosis differs from the recommendation then, if nothing else, it at least forces the doctor to think more deeply about his or her decision.
- Financial Services – Ahem, think about nearly every mortgage broker in 2007. That was an epidemic of confirmation bias. “I know that traditionally a borrower needs income, a job and assets to pay their mortgage but things are different now. I’m going to go ahead and approve this NINJA loan. What could go wrong?”
- Government Services – While case workers need some flexibility to use their judgment for individual cases, the core of making decisions regarding benefits must be consistent, often by law. If someone sees a different case worker, they should get the same answer. The answers also have to be transparent, auditable and defendable so that agencies can answer questions for citizens such as “Why didn’t I qualify for that program?”.
Automated decision making can take the irrationality of day-to-day decisions that impact you as a patient, borrower, customer or citizen resulting in more consistent treatment. That is a good thing for both you and the organizations that serve you.