Type to search

HUB-Supply Chain Supply Chain Industry Updates

Your Demand Forecasting Might be Biased. Here’s What You Can Do About It

Share
iStock/Roman Didkivskyi

Over the past few decades, we have been overwhelmed with new technology, a preponderance of options for managing the supply chain and a ceaseless list of new tools to consider. With this focus on new technology, many organizations have forgotten one piece of the supply chain that remains unchanged: people.

Every supply chain involves people at some point, whether it is in the action to take following the creation of a forecast, deciding what data to input or manually adjusting the forecast. Therefore, it is critically important that behavioral economics principals are applied to your supply chain with as much enthusiasm as the latest machine learning offerings.

One such principle is that of unconscious biases and heuristics. Unconscious biases are ingrained ideas and assumptions that we use to make decisions without realizing that these influences are at play. We all have them, and while each person’s unconscious biases are different, there tend to be some similarities among people of a single group. Geographic region, gender, age, religion and socioeconomic class are a few examples of groups whose unconscious biases may overlap.  

Similarly, heuristics are cognitive shortcuts. Because we make more than 30,000 decisions a day, they are an effective way to rapidly navigate this complex choice landscape. Rather than taking in new information and switching from our cognitive brain to the prefrontal cortex to process and ponder the new data, people with experience use heuristics to make instinctive decisions.

This is the result of a surrogate of the prefrontal cortex knowledge in the cognitive brain, allowing you to think, “I’ve seen something like this before and I’m going to apply the lessons I learned from that situation to this scenario.” 

A Trio of Biases

These biases impact everyone, including demand planners and their work. We developed a tool to understand how unconscious biases and heuristics impact demand planners. We found some very interesting themes, particularly with the framing effect, overconfidence bias and cluster illusion bias.  

One of the most concerning findings was that half of our research study respondents altered their opinions about the same data when it was framed in different way (made more or less appealing). This is known as the “framing effect,” which is best illustrated with this example: Imagine you had the option to lose 10 out of 100 lives versus saving 90 out of 100 lives.

Most people would feel more comfortable saving 90 lives, despite the outcome being the same. If demand planners are affected by the framing effect, then how data is presented to them can have huge impacts on their forecasts.

It should also be noted that the framing effect was strongest in respondents with high “feeling” and “perceiving” personality characteristics, as defined by Jungian Typologies (made popular by the Myers-Briggs test). Being aware of personality types and hiring a diverse group of demand planners can help offset these biases.

Demand planners were also 10% more likely to exhibit overconfidence bias than non-demand planners. That is, they were more likely to have an unjustified certainty in their opinions, and this was particularly true among less experienced demand planners. This is consistent with the Dunning-Kruger effect, which shows that overconfidence decreases with experience. Helping demand planners, especially inexperienced ones, become conscious of this bias can have positive impacts on your team and forecast.

Finally, when correlating personality types and biases, we found that the cluster illusion bias was highest in responders with strong “sensing” characteristics. The cluster illusion bias is the tendency to erroneously conclude that randomly occurring data distributions are systemic (non-random). If your demand planners are finding faulty patterns in data (seasonality being a common one), this could result in inaccurate and overconfident forecasts.

How to Battle Bias

This does not mean that you should never hire demand planners with these personality traits or biases — it’s impossible to find someone who doesn’t exhibit biases. Rather, your organization should take these four steps to reduce and counteract the effects of bias:

  1. Make it conscious — One of the reasons we are affected by biases is that we don’t realize we have them. By having your demand planners learn about their own biases, they can identify and counteract them in their work. (Try this bias tool to learn your top biases.)
  2. Measure everything — Don’t make assumptions; measure the value of every input into your demand planning. Consider the forecast-value-add analysis as a framework to facilitate this.
  3. Train your team — Creating customized best practices training for your team, including a standard decision-making framework, improves organizational performance.
  4. Value diversity — Some personality types are heavily influenced by biases, so composing a team with differing viewpoints, experience and backgrounds benefits organizations immensely.

With those four recommendations,  your demand planning team should be better able to minimize the impact of biases on your demand forecast.

Jonathon Karelse is CEO and Y Nguyen is a senior associate with NorthFind Management.

Tags:
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]
[(args.length - 1)]