A Manager’s Guide to Human Irrationalities

Dan Ariely is one of the world’s foremost experts in behavioral economics, the study of how people truly behave when making business and financial decisions. His research has proven time and again that we all are not the sensible decision makers that we might like to think we are, and his recent book on the topic, “Predictably Irrational: The Hidden Forces That Shape Our Decisions” (HarperCollins, 2008), debuted on The New York Times Best Seller list when it was released earlier this year. Ariely is a professor of behavioral economics at Duke University and former Alfred P. Sloan Professor of Behavioral Economics at MIT, and through his Web site (www.predictablyirrational.com) he provides updates of his fascinating experiments and invites readers to participate in them.

Ariely’s insights should make executives think twice about the wisdom of the decisions they regularly make — as well as the inner processes they rely on to make those decisions. Why, for example, will managers veto a 10% cost increase for a $1 million project while thinking nothing of a 1% overrun on a $10 million budget — even though the actual amount is the same? Why will they often agonize trying to choose between two close candidates for a job when they’re frequently better off just flipping a coin? Why do they try to dominate a market when, in fact, a product monopoly sometimes leads only to consumer confusion, resulting in slow sales?

Recently, Ariely sat down with Alden M. Hayashi, a senior editor with MIT Sloan Management Review. Their conversation begins with a discussion of the implications of behavioral economics for product pricing. Specifically, Ariely explains why a product’s price must take into account various irrational human behaviors, including the principles of anchoring (when the initial price of something has undue influence on our future decisions) and relativity (when we infer the value of a product from the prices of similar offerings). The following interview has been edited for length and clarity.

How to Price Products

SMR: Your research suggests that, when selling a new product, companies should always compare it with something that the customer is already familiar with, even if the product is so novel that there really isn’t something similar on the market.

Ariely: Absolutely, for two reasons. First, without such comparison, the “space” for a new product in people’s minds is ill-defined, making it very hard for people to figure out how to place a value on something. We just don’t know how to value things in isolation. The second thing is that we are mainly creatures of habit, and decisions are actually quite tough. How many times a day do we really want to contemplate buying something by analyzing everything, thinking about the opportunity cost and so on? So we rely on our past decisions, including comparisons to other products.

SMR: In other words, we’re essentially lazy and we’re not going to do a rational, thoughtful analysis for every product we buy.

Ariely: I am not sure I like the word “lazy,” but, in general, thinking is difficult, and we try not to overdo it. Take, for example, TiVo. What’s the value of TiVo for you? How do you compute that? Do you take into account how many minutes of commercials you’re saving, multiply that by your income per hour, deduct from it the breaks you get to go to the bathroom and take a snack, and so on? That would be very, very hard to do. So instead, you rely on your past impressions and use that to infer value. If companies want to understand how people make decisions about their products, they have to take such decision processes into account.

SMR: So, of course, the trick for companies is to figure out what to compare their new product to.

Ariely: Imagine two universes. In the first, TiVo is compared to a VCR and is introduced at $200. In the second, it’s compared to a computer and is introduced at $1,000. Then imagine that in both worlds the price goes to $500. In the first universe, people will presumably be outraged and nobody will buy it. In the second, people will think it’s a great deal. And that’s why the principle of relativity is so important, especially for new products — because we just have such a hard time computing the real value of products and services.

SMR: And in the comparison you can actually define the product.

Ariely: Yes, and the thing you need to understand is that that definition will last for a long, long time. We can speculate, for example, about the new iPhone. I haven’t talked to Steve Jobs about this, but let me speculate about what he did. He put this iPhone out at $600 and immediately reduced it to $400. Now, this could have been a mistake, but it also could have been a smart trick because the question to the consumer at that time was, what is the comparison price? All of a sudden, something can look like a great deal at $400 when it was $600 just a few weeks earlier. If Apple had introduced the iPhone at $400, it would have been difficult to figure if the iPhone was worth it. But the initial $600 price and then the $400 helped, I think, create a very high price point in people’s minds. Relative to the $600 price, the $400 price seemed like a deal. And now that the iPhone is being offered at $200, it looks like a fantastic deal because we still have these very high prices sticking in our minds.

