Polling random people can be "incredibly useful as a persuasive attempt to tell people what to do," but not really as a way to find out information.<br /><br /> <br />Question: What kinds of things to companies do that are irrational?Dan Ariely: One of the things that I think companies do <br />to a large degree that makes very little sense, is the use of focus <br />groups. So think about what a strange idea is that we take 10 people <br />who know basically nothing about your project and you put them in the <br />room and you let them talk for a while and then you take the, whatever <br />they came up with, as a consequence of these two hours of random <br />thinking and you base your strategy on it, to a large degree. And I <br />don't want to say that focus groups are always useless and also <br />uninformative, but I think that taking this data and then relying [on] <br />it as extensively as companies do, I think it's crazy in many ways. <br />And the reason I think companies do it is because when these people <br />create a sentence or an idea, it's very easy to say, "Joe, focus group <br />17, said this," and it can help you create and formulate an idea around <br />it. When if you said "87% of the people said X," it just doesn't have a <br /> face, it doesn't have a desire, it's just not as concrete and <br />therefore, people are not as excited about the notion. So I think the <br />focus group is incredibly useful as a persuasive attempt to tell people <br />what to do, but as a way to find out information, it's not as useful as <br />people think it is.The other thing that really puzzles me is <br />that companies, how little they understand about how incentives really <br />work. So the biggest expenditure for any company is salaries. And you <br />ask people, "What do you know about the relationship between salaries <br />and performance?" The answer is nothing. Right? And the question is, <br />why don't they study anything? Why don't they study? Why don't they <br />try different performance, different incentive level and see how it <br />influences performance. And they often say, well, there are all kind of <br /> laws and if we have, pay some people one way and some people another <br />way, we might be sued. But in the last two years, with all these <br />questions about bonuses and so on, I've gone to many companies and said, <br /> "Why don't we at least do some surveys, why don't we at least see how <br />happy people are in the months where they give them, you give them the <br />bonus and how productive they are?" And the answer I got 100% of the <br />time is that people are really miserable in bonus season and because of <br />that, they don't want to ask them any more questions. And I said, "If <br />people are miserable in bonus season, shouldn't it mean that you should <br />study it, understand it, and try to prevent it for next year?" They <br />say, "Well, well, not this year, this year was really miserable, maybe <br />next year, come back to us next year." And if you think about it, it's <br />really incredible.Now, how do companies decide about <br />compensation? They look at what other companies are doing and try to <br />equate compensation to what other companies are doing it, but it's a <br />place of the blind leading the blind, right? And then if we pay 5% <br />more, other people would start doing it, but the real question is, how <br />do you pay and how do you get people to care about the work and become <br />more productive and be happier at their work? And the answer is not <br />simply that more money is better. More money can be part of the <br />equation, but it's also the question of how do you give that money? Do <br />you give it as bonus, do you give it as a fixed salary, do you give it <br />as part of the benefits? Maybe you give benefits for the gym. Do you <br />send people on vacation to the Bahamas? And it turns out that there are <br /> ways to use money that is economically less efficient, but actually get <br /> people to be more motivated, care more, and actually become more <br />productive.Recorded on June 1, 2010Interviewed by David Hirschman
