“Work alone. You’re going to be best able to design revolutionary products and features if you’re working on your own. Not on a committee. Not on a team.”
– Steve Wozniak
In 1963 Marvin Dunnette, a psychology professor at the University of Minnesota performed an experiment that challenged conventional wisdom, yet few people know about it and even fewer have learned from it.
Dunnette gathered 48 research scientists and 48 advertising executives, all of them from 3M, and asked them to participate in both solitary and group brainstorming exercises.
He divided the group into twelve teams of four. Each foursome was given a problem to brainstorm, such as the benefits or difficulties of being born with an extra thumb. Each man was also given a similar problem to brainstorm on his own. Then Dunnette and his team counted all the ideas, comparing those produced by the groups with those generated by people working on their own.
The results were startling and counter-intuitive. The men in 23 of the 24 groups produced more ideas and of better quality when they worked on their own rather than in a group.
Since then some forty years of research and studies have shown that, almost always, performance gets worse as the group size increases. The “evidence from science suggests that business people must be insane to use brainstorming groups”, wrote the organizational psychologist Adrian Furnham.
When it comes to forecasting, collaboration is the accepted conventional wisdom. Consider retail supply chain forecasting. Most of us believe that collaborating with multiple people, potentially including the supplier, produces a better forecast. Yet, Dunnette’s research would suggest otherwise.
Perhaps better results could be achieved if forecasting was solely left to the retailer? And, further, to one person who is focused on a specific category of products. Then, using the Flowcasting process, this forecast of consumer demand would be translated into all other demand, supply, capacity, and financial plans for all partners in the supply chain.
But wait, you say, having the input of lots of people can materially improve the quality of the forecast. Sure, we all say that, but how specifically? It’s just accepted practice that if multiple people work on the forecast, then it will be better. Just like group brainstorming produces better results.
Do we think the forecast can be improved for promotional forecasts by having more people collaborate? Again, how exactly?
If you think about forecasting consumer demand isn’t the retailer closer to the consumer? Haven’t they got all kinds of information regarding past promotions for the item in question and similar items (likely from another supplier)?
Don’t they have a complete view of other marketing and promotional activities that might impact specific products being forecast at the same time? And isn’t this information available to the demand planner accountable for the forecast?
If two people were tasked with developing a forecast of consumer demand for a promotion and had the same inputs and history as a starting point, how different would their forecasts really be?
It’s a question that’s seldom asked.
Maybe, retail supply chain forecasting would be better served with one forecast, created by the retailer, owned, and managed by one person – and with the advances in Artificial Intelligence and Machine Learning, managed by reviewing only a miniscule number of exceptions that the machine couldn’t understand yet.
Maybe, though, we’re all wrong.