SMR: Many outsiders have assumed that Apple made a mistake in initially offering the iPhone at $600 because there were so many outraged customers who had paid $600 only to see the price drop to $400.

Ariely: We can argue whether Steve Jobs made a mistake or whether he is very clever. I think it’s interesting to think about it in this way: You introduce something at a very high price point, you create a value statement, and then you say, “Sorry, I didn’t mean it.” But still the high price point is stuck in people’s minds; everybody remembers it was $600.

SMR: On the other hand, what about companies that set the initial price of something too low, even offering a product or service for free in order to encourage people to use it? Isn’t that why so many online publishers are facing such great difficulties, because they initially offered their content for free and then consumers couldn’t move past that anchoring point?

Ariely: The truth of the matter is that it’s very hard to realize the value of something even after you’ve used it. Say you use e-mail. How valuable is it to you? Sure, if something is free then people will start using it. But what companies don’t realize is that the mapping of utility to money is very difficult. People won’t say, “This is so great. I’ll pay $20 for it.” Instead, they’ll say, “I used it for free all along and now you’re charging me? I’m not interested.”

Price As a Product Feature

SMR: Many managers determine pricing by figuring out costs plus a certain amount of markup. But you’re saying that pricing is much more complex than that. In fact, pricing has the power to affect a customer’s perceived value of products.

Ariely: Yes, price signals quality to people in a nondeliberative, conscious way.

SMR: Is price like a product feature then?

Ariely: That’s a good way to think about it. And because price is a product feature, we all make general inferences about a product from its price. If something is cheap we expect it to be cheap, and the amazing thing is that we can even experience it as cheap. As a side point I should mention that we did some experiments in which we gave people cheaper painkillers, and guess what — relative to people who got the expensive ones they felt more pain. The key thing here is that there are many experiences in our lives that are ambiguous, and no product is perfect. If you got a discount for something, every time something un-perfect happened you might say, “Oh, it’s because the product was cheap.” And you would pay more attention to those un-perfect things, making the whole experience a self-fulfilling prophecy.

SMR: So managers need to consider how they want their product to be perceived when setting its price.

Ariely: If you’re the cheap product in the market and you want that image, there’s no problem. But if you want a different image, you should consider very carefully your pricing strategies. Let me return to the Apple example. I think the new price for the iPhone is too cheap. I have one and the touch screen is incredible. But all of a sudden you can buy the phone for $200 and I think that this price will diminish the value of this amazing touch screen.

SMR: So the cheaper price will affect how consumers experience the iPhone’s touch screen?

Ariely: I used to think of that touch screen as an amazingly expensive feature. Now I think of it as a much cheaper feature. The interesting thing here is that, when people don’t think about it, a pricing discount translates into lower efficacy. But if we get them to think about it carefully, it doesn’t have that effect.

SMR: So with the iPhone, because Apple didn’t make greater efforts to explain the lower price, consumers will somehow think the quality is less than that of the original product?

Ariely: That’s right, at least in some very primitive, nondeliberative sense. People would not say, “Apple made this worse,” but there’s a chance that they would think of it as second best. And maybe they will not handle it as carefully, and as a consequence maybe it will break faster.

SMR: Should Apple have better explained the lower pricing, for example, by saying that the manufacturing cost has decreased because of the technology learning curve, all the while stating that the quality of the product is the same? In other words, you want customers to consciously think of why a product is now being offered at a discount.

Ariely: That’s right.

Too Few (Versus Too Many) Choices

SMR: The topic of relativity raises the question: is having a monopoly necessarily the best thing in the world. That is, does Coke need Pepsi and Pepsi need Coke because they provide that frame of reference for each other?

Ariely: When your product is by itself, consumers will have difficulty evaluating it, and as a consequence, fewer people will buy it. But when there are two products, then people can compare, and your market share can potentially increase.

