Hacker News new | past | comments | ask | show | jobs | submit login
Brainstorming Doesn't Really Work (newyorker.com)
139 points by gruseom on Feb 14, 2012 | hide | past | web | favorite | 40 comments

The no-criticism rule is probably more important in a hierarchical organization. At BBDO in the 1940s, I imagine junior creatives were intimidated by senior partners and the challenge was to get them to speak freely. The experiments that show no benefit were done on groups of students, all roughly peers who weren't afraid to utter a silly idea.

The article suggests that brainstorming was invalid all along, but it's more fair to say that it had a time and place. Society and business have evolved towards equality and inclusiveness, so the no-criticism rule is less important.

I think your point might be valid, but it would be nice to test it. The obvious place to try it might be Japan which still has a pretty hierarchical corporate culture.

Perhaps a stronger argument along the same lines is that in a business setting anything that breaks away from the highly dysfunctional mechanics of many business meetings (which are not purely due to hierarchy) has to be a Good Thing.

The linkbait title does not correctly convey the contents. The actual conclusion is: brainstorming works, but only if you critically evaluate the products of the brainstorm session.

This matches both my experience and what my girlfriend was taught concerning how to conduct brainstorm sessions, which includes the variation of first having everyone brainstorm alone, then combining and grouping the outcomes and then evaluating the outcome.

> the variation of first having everyone brainstorm alone, then combining and grouping the outcomes

This. A thousand times this.

I find the biggest common issue with brainstorming is that you end up with an incoherent amalgamation of things that people think are individually important. People have a hard time evaluating options unless they are part of a coherent whole. Brainstorming beforehand, alone, allows them to figure out what's important and then imagine a coherent story around that thought. Otherwise you end up with solutions that technically contain individual important details but make no coherent sense as a whole. It's like saying "well, Suzie like dragons and Jimmy like potatoes so we put them together in a story." If Suzie had brainstormed before she could propose a backstory and context that the group finds more compelling than Jimmy's context about potatoes and they could have a more fruitful discussion than "I don't like potatoes," or "Dragons are dangerous."

Edit: Now that I think about it, this might be at the core of what the article describes as the optimum collaborative environment for innovating - people who exist separately and have their own coherent problems and solutions to worry about such that when combined they lead to even more innovative solutions for problems they didn't initially realize had any overlap.

> the variation of first having everyone brainstorm alone

This would be called "thinking" by most people wouldn't it?

Calling it a variation rather than the complete opposite approach seems like people just like the word brainstorming.

Your conclusion is if you decide to call whatever works "brainstorming" then brainstorming works. If you study the specific technique that has been called brainstorming for decades it doesn't work. Like the title says.

Your normal mode of thinking involves quickly dismissing many ideas, thoughts, hypothesis, etc., which is necessary to actually get stuff done. Brainstorming isn't a good idea when you're trying to think of a way to cross a road quickly.

Brainstorming let's you freely associate by postponing your own judgement. It gets you into corners of your mind you wouldn't visit if you were actively trying to solve the problem. Of course that is also a way of 'thinking': the point is that you get into a different mindset and think in different ways.

Fair enough, is it common to describe this task as brainstorming when it's done solo?

I only really hear it used when it's a done as a group activity.

As far as I know that is common, but I would also still describe it as a group activity, even though the initial idea generation is done solo. The process could for instance be:

* Do a short exercise to prime everyone to 'let go of their default mindset' (you can leave this out if everyone is experienced in doing this, although it should usually be fun)

* Have everyone put their own ideas to paper, within a reasonable amount of time, while together in the same room, but without communicating about them

* Together, group the results: determine overlaps and outliers.

* Then discuss the results together

In the last phase, the first exercise and the resulting mindset still has some effect: people will be less protective of their specific ideas, because of the mindset that generated them, which can easily be blamed for 'silly' ideas. Some outliers can therefore be dismissed easily. On the other hand, people will more seriously consider ideas they might otherwise have dismissed as 'silly' right after they were uttered. Some outliers may prove to be very valuable and a group brainstorm wouldn't have resulted in them.

