Sometimes, It's Better to Brainstorm Alone

Posted on Harvard Business Review: February 4, 2010 2:20 PM

Let's say you're assigned to come up with an incredibly great Super Bowl commercial for your company. The higher ups give you a comfy room and a half dozen really smart people to get it done. You gather them together, get an unlimited supply of veggie chips and bottled water, and let them brainstorm away. Pretty soon they're talking fake koala bears and stock-trading babies (you can probably tell I was watching CBS's "Super Bowl's Greatest Commercials" last night), and you're starting to daydream about the big bonus you're bound to get when your spot becomes the talked about Super Bowl ad.

But wait a minute. Why do they keep talking about koalas and babies? Can't they come up with something better than riffing off of what's already been done before? Where are all the fresh ideas inside those really smart heads?

Maybe a group isn't the way to get to a great ad. Should you instead divide up the chips and water and give each one of your team members his or her own comfy room so they can work alone, and in parallel? How do you know if the outcome will be better — or worse depending on which approach you take?

I've been reading the very thoughtful comments to my previous post with great interest. As long as brainstorming sessions aren't a sham, people really do like them. As Mike Brown says, "for anyone who's tried to sit at their desk by themselves to think up creative ideas," brainstorming is "a tremendously beneficial process." Bill Baker adds that "most of us sharpen and improve creative ideas by having someone to bounce them around with."

How can that attitude be reconciled with the abundance of psychological research over the past 50 years showing that a brainstorming team generates fewer ideas than the same number of people working solo?

Stylianos Kavadias of Georgia Tech and Svenja C. Sommer of HEC Paris, propose and answer to this question in a recent issue of Management Science. They point out that most research experiments on brainstorming involved searching for solutions to ridiculously simple or very complex problems—and those are just the types of situations in which brainstorming loses out to solo thinking. In dealing with highly complex problems, brainstorming can have a stifling effect, dangerously limiting the number of proposals that get serious consideration. You've probably experienced this yourself: Your glimmer of an idea quickly fades as other group members talk about their ideas.

But by breaking down the ideation process into its factors and modeling it mathematically, Kavadias and Sommer found that when the target problem is a "moderately" complex issue that involves multiple organizational functions, brainstorming is more effective than solitary musing. In those situations, "brainstorming groups can exploit the competence diversity of their members," and participants can build on one another's suggestions—just as Mike and Bill point out.

So what do "moderate" and "high" complexity mean? I'm looking at Kavadias and Sommer's paper, and I'm seeing a lot of integral signs and sigma signs, so I can't help you there. This is academic research, after all. But suffice it to say that moderately complex problems involving cross-functional expertise are as common in business as pushpins and bagels.

I'll be thinking about all this on Sunday as the Colts and Saints try to outwit each other on the field and advertisers try to outwit each other—and us—in the TV timeouts. When you're trying to figure out how to unleash the creativity of really smart people, you can't just assume that brainstorming or solo thinking will be best. You have to consider the complexity of the problem. More next time.

Before it's here, it's on the Bloomberg Terminal.