SMR: In other words, in a monopoly you might have 100% of the market but the size of that total pie might be small. When you have competitors, you could end up with just a slice of the pie but that piece might be larger than the entire pie of a monopoly.

Ariely: Yes, exactly. There’s an interesting story here of a company that used to sell birth-control pills in third-world countries. It had very little success with one of its products, and so it introduced another one. But it didn’t take the old one off the market, so it basically had a pill that was too cheap and then introduced another pill with an American-sounding name that was a bit more expensive. That second product did better, and over time it also pulled up the sales of the first pill that hadn’t been doing so well. So it increased the total market. You can think about many products like that.

SMR: So would you agree with a GM strategy in which you create all these product brands so that consumers can compare Buicks with Chevrolets and Cadillacs and Pontiacs and so on?

Ariely: I am not sure about that, because what you don’t want to do is create too many choices and get consumers confused.

SMR: Is there a rule of thumb to help companies walk that fine line between offering customers too few choices versus too many? Because, as you mentioned, when there are too many options, then you end up with the “paradox of choice,” and people get confused and often make bad decisions.

Ariely: Sheena Iyengar and Mark Lepper [researchers at Columbia and Stanford, respectively] did this very nice study. They went to a grocery store and set up a booth in which they presented people with either six jams or 24 jams. Of course, many more people got attracted to the bigger selection: more jams, more color, more activity. But then people were given a coupon that was good for any jam in the store. [Note: The store itself had more than 300 jams.] What they found was that, for the people who saw just six jams, 30% of them ended up buying jam. For the people who saw 24 jams, only 3% of them bought jam. Now, jams are not a complex product — it’s just fruit and sugar — but it turns out in some important way that 24 are too many.

SMR: In other words, people are easily confused?

Ariely: But there are two important factors. The first is whether there’s a default, so that one doesn’t really have to choose. When people enter the store and have “jam” on their shopping list, they’ll go ahead and buy jam whether they see six or 24 jams. But confusion is very important when there’s a default of not doing something, such as when an item is not on your shopping list. The second factor is the amount of effort that people want to put into making a choice. It turns out that jams just aren’t that important to most of us. But if you’re dealing with something important, people will pay more attention, and they’ll be more likely to persist despite their confusion. So you can’t make a blanket statement and say that eight products are too many. Eight jams might be too many but eight cars might not be. It all depends on the defaults and on people’s motivation.

SMR: And I suppose another factor is the ease of comparison. When cell phone providers first came out with their service plans, it was almost impossible to compare them. In that case, having just three plans was too many.

Ariely: Yes, call it the “complexity of the choice.” How difficult is it to compare products? How much do you need to know to compare them? How important is the decision, and what is the default? All of those factors together will determine the complexity of a decision. So, for example, when choosing between pens, it’s great to have different options until, say, you reach 10, and then it becomes worse. With other products, like cars, the number will be different.

Hiring the Right People

SMR: You’re a huge proponent of experimentation to test our assumptions because, as you’ve so often shown in the past, our assumptions can be totally wrong. But in business, managers regularly make decisions based on beliefs that haven’t really been tested with any sort of rigor.

Ariely: I’ll give you a good example: the interviewing process used to hire employees. A few years ago, I was walking around the corridors of MIT and Bain Consulting was offering free pizza, so I went in. So there I was with MBA students, and we heard a lecture about the Bain interviewing process. And then I asked, “How valid is your test — the interviewing process?” And the guy said, “We grew 20% last year.” I said “That’s wonderful news, but how valid is your test?” He said, “McKinsey does the same thing.” Again, I said, “I’m very happy for you, but that is not an answer.”

SMR: It probably didn’t cross the guy’s mind that Bain’s interviewing process might be invalid.

Ariely: I told him, “If you hired me to give you advice I would say that the only resource you have is your people. And you might as well figure out how to sort the good people from the bad ones. In terms of your executives’ time, it costs you tens of thousands of dollars to interview somebody. So you should at least check to see how good your interviewing is. If you don’t, then you should save the time and effort when you come to MIT and just assume that everybody is good.” The guy got really upset with me, but the point I was trying to make was actually very basic: To really test how good their interviews are, they must check their prediction ability and, even better, they should from time to time hire people that they don’t think will work out well.

