The following is a conversation between Uri Gneezy, Professor of Economics and Strategy & Chair in Behavioral Economics at UC, San Diego Rady School of Management, and author of Mixed Signals: How Incentives Really Work, and Denver Frederick, the Host of The Business of Giving.
Denver: Incentives send powerful signals that aim to influence behavior, but often there’s a conflict between what we say and what we do in response to these incentives. The result, mixed signals. And that is the title of my next guest’s new book. He is Uri Gneezy, a professor of Economics and Strategy and Chair in Behavioral Economics at UC, San Diego Rady School of Management, and author of Mixed Signals: How Incentives Really Work.
Welcome to The Business of Giving, Uri.
Uri: Thank you. Great to be here.
Denver: I believe it was Warren Buffett’s partner, Charlie Munger, who said, “Show me the incentive, and I’ll show you the outcome.” Do we even realize the profound impact that incentives have on our behavior?
Uri: We should. I think that many times we don’t, but we should, and that’s what brought me to study incentives because I think it’s always fascinating to understand them.
“…I was always curious about why people do what they do. And that’s true for other people, and it’s also true about myself. So very often, I found myself doing stuff that they didn’t understand why I did it. And then I tried to trace it back to the incentives that they have. An incentive is definitely not just money. Money is just one example of incentive. There are many others.”
Denver: Yeah, well, let me ask you that. What did bring you to this subject? This is not somebody that you’d wake up, when you’re seven years old and say, Hey, I want to study incentives and become an expert on it because they are so important, but they are perhaps under-reported. What drew you to this field?
Uri: So I, actually, I think that I did. I didn’t know that it’s called incentives at the time, but I was always curious about why people do what they do. And that’s true for other people, and it’s also true about myself. So very often, I found myself doing stuff that they didn’t understand why I did it. And then I tried to trace it back to the incentives that they have.
An incentive is definitely not just money. Money is just one example of incentive. There are many others.
Denver: Mm-hmm. Mm-hmm. A real famous example of an incentive that maybe had some unintended consequences was the daycare experiment. Why don’t you tell listeners about that? It is just a great way to set the grounding for this.
Uri: Sure. It’s one of my favorites. So I used to study… I was interested in how incentives can backfire. With Aldo Rustichini, we thought about it, and then my girls, my two daughters were in a daycare in a suburb of Tel Aviv. That’s where we lived at the time. My wife and I went to lunch in Tel Aviv very often, and then one day we were late, we needed to pick them up by 4:00 p.m. And I drove like crazy because that’s what you do; you need to be there on time.
Uri: Then the principal introduced a small fine of $3 for parents that come late to pick up their kids. Again, we were in Tel Aviv. Surprise, surprise! Again, there was traffic. This time, I didn’t drive like crazy because I’m not going to risk my life for $3, right? It completely changed the meaning of: what does it mean to be there on time?
Before, that’s what you need to do, and now it was… they put a price on it, and the price was very low. So, yeah, let’s go with it. Let’s drive a bit more safely and be there a bit late picking up our kids, pay $3, and all is good. And that’s one of the things that led me to thinking about the information that incentives send. Before that, I didn’t know how bad it was. Now that principal told me: it’s $3 bad. Well, then it’s okay.
Denver: Yeah, yeah. Unintended consequences. As you say, we sometimes don’t think about it. We just think really one step ahead, and very rarely do we think four steps ahead because then that’s what really is going to change the behavior.
Well, what are some of the common mistakes that organizations make when implementing incentive programs?
Uri: So we start from … maybe where we should end, but I think that that’s the interesting part. We need to use common sense. So I used to… my line on this is that every company has a CEO and CFO and all the nice functions.
They don’t have a common sense officer, someone that will look at the incentives that they design and say, Wait a minute, wait a minute! If you introduce a $3 fine, maybe you’ll collect some money, but people are not going to stop coming. You want them to stop coming? Make it a $10 a minute fine. So it’s not that fines don’t work.
If they’ll execute any parent who comes late to pick up their kids, that will work, right? But sometimes you cannot do that. You can only give small fines in this case. Make sure that you understand what, how is it going to change the perception of people. And the common sense officer, like that someone that will look at the incentive and will say,” Just a second, we are sending the wrong signal over here,” is really important.
