After six months at Google, a new manager was hired for my team. He instilled/beat into his team the value of the obligation to dissent. If everyone agrees on a subject, then the group hasn’t done enough thinking.

My manager subscribed to the idea that Peter Drucker captured when he wrote about a GM executive who once told his staff after all of them agreed on an issue something along the lines of  “If we all agree, then let’s adjourn for the day until we have informed and thoughtful points of view.”

This is the thesis of a New Yorker article entitled “GroupThink.” In order to have the best brainstorming sessions, those with many high quality ideas, people must be encouraged, even cajoled to disagree. Studies by teams Yale and Berkeley have demonstrated the notion of non-critical brainstorming, first proposed by a BBDO ad executive, is bunk.

Groups that are encouraged to analyze others’ ideas, finding strengths and weaknesses, generate more ideas of higher quality. In fact, they also keep thinking about the problem. After all, good analysis requires deep thought and engagement. Palliative treatment of ideas doesn’t encourage scrutiny.

The best way to start a brainstorming session is to begin by creating inaccurate or obscure associations. For example, if asked to say the first word that comes to mind after I said “blue”, instead of saying “sky”, it’s better to say “jazz” or “Picasso.” These unexpected links free our minds from well-trodden thought pathways.

Brainstorming is an essential skill for startups. Whether building product features, determining go-to-market, or hiring, make sure to brainstorm properly – and critique those ideas. The article is well worth reading because it highlights physical manifestation of the critique process – the famed MIT Building 20.