SMR: So your advice for Bain would be to invest in running an experiment to test how good its interviewing process is.

Ariely: Yes, I think they have to start being more rigorous and systematic in their approach. Years ago, Daniel Kahneman, one of the fathers of behavioral economics, went to the Israeli air force and actually convinced them to accept for pilot training some of the people who failed the pilot test. Now think about that: It’s incredibly expensive to put a pilot through training, and every year they take a few people who didn’t do well on the test and put them through that training. It’s exactly like hiring the people you think will not do well. In one sense, it’s very counterintuitive. But in another, it’s the only way for you to ever improve your test.

SMR: That’s a tough sell for companies — having them purposely hire people whom they think will do poorly — because bad employees can wreak so much havoc. But your point is: How else will you learn unless you experiment?

Ariely: That’s right. Imagine that you’re in Las Vegas and there are two types of slot machines. You’ve played with the first type the whole night, so you know how much money it gives, but you’ve played with the second type only a few times. So you don’t know if the first type is better than the second. If you’re going to be in Las Vegas for weeks, isn’t it better to lose some money in the short term to figure out which type of slot machine is better? Presumably, Bain is going to be around for a long time. Shouldn’t it figure out how to optimize its interviewing for the long term?

SMR: Well, Bain may be around for a long time, but its current executives might not. They might see themselves in a relatively short time frame; so to them, the investment in experimentation might not be worth it.

Ariely: Generally speaking, that’s part of the problem in corporate America. People have intuitions, which are sometimes wrong, and there is no requirement to put these assumptions to a rigorous test. Now, experimentation is only good for the long term. It’s a short-term loss for a long-term gain. I think many companies have a short-term perspective, and that’s why they don’t invest in learning. But the other thing to understand is that current hiring practices are awful; many companies might be better off going to a university like MIT and hiring students at random.

Toward Better Group Decisions

SMR: With respect to group decision making, I found your research on how people order at restaurants fascinating. You showed that when a waiter takes orders sequentially from people in a group, after the first person orders a particular entree, the others in the party will be less likely to order that dish because they don’t want to be seen as copycats. I wonder if that happens a lot in business meetings in which people give their opinions about a proposal by going sequentially around a table.

Ariely: That’s right. It’s unbelievable how everybody feels the need to say something in meetings, even if it’s only very slightly different from what someone else said. A few years ago, we created a piece of software we called “anti-groupware” in the MIT Media Lab. We wanted to investigate why, when you take good people and put them in a room, they make bad decisions as a group. And the question is: How can you use technology to prevent that? How can you, for example, eliminate excessive conformity and obedience?

SMR: What does the software do?

Ariely: We had multiple versions, and one of the problems we investigated was the presence of an authority figure: When your boss says, “I think we should do X,” how likely are you to vote for a different option? Another thing we investigated was conformity: If five people say the same thing, will you be more likely to follow suit? So we created basic technological tools for voting. People could all vote anonymously and we investigated different mechanisms. For example, I might be able to allocate my vote to you if I think you’re a big expert on the particular topic at hand, but you wouldn’t know that I had allocated my vote to you, etc.

SMR: Did the software allow changes in the rules so that, instead of needing a majority of votes, you need three-quarters?

Ariely: Exactly, yes. We also created situations in which people could vote, see the outcome and then change their vote, and this could occur over a few rounds. And we also tried to get people to figure out how much they cared about a topic. For example, we could present them with 10 things they needed to make decisions about and give them 100 points, and then they would have to allocate those points among the 10 issues. [So if someone felt very strongly about issue six but cared nothing about the other issues, he could allocate all 100 of his points to issue six.]

SMR: So the software essentially tries to help people avoid the different traps of group decision making.