“So there is one thing that you tell your kid or anyone else, and a different thing that you do when incentives present themselves, and you need to be careful about this.“
Denver: Yeah. Yeah. Also, I think sometimes incentives can send awry some of the great declarations we make as you did, Dad, with your daughter down in Disney World… when you have pressed one thing, but an incentive pops up, and you behave in a different way. Tell listeners that story.
Uri: Exactly. That was actually my son, he’s a bit younger.
Denver: Son. Mm-hmm.
Uri: He just turned three and started talking. That’s what kids do.
Uri: Which is great, right? But they also start lying because, again, that’s what kids do when they start talking. And we told him, “Look Ron, only bad people lie, you shouldn’t lie.” Then we went to Disney World, like you said, and the cashier, the nice cashier asked me: How old is the kid? When I saw the sign that said, “Under three years old, it’s free. Over three years old, it’s $117.”
So I said he is almost three, which wasn’t a lie because he was almost three just from the wrong side, right?
Uri: And that was fine. I didn’t pay for him. He went in, and then he pulled my sleeve a few minutes later and asked me, “Daddy, daddy, you told me that only bad people lie, and you just did.” So there is one thing that you tell your kid or anyone else, and a different thing that you do when incentives present themselves, and you need to be careful about this.
Denver: That’s a great point. I think it’s very true with values in an organization as well. Values are easy to profess, but when you have to turn down an opportunity or money to be able to adhere to those values, you’ll really find whether an organization lives by those values, or whether they’re just on a chalkboard.
Let’s go back to Tel Aviv a little bit. Oh, I’m sorry. Go ahead.
Uri: No, that’s exactly the point. That it’s one thing to say that you care about whatever honesty you care about. We might talk about creativity or quality or whatever, but then incentivize something very different.
Denver: Since you brought up the streets of Tel Aviv and how hectic they can be, let’s talk about another incentive, or a pair of them, having to do with bus drivers.
Uri: Right. So in the street, I visit Tel Aviv lot. That’s where I’m from. And now we live in San Diego, but I come here quite often. And on the street next to our house, there is a very interesting experiment. There are two types of, you can call it bus drivers. One of them is paid per hour. These drivers are very polite. They stop at the station, let the people go in; everything is good. And the other ones drive like crazy.
And the difference is that one is paid per hour; those are the polite, and the others are paid per rider. And if you’re a paid per rider, you really have incentives to pick up as many as you can. They often start driving before the passenger actually sits down and they’re much more… they need to make money. Exactly.
Denver: Yeah. Sounds like some airlines.
Uri: Yeah, exactly. So basically the company can tell them as much as they want, that quality is important in customer service, and safety is the number one priority, whatever. But then if they pay them per rider, it’s not going to work this way. Now the way to fix it is: think about Uber. The drivers have incentives to be fast, which is great. It’s not something bad.
They have incentives to think where to go, when to go. That’s good that the company wants it, but they added another layer of incentives, which is the rating. If a driver will not drive well, will be too aggressive, the car will be dirty or whatever, I’ll give them one or maybe even complain to the company. And that’s going to affect them, negatively affect them.
So now, by adding another layer of incentives, you basically were able to get the good thing from both worlds. You get drivers that really think strategically and want to pick up as many riders, but also need to think about the quality of the ride.
“…distinguish between two types of signals that we have. One is public, the social signal that I give to my neighbors, to my coworkers, even to strangers. I want them to see me in a good light. And there is a signal that I want to send to myself. This one is a bit harder to understand, but very often we don’t really know how good we are, and we look at our actions and learn from them how good we are.”
Denver: Yeah. Yeah. That’s really brilliant. And I guess that’s where the common sense officer comes in again, when you have to begin to think of incentives and just say, okay, and then just go that next step and say, but if we do this, then we’ll get, as you just said, the best of both worlds. And that would make so much sense.
Denver: You conducted a field experiment in a Pakistani buffet restaurant, which is something which I think a lot of listeners are familiar with, which is “pay what you want.” Share the insights you had from that.