Brainstorming, as I understand the term, is used to describe this entire process of idea generation and evaluation. Not just the phase/mindset when generating ideas and not just when that happens in a group together. Perhaps this school of thought is more regionally limited than I know and this is indeed not generally known as 'brainstorming'.

It doesn't come as a surprise to me that the group dynamic of bringing lots of talented people together in the same building is superior to brainstorming. However, that seems like a false choice to me. Brainstorming is often used to solve a problem now; you don't always have time to wait for something genius to emerge (which might not even be related to your problem). Brainstorming tries to emulate that situation by having people interact more.

That said, in my experience the quality of a brainstorm session depends a lot on who are participating and/or who is facilitating. I find brainstorming works well when the problem space is clearly defined and people with different functions or backgrounds participate. As the article says on the interaction between Chomsky and Halle: “We became great friends,” Halle says. “And friends shouldn’t be shy about telling each other when they are wrong. What am I supposed to do? Not tell him he’s got a bad idea?” For this to work, people shouldn't be afraid to tell each other things that can be shot down. For friends this is natural, but for random people put together in a team it often isn't.

In a way I think brainstorming is a way to simulate the sort of discussion that would take place in a group that the article refers to as having a high Q. This means both feeling free to give your opinion, but also leaving room for discussion. Maybe brainstorming focuses a little too much on the former, though in my experience a good facilitator will also stimulate discussion. Personally I'm planning to look into the 'debating' technique the article describes as well.

> Brainstorming is often used to solve a problem now; you don't always have time to wait for something genius to emerge (which might not even be related to your problem). Brainstorming tries to emulate that situation by having people interact more.

That's not what I got out of it at all. In the U.C. Berkeley experiment, simply giving the same number of people the suggestion to "debate" caused them to have far more ideas, in the same amount of time. Brainstorming doesn't have people interact more -- it has people interact in a specific way (no criticism), even though that has since been shown to actually be harmful to developing ideas. It's not to save time.

Matches my experience. To have one good idea, you must not have one hundred terrible ideas.

It reminded me of your evaporative cooling essay. http://lesswrong.com/lw/lr/evaporative_cooling_of_group_beli...

Perhaps communities that are quick to kick out oddballs/trolls may reduce noise and distraction. However, by eliminating dissent and criticism, even invalid dissent, this newyorker article suggests that the community also might be harming creativity of comments by remaining members. Eventually creativity becomes so low that the forum members get fed up with the monotony and leave, and the forum dies?

A hundred terrible ideas would drown out a few good ideas, but a positive number of horribly wrong or even offensive ideas might be optimal for generating the most good ideas.

Not to hijack the thread, but your linked article is excellent.

This is one reason why it's important to be prejudiced in favor of tolerating dissent. Wait until substantially after it seems to you justified in ejecting a member from the group, before actually ejecting. If you get rid of the old outliers, the group position will shift, and someone else will become the oddball.

Wise words indeed.

Looks like a reinforcing feedback loop to me.

Does not match my experience: to have a good idea, I have to communicate with others and have them give me a dozen other takes on the problem, from which an optimal amalgamous solution will come up.

I would caution against taking results derived from random groupings of undergraduates, thrown together for the purpose of one classroom/study, and generalizing those to workplaces.

At workplaces, recruiting, self-selection, market forces, and management have all changed the group composition and motivations significantly.

Scott Berkun rails against Jonah Lehrer here.


I am really beginning to deeply loathe link-bait titles as they are guaranteed to introduce straw men and unnecessary arguments into the comments.

Haven't had much experience of brainstorming in a group session, but I find it immensely useful individually. Any time I'm stuck for ideas for blog posts, web-based projects, libraries etc, I timebox 20 minutes and spit as many ideas out into a buffer as I can manage. Most of the time I get junk but occasionally I discover ones worth testing.