Ariely: Groups are important to share information, but when people need to make decisions, they’re not good. So we basically created a way for people to make better decisions in a group, and we were hoping that some companies would use it, but nobody was willing to. We thought the software was very good, but no company ever wanted to try it with us in a real experiment.

SMR: Why the reluctance?

Ariely: I think group decision making is one of those things that companies have huge confidence in. Groups are not helpful in getting people to make better decisions, but they’re helpful in getting people to feel more confident about the decision they’ve made. So people have this intuition that group decision making is just fantastic when it’s actually not.

Managing Social Norms

SMR: Another area in which some companies have trouble is managing unspoken social norms.

Ariely: People have fantastic intuitions about this in their personal lives. Consider this: Imagine you’re on a date with a woman. You go to a movie and dinner, and as you’re escorting her home you lean over to kiss her and say, “Interestingly, I’ve spent $150 on the date so far.” You haven’t actually said anything that she doesn’t know. The prices are printed on the movie tickets; menus have prices. But the moment you mention what you’ve spent you’ve altered the relationship from a social to a financial one and in the process have cheapened it. And I think that while everybody understands this in their social lives, they don’t understand it in the employer-employee relationship. And I think it’s a crucial lesson.

SMR: But it’s difficult because companies have to do this delicate dance. On the one hand, they pay employees to do work, and so the relationship is basically an economic one governed by the market. On the other hand, they want to motivate employees to go beyond the call of duty so they want workers to feel that the employer-employee relationship is more than just economic.

Ariely: Part of the trick is obfuscation. If I pay you per hour or I pay you per word that you write, it’s very clear that our relationship is economic. But if I pay you monthly or yearly, then it’s not as clear. And if I don’t pay you with cash in an envelope but simply move the money to your bank account, then the relationship is even less direct. And what if I also give you health care benefits and so on? By the way, when companies give employees paychecks, they’ll often state how much money they spent on health care benefits. They say, “Your salary is $80,000 but we spent another $12,000 a year on your health care benefits.” But that’s just telling you that your total compensation is $92,000. And now if another company offered you $93,000, then it’s very simple for you to see that you could be earning more elsewhere.

SMR: And your motivation in your current job would drop.

Ariely: Imagine that I was your employer and I offered you a thousand dollars in cash or a vacation to the Bahamas. Now presumably you would be better off taking the cash, because you might want to go to the North Carolina shore instead of to the Bahamas. Or you could skip the vacation and buy a new iPod. So you could optimize the reward you receive by taking the cash. But under which condition — the cash or the Bahamas vacation — would you be more receptive to my asking you next Monday to stay longer in the office? Under which condition would you spend less time browsing the job ads on Monster.com? Under which condition would you be more loyal? I think it’s the vacation, and that’s the trick: There are things that are financially inefficient but are socially and motivationally efficient. By the way, isn’t it amazing that the thing that Microsoft fears most is an organization of people who work for free? I mean, just think about that.

SMR: So do you think that businesses will move toward organizations like Linux that are more community based than market driven?

Ariely: I think it’s about the meaning that people derive from their labor. Work is incredibly important to people — the value and meaning they get from it.

Let me tell you about an experiment in which we asked people to build Lego robots. We paid them $3 for the first robot, $2.70 for the next, $2.40 for the next and so on, and people could decide when they wanted to stop. In essence we were asking, what is the marginal value of your labor? For one group of people, when they finished a robot we’d ask, “Do you want another?” and if they did, we’d give them the parts and take the completed robot and put it in a box. For the other group of people, when they finished a robot we’d take it and disassemble it in front of their eyes, and then we’d give it back to them to rebuild. We called this a Sisyphic condition, after Sisyphus. In essence, both groups of people were doing the same thing: They were building robots for money. But the reality was that the people who saw their work destroyed hated the task. The other people were able to find small meaning in even something as trivial as building Lego robots, and that sustained them. Now, that for me is an incredible insight — to understand how little meaning you need to add to people’s work to increase their motivation. You want employees to feel that they’re doing something important, that they’re making a contribution, and from that perspective Linux has a much better story than Microsoft.