Uri: Right. So “pay what you want” sounds crazy, but it works in many cases. When I care about the institution, when I want to do well with some values that they have, I’ll pay something. And this day, the Pakistanian restaurant experiment is actually interesting because it can distinguish between two types of signals that we have.
One is public, the social signal that I give to my neighbors, to my coworkers, even to strangers. I want them to see me in a good light. And there is a signal that I want to send to myself. This one is a bit harder to understand, but very often we don’t really know how good we are, and we look at our actions and learn from them how good we are.
Uri: So this… in this pay-what-you-want restaurant, the normal thing was that when you’re done eating, you paid the waiter, and you paid as much as you want. And then people paid something; they paid like six euros on average. And the reason was probably that they didn’t want to look like jerks, right? You went there, you ate; you need to pay something.
Uri: Now, we contrast that with instead of paying the waiter, they put the money in an envelope and put the envelope in a box at the exit of the restaurant. Now, no one sees how much they put, and they can put zero, and no one would blame them about it because no one will know. But now comes the self-signaling. If I put zero, I’ll feel really bad about it.
If I don’t tip someone, I feel bad about it. I need to tip. And in the U.S., we got into you tip 15%, 20% as a routine, but in many places you don’t. But I feel good when I tip a server that did a good job. And what we found was that when they put the money in an envelope, they actually paid a bit more even. And that shows the power of self-signaling.
And it also shows that if there is a social-signaling, we call it “crowd out the self-signaling,” because if you give such that someone else sees you, you don’t get the benefit of feeling good about yourself because you also think: Well, I’m doing it because I don’t want the waiter to think badly about me. So that’s the power of self-signaling versus social-signaling.
Denver: Oh, that’s really interesting. Well, let’s take that because this is a show that really caters to nonprofit organizations. So I’d be curious as to how you see self-signaling and social-signaling playing in, let’s say, blood donations.
Uri: Blood donation. Blood donation is great. So if I may, let me give you another example. Imagine that in the morning you live in a cold place. You see your neighbor going to the recycle center with a large bag filled with a hundred soda cans. You’re probably going to say, Well, she’s great. She could have trashed everything, but she decided to do that.
Denver: Good woman.
Uri: Exactly. That’s what she’s going to feel about herself as well. That’s the self-signaling. Now imagine that the city decides to pay 10 cents per soda can that is recycled. Now you see her. Now you’re not going to say “good woman,” you’re going to say “cheap woman” because…
Denver: That’s correct.
Uri: …for $10, I mean… and she might not feel that good about herself. So that’s the background.
Now think about blood donation in that spirit. The blood donation is really interesting because many people like us donate blood. All donation, no money, but then up the chain, you get the blood bank is selling it to hospitals. There are billions of dollars changing hands with blood, but still we don’t pay the donors.
And the reason is similar to the recycling example that says that you go to donate blood; you’ll feel good about yourself, right? You spend a couple of hours; someone sticks a needle into your arm, but that’s actually good because you feel that you contributed. It wasn’t easy. If it was easy, you wouldn’t feel that good about yourself.
Uri: You might even brag to your friends at work and say, “Well, sorry I missed the morning session. I donated blood.” That’s great. Now imagine that the blood bank is paying you $50 for doing this. Well, for $50, I’m not going to waste half a day and suffer, right? So that’s bad. And I’m definitely not going to show off to my friends about it.
Now, what a blood bank could do, which many of them actually do, is give you, say, a coffee mug with the logo of the blood bank. Now, every morning… that cost, by the way, much less than $50, but now every morning when you drink coffee, you’ll say, Well, Denver, you’re a good guy. You donated blood. It will remind you.
And you don’t have to tell your friends. You can bring the coffee mug to the meeting and put it on the desk. Everyone will see that you’re a good person, right? So it sends the signals that you want. And by the way, if you want to stick with money, you can still pay people for donating blood, but you need to call it something else. You can tell them, Look, that’s for the Uber that you took here. That’s fine. That…
Denver: You know, I… yeah, go ahead.
Uri: That will not crowd out the signaling because, yeah, clearly, they are covering my cost. They’re reimbursing me for this. That’s fine. So really, the labeling and the type of incentive is really important.
Denver: Yeah, really, really interesting. I can see tomorrow maybe going to work with my blood bank mug in one hand and my National Public Radio bag in the other, so everybody knows that this guy is not a good guy, he’s just better than… he’s unbelievable.
