Blog: The Data Center
I was reading one of my favorite economics blogs recently (marginalrevolution.com) and came across a link to a Harvard Business Review article by Daniel Kahneman, who won the Nobel Prize in Economics in 2002. The article is titled: “Noise: How to Overcome the High, Hidden Cost of Inconsistent Decision Making.” I would highly recommend you read the whole thing.
The crux of the article is that people within an organization can vary widely in their decision-making both individually and between individuals. For example, programmers given the same exact information on different days to estimate the time to complete a given task differed by 71%, on average! In another example, when pathologists made two assessments of the severity of the same biopsy samples on different days their results only correlated at 0.61 (out of a perfect 1.0). Both of these measurements were conducted using the same people. Differences between different individuals vary even more.
Unlike humans, a rules engine will return the same output if given the same inputs each and every time. By definition, a rules engine will noise-free. As we see at Decisions every day, it is possible to parameterize many of the common decisions made within an organization on a daily basis with improved and consistent outcomes. As Kahneman’s article states, people have competed against algorithms in contests of accuracy for over 60 years and the algorithms were more accurate in half the studies. In the other half of the studies it was a tie (which is a win for the algorithm as they are more cost-effective).
The most surprising result from the article (although not to us at Decisions) is that simple rules that apply equal weight to all the inputs were just as valid as complex statistical models built with lots of outcome data. Here is the key message from the article:
“The bottom line here is that if you plan to use an algorithm to reduce noise, you need not wait for outcome data. You can reap most of the benefits by using common sense to select variables and the simplest possible rule to combine them. Studies show that algorithms do better than humans in the role of decision maker.”
When you stop to think about how many decisions each and every one of your employees makes on a given day, how many of these decisions do you think would be made the same way tomorrow? Would you like to understand how to make noise-free decisions in your organization? I’d love to discuss the possibilities with you. Decisions can help. firstname.lastname@example.org