Good at Forecasting, Bad at Decisions
New research shows that groupthink gets in the way
Quote of the Moment
Management will continue to use techniques that don’t work, instead of adopting techniques that they don’t understand.
| Eric Bonabeau
New research — presented in Some Questions Benefit from Group Discussion. Others Don’t. by Joshua Becker, Douglas Guilbeault, and Edward “Ned” Smith — suggests that groups can be good at forecasting, but fall short when making final decisions.
Research on the concept of “collective intelligence” has shown that in many cases, groups tend to come up with more accurate estimates after discussing a question than individual experts do on their own. However, a new study found that while this holds true for quantitative questions — i.e., “How long will the project take?” — groups are actually less accurate than individuals when it comes to yes/no questions, such as, “Will the project be done before the deadline?”. Based on this nuanced distinction, the authors offer three strategies for managers to reap the benefits of group deliberation without falling prey to its downsides: Focus teams on discussing data, not predicting outcomes; separate “How Much?” questions from “Yes or No?” questions; and continuously capture data on group dynamics and team members’ strengths and weaknesses to inform future decision-making.
The predicting outcomes problem is an example of how social dynamics can lead groups to converge in their thinking, even when that convergence heads in the wrong direction. As the researchers elaborate:
During discussion, managers should encourage people to share relevant information, such as personal experiences, facts, and data — but not numerical estimates or decision recommendations.
However, managers should do their best to steer groups away from straw polls and public predictions. Comments such as I think it will be late, or, I think it will be done in three months tend to increase the similarity of opinions regardless of the reasons why group members feel one way or another, ultimately leading to fewer accurate votes when the time comes to make a decision.
And the Yes/No problem is stark:
One way to take advantage of the benefits of collective intelligence while still optimizing for accurate decision-making is to explicitly separate the forecasting discussion (how much?) from the decision (yes or no?). Consider asking a group to discuss a quantitative forecast, then take the average of their predictions and leave the final decision to a manager.
Or instead of a manager, a separate decision-making group that does not discuss the forecasts, but simply votes based on the crowdsourced forecasts.
An interesting footnote:
Both in our recent research and in our prior work, we’ve found that people who are more stubborn — that is, those who tend to make smaller revisions to their initial estimates after group discussion — also tend to be more accurate. By tracking data on both the evolution of group members’ opinions and ultimate project results, you can learn over time whose opinions to put the most stock into.
Ah! Stubbornness is a sign of wisdom, it seems. The researchers suggest tracking individual performance in the forecasting of different sorts: one person is good at forecasting cost, another at predicting delivery dates. Keep track.
Makes me wonder why companies don’t run prediction markets internally.
In Dopamine, Motivation, and the Science Behind Not Giving Up, Steve Dunne interviews Shivvy Jervison her research on human motivation:
Dunne: You also discussed how one unintended consequence of remote working during the pandemic has been greater decision-making efficiency. Tell us how you see that?
Jervison: Despite a tumultuous period for business and humanity, this phase has actually created some rather unexpected opportunities. Think about how many meetings on average it’s taken, within your teams or even in terms of your wider business line, to reach a significant business decision. Are you thinking eight, 12? Does it feel like 15? Well, it might surprise you to learn that a vast study by Deloitte found that it used to take a staggering 18 to 25 meetings for a core business decision to take place. But, since mid-2020, that figure is now seven to nine.
When a crisis hits, some of the aspects of an organization that were set in stone can be unlocked, because of the pressure a crisis injects on business-as-usual and the need to get the job done.
Sarah Kessler rounds up various surveys and reports on worker sentiment re: back into the office.
Conducted by Willis Towers Watson, the survey [see this] polled nearly 1,000 companies that together employ almost 10 million people:
52 percent plan to have vaccine mandates by the end of the year (including 21 percent that already do).
78 percent plan to track employees’ vaccination status (55 percent already do).
17 percent are considering health insurance premium rewards or surcharges to encourage vaccination (2 percent already do).
After the pandemic, one of the things that workers can probably count on is less business travel, according to a survey out Tuesday by Bloomberg of 45 large companies around the world:
84 percent of companies plan to spend less on travel after the pandemic, with a majority of those planning cuts of 20 to 40 percent of their prepandemic budgets. Put another way, all of those Zoom meetings aren’t going away.
Nothing will ever be the same.
Uber drivers are employees, Dutch judge rules — The slow erosion of the premises of the gig economy, resting on the backs of freelancers, continues.
Revolt of the NYC Delivery Workers — How delivery workers have spontaneously created loose associations to counter theft and confrontation by thieves trying to steal their electric bikes, filling a gap in what should be provided by the city’s police or the firms that employ them.
In Why so many workers have lost interest in their jobs, the BBC’s Kate Morgan makes disaffection seem like the workers’ fault, not the companies they work for, or the economic system that treats them like furniture.