Denver: Uri, what happens…what’s the impact when you remove an incentive that you have been using to try to shape behavior? And, of course, you think of kids… you think of your children, and maybe every time your child finishes their homework, they get a bowl of ice cream. And you do that for a while, and then maybe one day you decide to stop. What is the impact of removing an incentive?
Uri: In general, in education. So I’m a professor. I have a very relaxed life. No one is mad at me, usually, until I try talking about incentives with educators. And then for good reasons, they are really upset because they say, “Look, we don’t want you to do the homework. We want you to learn to enjoy, to have intrinsic motivation, to learn and to do all the good stuff.”
And if I’ll pay you, just like you said, if I’ll give you money now; maybe you’ll do the homework, but you won’t do it in the long run. And that’s a very touchy situation. But there are ways to look at it. So my mother had a great solution to this. As a kid, I used to read these trashy Western books that everyone asked my mother, Why do you let him read this? And she said, Well, as long as he enjoys it, I don’t care.
If he’ll learn how to enjoy reading, he will continue reading later. And you can think about Harry Potter that had this effect on all generations of kids that started reading because of that. Now, in certain schools, very often, at least when I was a kid, I had to read Dostoevsky. He is the best. But as a 16-year-old, I couldn’t care less about a thousand-word book, right?
Denver: For sure.
Uri: Now, I’m 56. Now, I can read it. So I think that in many cases, we ignore the fact that we should find some activity that people enjoy doing. And that might get them to enjoy reading after that in this example. And for that, we might use incentives. So if I can use incentives in a way that will actually increase your intrinsic motivation to do something….
So instead of incentivizing you by punishment, for example, if you don’t do your homework and read Dostoevsky, I can tell you, “Look, you need for homework… you need to read a book and tell us about it later on. Choose the book.” I think that that way, incentives can actually help if I incentivize you to do something that you actually learn to enjoy.
“Don’t punish people for trying, because if you punish people for trying, they’re not going to be creative.”
Denver: Yeah, yeah, yeah. Instill that love of reading, and eventually it’s going to have a good long-term impact. Don’t frustrate them because then they won’t read anything at all.
How important is it for organizations to line up their incentives in sort of an intelligent way? Because I could see that you could have different incentives within an organization, and even if each of them was thought out, there could be mixed signals in the kind of incentives that you’re giving. How does a company or an organization need to think to make sure that the kind of place you want to have the incentives are all aligned, that you’re sending out a singular message?
Uri: So, imagine a company that tells its workers to be creative. Creativity is really important in many organizations, and that’s usually the message: Be Creative! But then they punish you if you fail. So punishment could be no promotion, could be no bonus, or even firing you. Now, being creative means increasing the variance, right?
The more creative I am, the higher is the chance that I’ll get something really interesting. That’s why companies, organizations try to incentivize it, but it could also end up with a failure, more likely to end up with a failure. The variance is just larger. Now, the company shouldn’t punish you for failing. The company should look: What happened? Why did it happen?
And once you understand why did it happen… if it happened because I was lazy or because my idea was really bad, then punish me. But if it happened very often because my intuition was wrong, or something in the way people reacted to it didn’t work, or the physics didn’t work if you are in R&D, just learn from it and make sure that you move on. Don’t punish people for trying, because if you punish people for trying, they’re not going to be creative.
“Very often, people are really creative at gaining the incentives. Very often, people are going to be insulted by the incentive. You need to see what happens with a small group, adjust it as needed, and only when you are ready, roll it out for everyone. Don’t assume that all incentives are created equal, they’re not.”
Denver: Yeah, they’re not going to try. Do you advise… I mean, is there a mental checklist that companies need to go through? Because I got to tell you my experience. When I look at incentives… and it’s nothing like yours, needless to say, but they’re kind of spur of the moment… we need to get a certain behavior… they get together; they come up with something.
And it’s really done in a haphazard way, looking for, in more cases than not, some kind of immediate result. And I didn’t know how you would frame it if you were talking to a company or an organization to say, “No, you’ve really got to think about this in a systemic way.” What advice would you give them in terms of how to approach incentives?