The whole confusion seems to be a lack of information in the arguments as to what exactly they are brainstorming about. Is it a new movie/book idea or solving a programming problem, creating a software architecture, planning an invasion or just discussing how to improve a process etc... All of these are different situations and in some cases, brainstorming might work better, others are best left to individuals. You could try and guess which is good for which as a practice and you will still probably have some subjective cases where one might be better than the other. So, lets not make generalized statements please.

Usually every brainstorming session ends up, at least in my experience, with a "vote" on the ideas generated. The vote is not substantiated by anything. No need to justify.

I fail to understand how voting for stuff (which is basically the dictatorship of the mainstream opinion) is going to get the most creative solutions on the table. That defeats the whole purpose, and makes brainstorming fail to deliver on its promises.

I have never experienced voting on the outcomes of a brainstorm session and that makes no sense at all to me. It's about generating ideas and it will, hopefully, include ideas that sound nice initially, but are bad when given some more thought. There should be a round to separate the wheat from the chaff.

In my experience (again), we do have two rounds to separate them. But both of them are voting-based rounds. One vote to narrow down the 30+ ideas to like 5 or so, and a second one to reduce it to one or two :) The outcome was very minimal compared to the overall time wasted on the process (5-6 people in a room, for one hour, ending up with a not-so-interesting idea... there are better ways to spend a team's time).

'most creative' is not necessarily 'most effective' or 'most appropriate'. Look at Rube Goldberg machines.

I see your point, but since "brainstorming" is used also in "design" environments, we are not always talking about efficiency here, but simply straying away from the too obvious ideas and avoid repeating the same old stuff again. I guess it really depends what you expect from the brainstorming session.

I've used group brainstorming as an effective technique for large writing projects. Get together as a group, brainstorm an outline for each section, then assign the sections out to small writing teams.

Accomplished several fairly large projects this way. I think it works because even though people aren't necessarily brain storming for their section, they can bring valuable insight to the person who has to do the writing.

I must say all this proves is that MOST people (random selection of people) aren't GOOD at brainstorming. Having worked with highly-trained and practiced brainstormers, I think you can train yourself to be good at coming up with ideas in a group setting.

However, the "intermediate Q proximity" idea is definitely spot on.

Bob Sutton from Stanford has studied this a long time and he gives a very interesting response to this article: http://bobsutton.typepad.com/my_weblog/2006/06/brainstorming...

Isn't a contradiction to say that uncritically listening to a groups ideas doesn't stimulate creativity and then to follow that up by saying that listening to other people's ideas, even very wrong ones, increases creativity?

The way I read it, it's more of a social interaction phenomenon than whether ideas are inherently good or bad.

Something about knowing that people in the group disagree with some idea that's presented, and the reasons why they disagree, might be more important for improving your creativity than your own isolated assessment that an idea you heard is wrong.

Perhaps it doesn't work for companies and bureaucracies, but it works fantastically well for me, with a lil light music and warm lighting in the background, a blank sheet of paper (unlined) and an ink pen.

I've captured a lot of my thoughts about this here: http://www.erictang.org/process/2012/02/16/brainstorming/

I always find it funny when people "disprove" stuff I use successfully every day with my friends.

I always find it funny when people believe anecdotes over actual science.

On-topic: these are the kinds of results from psychology that most people here should hopefully know already.

And I always find it funny when people don't actually read the article, because it doesn't denounce brainstorming at all, just traditional 'group brainstorming and immediately taking the results for granted'.

I always find it funny when people consider my "anecdotes" to be self-evidently delusional and utterly useless.

Guess I should just tell dozens of people that the years of frequent brainstorming we've engaged in has been a complete waste of time. Science says so!

That is not what dochtman is saying.

Why do you attribute success specifically to "brainstorming" and not to some other element of the process? e.g. sharing ideas and having a conversation with your friends in a fun and low-risk atmosphere?

I think the constant brainstorming over the years has had some kind of training effect on our brains, so that we are much better at coming up with ideas than before.

Which of course proves nothing.

Applications are open for YC Winter 2020

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact