Search the Site

Episode Transcript

My guest today, Barry Nalebuff, is a professor of economics at the Yale School of Management and co-founder of the beverage company Honest Tea. He’s an expert in the areas of game theory and negotiation. And he has a very big personality.

NALEBUFF: You know the old statement: Turn lemons into lemonade? I knew how to make alcoholic kombucha. And so, we created Kombrewcha. The motto was, “Get tickled, not pickled.”

Welcome to People I (Mostly) Admire, with Steve Levitt.

There are two topics that I especially want to talk with Barry about today. The first is Honest Tea. How in the world does a company started by a professor and a student in the garage end up being a multi, multi-million dollar business? And the second is his new book entitled Split the Pie. When I first heard he had a new book on negotiation, I wondered what new ideas could there possibly be on that topic? But then I read the book and honestly, I was shocked at how original and insightful his approach really was.

LEVITT: So, I first heard about your company, Honest Tea, from the economist, Austin Goolsbee, and it must have been in the early 2000s. And at first I thought he was joking or maybe confused. I said, “Wait, there must be two Barry Nalebuffs, because there’s no way the one who’s a game theorist at Yale that he started a tea company?” And eventually Austin convinced me that indeed, it was true. And our conversation quickly turned to the topic of what an incredibly bad idea it was to try to sell tea, because, really, even now, doesn’t it seem like a terrible idea to try to start a bottled tea company in a garage? Trying to compete for shelf space in grocery stores with Coke and Pepsi and Nestea? Trying to figure out logistics, production. How in the world could you think that made any sense?

NALEBUFF: Yeah, it should have been a recipe for disaster, is the quick way of saying it. What I thought was the case, was all the tea companies out there had failed basic economics. And so, I had to choose between practice and theory, and I thought theory was right here. And that gave us a great confidence that what we were doing would make sense.

LEVITT: So, wait, what do you mean by these other tea companies had failed basic economics?

NALEBUFF: Well, here are some principles of economics that I think we truly believe in: the first is declining marginal utility. So, the first teaspoon of sugar is great. It takes away the bitterness, the second adds flavor. But each additional teaspoon of sugar is less and less valuable, but keeps on adding the same amount of calories and cost. And so, why is it that all the bottled teas out there had 12, 14 teaspoons of sugar when really one or two makes the most sense.

LEVITT: So, why do you think they did that?

NALEBUFF: It was a challenge trying to explain that. I think that they were trying to make tea taste like soda rather than taste like tea. If you’re using super astringent tea, you need lots of sugar. So, it depends on the quality of tea that you’re using. Or they thought that this is what kids want. And the fact is that adults have different taste buds than kids. But iced tea is the second most popular beverage in the United States after water. And when people order iced tea in restaurants, they don’t put 14 teaspoons of sugar in, except perhaps some places in the South.

LEVITT: So, this is a first bit of insight that made you confident you could beat the big players at this. Were there other elements where you think they failed basic economics?

NALEBUFF: So, one of the more famous theorems in economics is called the babysitter theorem. And what that says is, no one goes out and hires a babysitter to eat at McDonald’s. The reason? If it costs you $60 for the babysitter and $10 for McDonald’s, you’re not going to pay $70 to get a Big Mac. It just doesn’t make sense.

LEVITT: So, instead you go to the fancier restaurant than you might otherwise have gone to because you’ve got this fixed-cost investment in the babysitter.

NALEBUFF: Once you have to spend $60 for the babysitter, you’re going to go out to theater, you’re going to have a really nice fine meal. So, now how does it apply to bottled tea? Well, in our business, the bottle is the babysitter. The cost of a bottle, a label, caps, going through production line, getting to distribution — that costs 60 cents just to get the thing on the shelf. It was filled with air. And so, why put a half a penny of tea inside? Why not put two cents of tea that’s really good quality tea?

LEVITT: So, the other companies were spending half a cent on their tea and you guys were spending two cents on your tea?

NALEBUFF: Exactly. And I had no idea there was such a thing as good tea and bad tea. But I ended up going to India to write a case study and there I experienced wonderful tea and realized that great tea is one of the world’s cheapest luxuries. For a couple of pennies, you can enjoy fantastic tea. And when you do, you don’t need to dilute it, you don’t need to disguise the flavor with a massive amount of sugar.

LEVITT: Is there more economic wisdom lurking in the background?

NALEBUFF: Hopefully your listeners are somewhat nerds, and they may remember the fact that in calculus, when you have a function and you maximize that function, the derivative is zero.

LEVITT: Okay. They probably don’t, but let’s keep on going anyway.

NALEBUFF: Essentially, that’s what we call marginal analysis. You want to try and find the optimal amount of sweetness. And if you do a taste test, you will go and with blind taste testing, figure out what leads people to like your product the most compared to other levels of sweetness.

LEVITT: And you actually did that with Honest Tea? You did blind test to figure out how much sweetness people wanted?

NALEBUFF: Well, no, we rejected blind taste tests.


NALEBUFF: Because it’s only focusing on one dimension. In a blind taste test, your mouth is open and your eyes are closed. We could flip that and have your mouth closed and your eyes open, which is what you might call a label reading contest.


NALEBUFF: And on that you really want to have very low amounts of sugar and low calories. And so, if we backed off the amount of sugar, we’d lose almost nothing in terms of taste. But we would save cost of ingredients and make a product with many fewer calories.

LEVITT: So, a really interesting concept from economics is that if you’re close to an optimum, if you’re getting the right answer, then it turns out deviating a little bit from that answer, almost never cost you very much. It’s when you’re far away from optimum and then you make a mistake, that’s where you can really pay much more. So, you’re saying, you don’t have to get the sugar exactly right. You can have a little less sugar and people will barely even notice. But sugar is not free. So, if we can cut back on sugar—

NALEBUFF: We’re saving calories and we’re saving on the cost of ingredients.

LEVITT: And you’re doing it in a way that will barely be discernible to the customer.

NALEBUFF: Exactly. And so, the optimum level of sweetness is not the one that wins the blind taste test. And that’s what all the other companies were out there doing. And that’s one reason they made their products all too sweet.

LEVITT: So, people probably followed this argument, and I think what will surprise them is that they might understand how an economist would go ahead and make this calculation, make it not as sweet, but if I remember correctly, you put that graph and that explanation right on the back label. You went and told your customers, “Look, we didn’t make this taste as good as we could for these various reasons.” Is my memory correct on that?

NALEBUFF: We made this product for people who study at the University of Chicago. We put the only product out there with utility curves on the back label. We also made this product for people who read labels, and wanted to understand what it is that we were doing. So, when they tasted the product and it wasn’t quite as sweet as they might have imagined it would be, or might have wanted it to be, they could realize there were other factors that should more than compensate for the little bit less sweetness.

LEVITT: So, you started this and at some point, you needed to go for external funding. Could you describe how you structured the deal for your outside investors?

NALEBUFF: So, what I find ironic, is that the typical entrepreneur will talk about, “This business is going to be worth a hundred million, 500 million. And so, would you please give me money to invest at a 1, 2, 3, 4, 5 million valuation?” I was like, well, wait a second, if you really believe this business is going to be worth a hundred or more million, why are you taking money at such a low price? Kind of a dissonance there? It doesn’t make sense.

LEVITT: Right. I agree.

NALEBUFF: And so, what we said is, “We really believe this business is going to be worth well more than a hundred million.” And the way we’re willing to prove that to you, is we’re going to give you the entire company upfront, and essentially claw it back to the extent that we are successful in making this business valuable. So, the investors put in a $1 million, they owned a hundred percent of the company. Which was very hard for them to argue with in terms of the valuation. But we said, “Once we double your money, we have options to buy a million shares at two dollars. And then we have options to buy another million shares at three dollars. And at $5, and at $10, and at $15.” So, essentially, as we continue to be successful in making this company worth a lot, we will get a share of the business. Effectively, what I’m saying is, the value of the company initially should depend on how well we do.

LEVITT: So, you gave away the business with the option to buy it back if it was successful. So, if the company had not done very well, just, you know, middling, then you would have done very poorly. You would have gotten essentially nothing. So, at the end of the day, in a big success for a million dollars — you raised a million, right? Your first raise was a million? Is that right? Or is this just hypothetical?

NALEBUFF: It’s a million and in the end, the share price went from $1 to $23. So, people got 23 times their money.

LEVITT: Okay. So, people got back 23 times their money and then how much of the company did you own back in this successful world?

NALEBUFF: So, we went from owning zero of the company to a little over 40 percent of the company, by the time we had been successful.

LEVITT: Ex-post, obviously you should have done your financing very differently, right? You should have found some way to mortgage your houses to get that million dollars and keep all the ownership — ex-post.

NALEBUFF: Yeah, although, actually, yes and no. We had taken out loans that were backed by our houses and that was sufficiently scary. It’s one of the reasons we ended up selling the business. Because effectively, everything was on the line. And having a little bit less on the line might’ve allowed us to ride it out even more and get it to the billion-dollar valuation rather than the multiple-hundreds-of-million valuation.

LEVITT: I see. Yeah, no, it’s true. The existential risk associated with having everything that you’ve built over your life invested in a company that could disappear overnight.

NALEBUFF: And this is yet again, a case of declining marginal utility. At some point, how much more is the upside worth? And the downside is really severe. And I’m very happy to have made my friends, my relatives, all successful. Gives me great joy.

LEVITT: So, we’ve been talking about the things that went right. The financing system worked out great, the massive growth. I think it’s often more informative to talk though about the mistakes, the failures. Are there any particularly compelling mistakes that you made that you think would educate people hearing about them?

NALEBUFF: Sure. So, as you say, you learn a lot more from your mistakes than your successes. And I was hoping to be totally ignorant here as a result, but that didn’t work out. The first mistake we made is that we didn’t truly understand what our business was about. The name Honest Tea, we thought the most important word was “tea.” And of course, the most important word was “honest.” And our first expansion was into teabags. Didn’t turn out so well. Turned out the teabag business, the dry tea, is a completely different distribution, production than the tea in a bottle. But when we figured out that the “honest” was the important part of our word, that led us to create Honest Kids, which is essentially, a lower calorie fruit juice in pouches, a healthier version of CapriSun. And that was a giant success. I think the sales now exceed that of Honest Tea.

LEVITT: My kids drink that all the time.

NALEBUFF: No wonder you’re such a good parent. Now, I would say that we didn’t necessarily go far enough in that understanding. So, while we got to Honest Kids, there’s Honest Food, there’s the Honest Company. And so, we could have been making the diapers, the sunscreen, all of the other products that Jessica Alba and her team are up to. We were there first, in terms of understanding the value of “honest.”

LEVITT: So, the concept of authenticity has a guy named B.J. Miller, who does palliative care. And he said that authenticity is the single most important thing to have when you’re talking about dying or trying to consult people who are grieving. I was just talking, a while back, with the economist, Seth Stevens Davidowitz, and he followed a very authentic track in his research, which let him not to get an academic job, but turned into really exciting stuff that made him a best-selling author. And you give the impression that this business is based completely on authenticity and not only so much on authentic ingredients, but some kind of authentic weirdness that’s emanating directly from you. Is that a fair assessment?

NALEBUFF: Well, by putting the word “honest” on your name, one, you’re implying that other folks out there are not as honest as you. And that led to a lot of people saying, “Okay, what’s so honest about your product?” Well, we were the first company out there to label how many calories in a bottle as opposed to calories in a serving. Because essentially it was B.S. out there, the bottles with 16 ounce., They would pretend their calories are half as much as they really were by writing calories per serving in two servings. So, we didn’t play that game. We had a product which used a small amount of erythritol and it had five calories. You’re allowed to round that down to zero. We didn’t do that. So, we were out there explaining exactly what it is they we’re up to, how we’re doing it, why, and I think people trusted us because we gave them the information that they deserved.

LEVITT: You sold the company to Coca-Cola. Did that not worry you that there’d be some conflict between this sense of honesty and working with a company, who all along the way, you must’ve been implying they were very dishonest?

NALEBUFF: I — not that they are dishonest—

LEVITT: But they were not labeling their calories correctly. I mean, they were doing all the things everybody was doing, other than you.

NALEBUFF: They were doing what the industry does. And the good news is, we were out to change that. Mutar Kent, the then C.E.O. of Coca-Cola said he didn’t want Coca-Cola to change Honest Tea. He wanted Honest Tea to change Coke. And as a result, we helped bring organic ingredients into Coca-Cola. We helped them expand their fair-trade activities. So, yes, if we could change Coca-Cola 2 percent, that was going to have a much bigger impact than anything that we could do on our own.

LEVITT: So, was Honest Tea your last start-up or have you got the bug now to do it again?

NALEBUFF: One of the products we made at Honest Tea was Honest Kombucha, which turned out to be accidentally alcoholic, and so we had to take it off the market.

LEVITT: Wait, how do you make a product that is “accidentally” alcoholic?

NALEBUFF: Yeah. So, kombucha is a live mixture, a symbiotic culture of bacteria and yeast. And what yeast do is turn sugar into alcohol. And they keep on doing it after you put it in the bottle. And so, when it left our factory, that product was below a half a percent in alcohol, which is the legal limit to be non-alcoholic.


NALEBUFF: but somehow on the shelves, it kept on doing its thing.

LEVITT: So, what happened?

NALEBUFF: So, a famous movie star was pulled over for drunk driving and she claimed it was the alcohol in the kombucha that had caused her to—

LEVITT: She must have had a lot of bottles?

NALEBUFF: Yeah. So, anyway, that led Whole Foods to do testing on its products on the shelves. And next thing we knew, we were at 0.7 percent and there’s no margin of error. And so, we had to take a national recall. We were sued for tricking people into drinking. It was a nightmare. And ultimately, we had to discontinue the product because we couldn’t figure out how to keep the alcohol level below a half a percent. However, you know, the old statement: Turn lemons into lemonade. I knew how to make alcoholic kombucha just the way I was trying to do it before. And so, with another one of my students, we created Kombrewcha, a slightly alcoholic version of kombucha. The motto was, “Get tickled, not pickled.” And that company was sold to Anheuser-Busch.

You’re listening to People I (Mostly) Admire with Steve Levitt and his conversation with Barry Nalebuff. After this short break, they’ll return to talk about the art of negotiations.

*          *          *

Morgan LEVEY: Hey, Levitt. So, our listener Matt writes, “I have the impression that economic results don’t generalize very well. What can economists do to get better at predicting and to convince the public they’re good at predicting?”

LEVITT: Okay, so it’s interesting because while many people have this view that the goal of economics is to make predictions, in fact, in my daily life as an academic economist, I am focused 100 percent, not on predicting the future, but in understanding the past. So, every academic paper I’ve ever written and virtually every paper any economist has written, takes a pile of data or a question or a puzzle from the past and tries to make sense, looking backwards of what happened. And so, it shouldn’t really be surprising that we’re not that great at predictions, given that’s not what we’re focused on.

LEVEY: Yeah, but I think you’re skirting around something. Isn’t the point of doing academic work so that we can learn something about our present time or make better policy decisions for the future?

LEVITT: So, economics has these two different pieces. There’s microeconomics, which tries to describe how individuals will behave. And then there’s macroeconomics, which is about how the overall system of the economy functions. And I would say, we’re really good at microeconomics. If you tell me how much a price is going to change, I could give you a really good estimate of how much the quantity will change in response — of what buyers will enter the market or what sellers will exit the market. I think we’re really good at that, especially as we’ve started to integrate behavioral economics and insights of psychology. What is just intrinsically a much more difficult problem is understanding the macroeconomy. It is this chaotic, complex system that ends up going beyond the power of what mathematical models can do. I feel pretty good about the state of economics for making microeconomic predictions, and I’m with Matt about macroeconomic predictions. But I also think it may just be impossible because macroeconomic problems are so hard that we may just always be in the dark.

LEVEY: Well, do you think economists and maybe academics in general should be more transparent in their uncertainty?

LEVITT: It’s probably not true of the economists who go on C.N.B.C., but among academic economists, I think they really do acknowledge their uncertainty. I remember right in the middle of the great recession, I was having lunch with one of the world’s premier macroeconomists. And I said to him, “You know, everybody is so angry at macroeconomics because you’ve failed to predict the great recession. How do you defend yourself against those claims?” And he said, “That’s not my job. My job is not to predict the future. My job is to understand the past and anyone who thinks I can predict the future, they’re crazy.” This is a Nobel Prize winning macroeconomist, and the thought that he would understand the actual macroeconomy in real time and be able to say something intelligent about what was going to happen — that was ludicrous to him. Maybe what we need is more honest academics being pundits, rather than the people who speak with great certainty when it’s not warranted. I think the problem is that the producers on the T.V. shows, they don’t want honesty from academics. They want a great story. And the great story is the one told with complete conviction and certainty, even though it’s not going to turn out to be true.

LEVEY: What we’re talking about really touches upon something that was brought up in the David Keith interview, which is, there’s a real difference between scientists and the role of scientists. And then politicians and policymakers. Scientists are there to do the science — to find things in the data and put it out there. And then it’s the politician and policy-makers role to tell us what to do with that. And we get in trouble when the line blurs between the two.

LEVITT: Yeah, and yet, it’s so enticing as an academic — you sit in your office all day long. No one wants to hear anything about what you’re doing and suddenly you get the chance to go in front of a big audience. And if you speak like an academic, you will never get another chance to go in front of a big audience. You only get to go in front of big audiences if you put on a different hat — an advocate hat or a policy hat, and it’s tricky. It’s complicated. And we haven’t really figured out how to do all that right — that’s for sure.

LEVEY: Matt, thank you so much for your question. If you have a question for us, our email address is That’s Steve and I read every email and we look forward to reading yours.

*          *          *

So, we covered Honest Tea. Now, I want to tackle negotiation. My one concern, though, is that Barry’s approach requires some computation. In his book, the numbers are right there in front of you. On a podcast, it’s going to take a little bit of concentration, but I’ll do my best to try to keep it simple and understandable.

LEVITT: You tell a story in your new book called Split the Pie about how you tried to save a few bucks by not hiring your lawyer to help you file for a trademark. And that seemed like a good idea until you then went to reserve the website — the domain name. And tell me about the rude surprise you woke up to when you try to then reserve the domain name.

NALEBUFF: Yeah. Well, when you file a trademark, it has to be public. And the reason it has to be public is because somebody has to be able to object to it. I’ve got a company called Real Made, which is overnight oats. And I stupidly did not buy the domain name before filing for the trademark. And literally, the day it became public, this troll, who I’ll call Edwar, — I discovered he had bought the domain name.

LEVITT: He had bought the domain name the second it became public. He snatched it up.

NALEBUFF: That’s right. He’s out on his computer, watching all the releases of the trademark applications. He tests to see, oh, have they bought the domain name dot com or not? Oh, look, this guy, he’s an idiot. He hasn’t done it. I’ll get him.

LEVITT: So, what happened next?

NALEBUFF: So, I write to him and he writes back saying, “Sorry, did not realize that this domain name was related to your trademark. Sorry, if this makes you feel bad, but for 2,500 bucks, I’ll sell it back to you.”

LEVITT: Okay. And roughly, what do you think it was worth to you, like if you had to pay —

NALEBUFF: $50,000.

LEVITT: $50,000. Okay. So, in that sense, it’s not the worst. He could’ve asked for a lot more. So, how do you respond?

NALEBUFF: He didn’t know what it was worth to me. And of course, let’s be clear, it’s worth zero to him. There’s no other value. Well, okay., I’m dumb, but I’m not stupid. And by that, I mean, I went and looked to see what was really going on here. And what Edward had done is called registration in bad faith. And there’s an organization called ICANN, which manages domain names. They have a dispute resolution process and for $1,300, I could protest what Edward had done. And I was, essentially, a 100 percent guaranteed to get the domain name back.

LEVITT: Okay. So, you definitely didn’t need to pay $2,500. So, then what did you say to Edward?

NALEBUFF: So, I write to Edward and say, “Hey, look, glad that you feel sorry about this. But, there’s no way I’m going to pay you $2,500. I’d rather pay ICANN $1,300 and you’ll end up with zero. So, if you truly feel sorry, I’ll buy it from you for $500. Otherwise, I’m happy to go to ICANN.”

LEVITT: The worst you can come out of this is down $1,300, and you made the offer to Edward that you will give him $500. So, you would end up down $500 instead of down $1,300. But Edward could not have been happy about that.

NALEBUFF: Well, you know, $500’s still better than zero.

LEVITT: So, he said, “Sure, I’ll take it?”

NALEBUFF: No, of course not. He comes back and says, “Yes, I know about the domain process. It’s money wasting. So, in light of that, I’m prepared to sell this to you for $1,100, but I’m leaving on my vacation, so, seal the deal. Let’s make this happen.”

LEVITT: Okay., so let me stop you for a second, because right now this sounds like a pretty standard bargaining. But your approach to negotiation turns much of the standard conventional wisdom on bargaining upside down. So, when are you going to employ that kind of special Barry Nalebuff logic into this particular transaction?

NALEBUFF: In the very next email. So, what has Edward has done, first, he’s done something called anchoring by starting with a high number. Now he’s trying to do a deadline effect. He’s saying that, “If you want to make this happen, you gotta make it happen now because I’m going away. And this is your chance. So, 5, 4, 3, 2, 1, shake my hand. Do this deal. Say yes.”

LEVITT: I would say Edward sounds like a very seasoned negotiator. Someone who’s read probably many of the leading books on negotiation and understands human nature and how to exploit frailties that people have to get a really good deal for himself.

NALEBUFF: Yeah. And there are more tricks up his sleeve as you’ll see.

LEVITT: Okay., great.

NALEBUFF: So, I come back and say, “Edward, here’s how I see the world: The reason I’m talking to you is to save the $1,300 fee from ICANN. And that what you’ve proposed at $1,100, is you’d be $1,100 ahead and I’d only be $200 ahead. Because $1,100 is just saving me 200 compared to the $1,300 I’d pay to ICANN. That doesn’t sound fair to me. Essentially, it’s like my offering you $200, so that I’d be up $1,100 and you’d be up $200. I wouldn’t expect you to accept that. And you shouldn’t expect me to accept your offer. What I’m prepared to do is split the $1,300, $650-$650. If I pay you $650, you’ll be $650 ahead of no deal. And I’ll be $650 ahead of no deal. And that’s as far as I’m prepared to go.”

LEVITT: Okay, so let me stop there, because there’s a bunch of interesting things floating around in the background. Now, the first one is that you started this message to Edward by identifying the size of the pie to be split. And that’s key to everything underlying what you do. But then you did more than that. You showed him why his offer wasn’t fair by proposing a hypothetical offer. Now, it’s important, I think, that it wasn’t a real offer. You didn’t go back to Edward and say, “$200.” You said, “I could offer you $200, but that would be insulting or unfair.” And then you said, “Let’s just split the pie. Let’s just try to make this fair.” Which is not, as far as I know, one of the things that when you read negotiation books, they tell you to try to do. Why not get more than half, but you’re willing to settle for half.

NALEBUFF: A lot of them do talk about the F-word, the fairness word, but they throw it around without actually ever defining it. There’s an expression, “Fight fire with fire,” but I say, “Fight fire with water.” And that if I flipped the 1,100 offered to a $200 offer, essentially I’m playing his game. And so, instead of doing the actual flip, I’m going to explain the hypothetical flip and why it’s unfair. And that way I’m hoping to put out fire, his ultimatum of $1,100. Well he comes back, and does what people expect you to in this case, which is, “Let’s try the old split the difference. You are at $650, I’m at $1,100. The average of that’s $875, but we’ll round up to a nice number — 900.”) And so, he says, “$900. That’s my final offer. If you can’t do this. We have no deal and note how much more I’ve moved. I’ve gone from $2,500 all the way down to $900. So, procedural fairness, you’ve only moved up $150.”

LEVITT: And did you give into that logic?

NALEBUFF: Total B.S. Because his move from $2,500 down to $1,300 doesn’t count in my book. The $2,500 was a totally arbitrary number. He could have started off with $47 million. So, until he gets below $1,300, I don’t count anything. So, my view is he’s moved from $1,300 down to $900, whereas I moved from zero up to $650. But at the end of the day, I didn’t make any counterargument. I didn’t answer his email. And then a week later he writes back, says, “We have a deal at $650. Please give me your information.”

LEVITT: The economist part of me sees the beauty of your approach, right? So, you figure out how much joint surplus there is, around which to make a deal. And then you solve this thorny problem of how to split the surplus by just essentially making the case that the only reasonable split is a fair one, 50-50. And that’s logical and defensible, but the emotional part of me, it’s pissed off. I’m pissed off. He’s not interested in fairness. Why should you offer them a fair deal? Why not offer him $100? And if he doesn’t take it, it’ll cost you $1,300, but look, at least Edward gets nothing.

NALEBUFF: I wouldn’t have such a good story, right?

LEVITT: There really, I think, is a tension in negotiation that at some level negotiations are economic deals, and at another level, the process matters, and there’s an important psychological component to it.

NALEBUFF: Well, the last thing in the world I was going to let this troll do is get an unfair deal, from my perspective. I really was prepared to hold firm at $650. So, yeah, I was willing to swallow my pride. I had made a mistake. I recognize that it was my error by not filing the domain name first. And yeah, I was willing to save $650, and let him have $650, compared to paying $1,300, but I wasn’t prepared to let him get more than half the pie.

LEVITT: So, in your view, every negotiation shares important commonalities. What are those?

NALEBUFF: The question is what is the negotiation really about? And the surprise is people are confused about that. It’s not about the total numbers. It’s about the surplus that the two parties can create by coming together, compared to what they can create if they don’t reach a deal. So, in our case here, I couldn’t save the $1,300 without Edward, but he couldn’t get any of those $1,300 without me. And so, we need each other equally. The negotiation was about that $1,300.

LEVITT: Negotiations aren’t about the total dollars that are out there. It’s about the extra that’s created by the coming together of these two parties.

NALEBUFF: And the second point is that once you correctly understand what the negotiation is about, the two parties have equal power. There’s no sense in which one side is weaker than the other, smaller than the other, is contributing more. They’re always the same.

LEVITT: Because without one of the parties there is no deal.

NALEBUFF: There’s no deal.

LEVITT: And so, then you fall back to your best outside option. And in that sense, say, when Honest Tea was negotiating with Coca-Cola, even though you were small, the idea was, whatever surplus will be generated by merging Honest Tea with Coca-Cola, your argument, I guess, would have been, you should split that equally between Honest Tea and Coca-Cola. That probably wasn’t Coca-Cola’s view of the world.

NALEBUFF: Actually, that’s exactly what we ended up doing. And we reached an agreement to split the pie in our first hour of our conversations together.

LEVITT: This is a much more intellectual approach to negotiation. It’s also kind of new age-y, right? So, part of your thesis is that you should put yourself in the shoes of the other person and understand their situation.

NALEBUFF: I think that people are looking for a framework for negotiation. That people who are nice, empathetic, curious, find themselves acting like jerks when they negotiate, because they don’t have a framework, and they are desperate for doing something that will allow them not to be taken advantage of when they negotiate.

LEVITT: It’s true, it does appeal to me because it’s no longer a testosterone-laden,“How do I take from the other person?” It changes the process, which I personally really like.

NALEBUFF: In fact, the way I suggest you start a negotiation is by having a discussion of how you negotiate. And rather than say, “Hey, my price is $2,500,” a la Edward. It’s, “Can we agree that our goal is to create a giant pie and split it? Because if we can agree on that, then we can focus our attention on making a bigger pie as opposed to beating each other up, playing the traditional ridiculous games and perhaps reaching no deal.”

LEVITT: So, I would love to go through an example — you have a good example, something simple we could work on?

NALEBUFF: Sure. Let’s do an interest rate example. We have Anju and Barat. Anju has $5,000 to invest on which he can earn 1 percent or $50. Bharat has $20,000 to invest on which he can earn 2 percent or $400. But if they combine their funds, they’ll have $25,000 to invest on which they can earn 3 percent or $750. Now, they could only get that 750 if they can agree between the two of them, how to divide it up. So, my question to you is how much of the $750 should Anju get and how much should Bharat get?

LEVITT: Okay, so walk us through that one.

NALEBUFF: Well, one answer would be just divided in two, $375 to each, but that won’t work because Bharat can get $400 on his own. So, that can’t be the right answer. The most common response is that Bharat has put in four times as much money as Anju, so therefore he should get four times as much. Bharat gets $600 and Anju gets $150. That’s the proportional answer. That’s what Aristotle would say. And that is the most common response out there.

LEVITT: But in your view, that’s not the right response.

NALEBUFF: It isn’t because it misses what the negotiation is about. It’s not about the $750. It’s about the additional value they create by having a deal. On their own, Anju can get $50 and Barat can get $400. So, collectively they can get $450 with no deal. The deal brings them from $450 up to $750. The reason to have an agreement is to get the extra $300 in interest. To get that Anju and Barat are equally essential. So, you divide it, 150, $150. That means Andrew gets $50 plus $150 or 200, and Bharat gets $400 plus $150 or $550.

LEVITT: How about you give me a chance to see if I’ve internalized the logic that you’ve got?

NALEBUFF: I’ve got a few for you. We have a merger case, and the company A is twice as big as B. And if they come together, they can save $6 million in savings by doing joint purchasing. So, A says, “I should get four and you should get two because I’m twice as big as you.” And what’s the response?

LEVITT: Well, they’re both equally important to grading that $6 million, so obviously you have to split that $3 million, $3 million.

NALEBUFF: Great. That was the warm up.

LEVITT: Okay, good. All right, so I passed the warmup.

NALEBUFF: A, it turns out has know-how. They can have a million dollars of cost savings they can apply to B. And A says, “That was cute there, Steve, about the three and three, but the know-how is our know-how. And so, we should get to keep that whole million.”

LEVITT: So, the know-how is worth nothing to A, if they don’t work with B.

NALEBUFF: They’ve already applied their know-how to their own operations, and they can improve B’s operations by a million dollars through their know-how.

LEVITT: Yeah. But without B being bad at doing what they’re doing, A’s know-how it isn’t worth anything. So, again, they got to split that 50-50.

NALEBUFF: So, what you said, and I think that’s the important part is what is the B is bringing to the table? It’s their inefficiencies. It’s their lack of knowledge. A’s know-how isn’t worth anything without B’s operations that have room for improvement. All right? That was the medium one. Now let’s go to the harder one.


NALEBUFF: These are two newspapers and because A is twice as big as B, A can bring 100,000 readers to B. And B can bring 50,000 readers to A. The 100,000 readers is worth 2 million. And the 50,000 readers is worth 1 million. So, A says, “Look, I’m bringing you twice as many readers, so I should get the 2 million and you should get the 1 million.”

LEVITT: Now maybe I’m falling down the job, but it sounds like, again, you need both A and B, so you still want to split that again, 50-50.

NALEBUFF: So, what you’ve done, and I agree with, is the high-level argument, but you also have to show people what’s wrong with the argument they’ve made. So, what’s wrong with the story that I’ve said that A is bringing twice as much to B? A is giving B twice as many readers as B is giving to A.

LEVITT: It seems like it’s true, but irrelevant, would be my take on it.

NALEBUFF: Not quite.


NALEBUFF: So, here’s the argument that I think you should make is, don’t look at A is giving twice as many readers to B. It’s, of the a hundred thousand readers that A is giving to B, B is giving them something to read.


NALEBUFF: And that’s why those two should be split. When B is giving A 50,000 readers, B is giving the readers and A is giving the content. So, the mistake is saying, “You’re giving me twice as many readers,” you could have just said in reverse, “I’m giving you twice as much content.” And it’s only by putting the readers and the content together that we create the value.

LEVITT: Okay, so that was good. I did pretty well on that one.

NALEBUFF: Two out of three. We give you an A at Yale for that.

LEVITT: So, many of the examples you have in the book have splits that seem to be much more surprising and counterintuitive. Do you have one of those up your sleeve that I could take a crack at?

NALEBUFF: Okay. This is the really hard one now. I’m giving two talks. I’m giving one talk in San Francisco and one talk in your hometown of Chicago. And if we can get the two parties who are arranging these talks to coordinate, I can do a triangle route rather than doing two round trips. And so, the question is, how should the airfare be divided between the group in San Francisco and the group in Chicago?

LEVITT: Okay. The contract says Barry will fly to San Francisco and the people in San Francisco will pay the cost of that flight. And Barry will fly from New Haven to Chicago and the people you’re speaking with have agreed to pay that cost. But you’re saying I got to do these two, and there might be some cost savings. How are we going to split the bill?

NALEBUFF: So, instead, I could do New York to Chicago to San Francisco, and San Francisco home.

LEVITT: Okay. So, what is a round trip to San Francisco? And what’s the round trip to Chicago?

NALEBUFF: The round trip to San Francisco is $2,000 or $1,000 each way. And Chicago, it’s $200 round trip or a hundred each way. And the last number you need is Chicago to San Francisco. That’s $500.

LEVITT: Okay, Chicago — I got to write these down. Chicago to San Francisco, that’s $500. Let’s pretend we don’t know anything about how Barry thinks about it. Let’s think of some of the possibilities.

NALEBUFF: On the triangle route, in all cases, I’m always going to Chicago and I’m always coming home from San Francisco. So, you could say, “Chicago pays $100 to get me there. And San Francisco pays $1,000 to get me home.” And now the question is, what do we do with the Chicago to San Francisco leg, the $500 part there?

LEVITT: Okay., and then it might be logical to say, let’s split that 50-50, but obviously, that’s a problem because Chicago’s going to say, “No way, because it only costs 200 to go round trip to Yale.” You can’t possibly ask them to pay them $350.

NALEBUFF: And what you see is, there’s like four or five different arguments. None of which actually makes any sense, but different people, depending on which side you’re on, would pick different ones of those to make.

LEVITT: Yes. Okay, so let’s try to solve this using the split-the-pie logic.

NALEBUFF: I’ll help you with some of the numbers again. If we do the round trip to San Francisco, it’s $2,000, and the round trip to Chicago is 200.

LEVITT: Okay. So, outside of making a deal, it’ll cost $2,200 to fly, but if we do the triangle route, it’s a $100 to get to Chicago. It’s $500 to get from Chicago to San Francisco. And it’s a $1,000 to get back from San Francisco to New Haven. So, that’s $1,600. And the cost of the two round trips was 2,200. So, that difference is $600. So, that’s the pie we’re talking about.

NALEBUFF: Exactly. And who was more essential to make that $600 pie? Chicago or San Francisco?

LEVITT: And they’re equally essential because without either one, there wouldn’t be a deal.

NALEBUFF: So, therefore, each side should be $300 ahead.

LEVITT: We know before, it was going to cost San Francisco $2,000. So, They should pay $1,700. And Chicago, they were only going to pay $200, they should get negative 100. So, you’re saying the solution is actually that San Francisco should not only cover the cost of the whole three-legged trip. They should actually subsidize Chicago for $100.

NALEBUFF: I’m afraid that’s the split-the-pie answer.

LEVITT: (laughs) Wow. That is counter-intuitive. That’s really interesting. Because you saw me pause, I thought I made a mistake. Yeah, that’s interesting.

NALEBUFF: Now a lot of people would say, “Timeout. If I were Chicago, I’d just be happy getting zero. I’d pay nothing and let San Francisco pay the whole amount.” And if I had to sell this to you or to one of our listeners, I wouldn’t say, “Hey, San Francisco pay $100 to Chicago.” I’d say, “Can we agree that the pie here is $600? That we create $600 of savings by doing the triangle root compared to the two round trips? Yeah? Can we agree that we’re both equally essential? Yeah. So, we should both be $300 ahead. Yeah.” Then I’d shake hands. And essentially what I want people to agree on is the principle and then numbers come out however the numbers come out.

LEVITT: This is not just about dividing a pie. It’s the fact that if you don’t think about the negotiations the right way, you’ll make deals you shouldn’t make, collectively, and you’ll fail to make deals that did have surplus, but because you didn’t negotiate properly, that’s left on the table. That’s really the crux of what you’re doing here, I think.

NALEBUFF: I would take even further if I may. So, one of the most famous experiments in economics is the ultimatum game. In an ultimatum game, one person makes a proposal to the other side of how to divide up $100. Say I get $80 and Steve gets $20. And then Steve can say, “Yes,” or “No.” He has no ability to make a counter. All he can do is accept or reject. And the experimental evidence out there says that if you don’t offer somebody a lot more than $10 or $20 — if you don’t get them close to half, they’re often going to say no and get zero rather than take what it is that you’ve given them, which is something.

LEVITT: Which is surprising to economists because the economists says, “Look, I don’t have any power in this game. The other person gets to split the pie. I can either take something bigger than zero or get zero.” Of course, I should take something bigger than zero, but that actually goes back to when you described your negotiation with Edward and I said, “Oh, my economist side loves what you’re doing, but my emotional side is really angry.” And clearly that plays into negotiations.

NALEBUFF: And my point is that a lot of people don’t reach deals because they think they’re being treated unfairly. And so, they turned down the 20, they get zero and the other side gets zero. And so, we think of negotiations often as zero sum, because the more I get, the less you get, but failing to reach a deal is zero-zero. And so, having a fair solution allows people to reach a deal and avoid the lose-lose outcome.

LEVITT: You teach these skills as a class to the masters, the M.B.A.’s—

NALEBUFF: To M.B.A.’s to undergraduates, to law students, to international relations — to anybody who will listen.

LEVITT: And how often do you get emails or phone calls where your former students thank you for some successful negotiation that they just carried out?

NALEBUFF: Oh, all the time. And I don’t get my 5 percent, you know, that’s what I’m looking for.

LEVITT: Is there a single best story from a student?

NALEBUFF: Oh, one of my former students was involved in a merger deal, and the two parties had agreed on all the terms. They were all set, and then it came time to paying the fees to the lawyers, to the bankers. And it turned out that created a $15 million tax savings, if the losses could be captured. And the company that was being bought didn’t have profits, so they couldn’t use them. But the company who was acquiring them had the ability to take advantage of those tax savings. And so, my student was representing the buyer and the seller says, “Look, we’ve got these losses, which create these 15 million of tax savings, so we want them.” And his C.E.O. was ready to, say, “Yes. Okay. They’re your losses. Yeah, you can have them.” And my student says, “Timeout, you need our profits to take advantage of these losses. And only by coming together, can we actually save this $15 million. So, we should split it seven and a half, seven and a half.” And they agreed.

LEVITT: Wow, that makes the price of higher education cheap — $60,000 a year, whatever we charge our students — $7.5 million in one deal.

NALEBUFF: In one deal. And what I like about it is, there’s no counterargument that you can hold your breath, but that’s not a principled approach. And so, when somebody provides logic, they can try the emotion, but I think, at the end, we need more Spock. We need more logic in negotiations.

You got a little taste of Barry Nalebuff’s approach to negotiation. What do you think? The book is called Spit the Pie and I highly recommend it. I was skeptical at first, but he won me over and I can’t wait till the next time I have to negotiate something to see how the methods work in real life. And if you’re inspired to try Barry’s approach in your next negotiation, send us an email to let us know how it turned out. The email address is That’s  Right on the heels of recording this interview, Coca-Cola announced that it is discontinuing Honest Tea. But never one to be deterred, Barry and his business partner will be launching a new tea brand called Just Ice Tea, which, if you put the words just and ice together, can also be pronounced “Justice Tea.” So, be on the lookout. Thanks for listening, and we’ll see you next week.

*          *          *

People I (Mostly) Admire is part of the Freakonomics Radio Network, which also includes Freakonomics Radio, No Stupid Questions, Freakonomics M.D., and Off Leash. All our shows are produced by Stitcher and Renbud Radio. Morgan Levey is our producer and Jasmin Klinger is our engineer. We had help from Alina Kulman. Our staff also includes Neal Carruth, Gabriel Roth, Greg Rippin, Rebecca Lee Douglas, Zack Lapinski, Julie Kanfer, Eleanor Osborne, Ryan Kelley, Emma Tyrrell, Lyric Bowditch, Jacob Clemente, and Stephen Dubner. Our theme music was composed by Luis Guerra. To listen ad-free, subscribe to Stitcher Premium. We can be reached at, that’s Thanks for listening.

NALEBUFF: I can’t tell you the number of times that we had to deal with just utter stupidity on the side of our customers. My favorite example of that is a distributor who wouldn’t carry Honest Tea because he wasn’t making enough money per case. He said, “I’m making four bucks a case with you. Whereas I make six bucks a case with Snapple.” And I’m thinking, how in the world is that possible? Our product sells for more money than Snapple. And then I realized Snapple comes in a 24-bottle case and Honest Tea comes in a 12-bottle case.

Read full Transcript


  • Barry Nalebuff, professor of economics at the Yale School of Management. 



Episode Video