Uri: So very often companies tell me: We tried incentives and they didn’t work. And my answer to this… it’s like going to a bad Japanese restaurant and saying Japanese food is bad. That’s not the right conclusion. You went to a bad Japanese restaurant, or you designed bad incentives.
Now, if you would do… let’s go back to engineering. If I’ll design an engine, a complicated engine, I don’t know about your knowledge, probably would say, Okay, thank you. You won’t think that you can design a better engine. But when we talk about incentives, because we all do it and we have experience with it, we think that we can do better without knowing the basics.
And that’s what I try to do in this book. I try to give you some tools to think about how to design it. Because it’s true, it’s not physics, it’s not rocket science, but there are some rules that you can learn. For example, think about the signals that you are sending. Are you going to upset your clients, workers, whatever it is, whoever it is that you are incentivizing, or are they going to be fine with it, right? Think about it this way.
And another important step to do is to do some kind of A/B testing after that. So don’t assume that it’s going to work. If you have a call center with 5,000 employees, don’t roll it out to all of them. Take a small group of say a hundred people. Give them the incentives. See what happens.
Very often, people are really creative at gaining the incentives. Very often, people are going to be insulted by the incentive. You need to see what happens with a small group, adjust it as needed, and only when you are ready, roll it out for everyone. Don’t assume that all incentives are created equal, they’re not.
Denver: Sounds like an officer of common sense right there again. Absolutely.
Uri: Exactly. Exactly.
Denver: Let me see if you have any thoughts about this… and get back to the nonprofit world for a moment. There’s some challenges that nonprofits are facing. One of them is the number of people who are giving to charity. About 20 years ago, maybe two out of three households gave to charity; now it’s less than one in two.
And the only reason the numbers are staying up overall is because we’re getting those mega donors giving multi hundred million dollar plus gifts. So you’re having more and more households not give to charity. And the same is true with volunteerism. Volunteerism has really dipped significantly here over the last 10 or 20 years.
How can nonprofits use incentives to boost these numbers? And conversely, what should they be wary of, again, to not do anything that’s actually going to dampen them any further?
Uri: So it’s really important to try and understand why people don’t give. So for example, one thing that we found in our research is that people are really worried about the overhead. So I give to a charity, and if I see that the CEO is flying with a private jet somewhere, I’m really pissed at that because I think that my money went to the wrong place.
And I think about it as an economist, right? So I understand that what I should care about is the impact of the organization, of the nonprofit. So say that you need to help kids learn something. I need to see how many kids are going to benefit from my dollar. But instead I look at the overhead that the company is using.
And that’s something that bothers me. And I don’t want to give my money to overhead, which again, as an economist, I understand the efficiency argument. And I also have the overhead argument myself, right? So I have this conflicting thing. What we found in our research is that if you get one of these mega donors to say, You know what? I’m going to cover all the overhead that is associated with, say, giving to this hospital.
Any dollar that you’ll give will go directly to the patient or to whatever it is that you’re interested in. I learned, I did my research. I have people that will look at it because most of us don’t have people that can look at the organization, say whether it’s efficient or not. If I give a hundred dollars, they don’t have money, time, or capabilities to do that.
Uri: But if, see that someone just gave a million dollars to this hospital to cover overhead of raising money, then I’m going to feel much better about it. Then I’ll feel that every dollar that I give actually goes to the cause that I care about. And the challenge in many cases is to convince these mega donors that instead of giving, instead of building a new building, they should cover overhead.
And then there is a multiplier effect because they’re going to be responsible for many more donations. So try to figure out why is it that people don’t give the money. And don’t try to explain to them that they need to think about efficiency, like some of my colleagues think that you should do.
Denver: Yeah. Yeah.
Uri: Don’t try to explain to me… I’m giving you money. Don’t try to educate me. Say thank you. Find what is it that will make you give more, but that will make me give more. And in this case, we found that if someone else is covering the overhead, people are much happier to give.
Denver: That’s an interesting perspective on it because often what we’re doing with incentives, we’re looking in a very micro way of trying to move here to there. And often, what you’re saying is that sometimes you have to take a step back and look at it holistically and think of it from a much larger perspective, and to change that which then will allow other incentives to work, as opposed to just trying to do something…
Denver: …in a very small way. How can organizations effectively measure and evaluate the success of their incentive programs in achieving the required outcomes?
Uri: So here comes the A/B testing. That’s what I said. So for example, with the overhead, in order to test it, we ran a field experiment with the company, with the nonprofit. We sent 10,000 letters to people in which we just asked for money. We sent 10,000 in which we said, we got a donor to cover… sorry, for a matching.
Every dollar that you’ll give, they will give a dollar, which we know is useful. And then we have another donor, another group of people to whom we said a donor covered the overhead, and that group paid much more. So do this kind of A/B testing and see what is it. Don’t assume that you understand what is it that people care about; really try to test it.
Denver: Yeah. Also one of the things when you can find that kind of donor to do that… and a couple organizations do it, charity: water, New Story or whatever, it does allow organizations often to pay salaries significantly more in those hard-to-get expertise… in terms of software design or whatever.
And often, you have to pay something competitive. But when you have that mega donor, you can get the best of the best, and that’s where you really find those long-term solutions when you have that kind of expertise within a nonprofit.
Uri: You just touched a really, really painful point. So I teach in a business school. Well, imagine that an MBA student five years later comes over and tells us, I’m making a million dollars a year as a banker. Now you’ll say, Wow, he is the man, right? He made it. Imagine that person that works in a nonprofit will say, Really? So much money. That’s not nice. That’s not good. I’m giving… right?
And then what happens is that you don’t get the best and the brightest always to go and work for nonprofits. Or if they do, very often they leave quickly because they want to make money, which is legitimate. I want to make money. You want to make money. That’s fine.
And exactly like you said, if you get a mega donor to say, Look, I’m going to sponsor this mechanism, this organization that will help with the raising money. Then that mega donor, she can say, Look, I want to have the best people, and I’m willing to pay for it. You shouldn’t care about this.
It’s not out of your donation. It’s out of my money. I understand this, and I want to get the best people, and I want to compensate them well. Why do we look at the banker that makes a lot of money as a hero and the person that dedicates their life to raising money and helping good causes as cheaters because they get lots of money? That doesn’t make any sense.
Denver: Yeah, and they’re working on the most difficult and vexing problems. They’re intractable problems. They’re harder than a lot of business problems in terms of solving poverty or whatever, so you sometimes do need the best brains in there in order to do it.
Let me close with this, Uri. Given the evolving nature of work and the changing expectations of employees, what are some of the emerging trends or innovative practices in design that you are seeing in incentives that you find to be promising, you find to be intriguing?
Uri: I think that the most important thing is the remote working. That’s what COVID did to the world. Really, it was a huge change. I believe that that would’ve happened, but it would’ve taken many more years.
Uri: And now I see my daughter is a software engineer. She works from home, and she is not less effective because she is not stuck in traffic or because she’s not drinking coffee with her friends, which is important, but it’s definitely not efficient in terms of things. So how do we get to a world… judging by your hair and my hair, we are used to going to work five days a week and being there, but the new generation maybe is more open to working remotely.
And then you can have, I don’t know, Thursday in the office in which you go, and you know that that way you’re not going to sit in front of your computer and write code maybe, but you’ll socialize and you’ll talk with people about problems; you’ll exchange. But the rest of the time, you’ll work from home and incentivizing this, understanding how to make this really efficient, that’s the big challenge I think that we’ll face.
Denver: Yeah. Yeah. Well, it’s a great book and it’s a topic that not enough organizations talk about and can really have a profound effect, that we talk so much about corporate culture and when… and I study it a lot. I talk to a lot of organizations. No one ever brings up incentives, and if they did, they could really have a profound impact on making it a better workplace and having a better organization.
The book, again, is titled Mixed Signals: How Incentives Really Work. I want to thank you, Uri, for being here today. It was just such a pleasure to have you on the show.
Uri: Thank you. I enjoyed it.
Denver Frederick, Host of The Business of Giving serves as a Trusted Advisor and Executive Coach to Nonprofit Leaders. His Book, The Business of Giving: New Best Practices for Nonprofit and Philanthropic Leaders in an Uncertain World, is available now on Amazon and Barnes & Noble.