Split the Pie by Barry Nalebuff
I finished this book in October 2023. I recommend this book 3/10.
You should read this book if you are looking for a book on negotiating written by a Yale professor and if you want to use a mathematical approach. This is a good book for someone who does not enjoy hackling but wants to be fair and split the winnings for both parties.
Get your copy here.
My notes and thoughts:
P11. You need to negotiate the difference between the two candidates. Example: Dealer wants $6000, buyer aims for $4000, you are then negotiating the pie of the difference of $2000.
P18. Find out what the real number is. You have to do some research. Seller will many times just throw a number at you.
P19. The starting point of any negotiation should be: What is at stake? While this might seem an obvious question, the answer turns out to be subtle. The scope of a negotiation is what the two sides can jointly create over and above what they can do on their own—the pie. To create the pie takes both parties. By definition, what is jointly created is what neither side can create alone. When viewed from this perspective, neither side is more powerful. This is the simple but profound insight that leads to the claim that the pie should always be divided 50:50.
There are three levers that are all equally effective in helping you get more: (1) make the deal a better deal by increasing the Total Value, (2) improve the value of your BATNA; (3) reduce the other side's BATNA. (BATNA=Best Alternative To a Negotiated Agreement.)
P39. Trying to determine which side contributes more to the pie is like asking which is more important to the Reese's peanut butter cup, the chocolate or the peanut butter? There's no answer to that question. You need both.
P45. Along with being a good deal for both sides, there's one last point I'd like you to notice about the Honest Tea deal. The two sides didn't have to know what the pie was in order to split it. They agreed to do an ex-post split of the pie—whatever it turned out to be. That is really helpful when you are negotiating in an environment of uncertainty. The two sides know there is a pie to be created, but neither side knows how big the pie will be, or each side has its own view, and they don't agree. They don't have to agree. So long as the pie can be measured after the fact, the two sides can agree to split it. That's a key insight we'll use in many of the upcoming examples.
P66. That's why, in my negotiation with Edward over the domain name, I started with a number that offered him less than half the pie. I guessed that he expected there would be some give-and-take, and so I played along. Edward was asking for more than the pie. Early on, I steered him toward the pie by framing things in terms of my BATNA (which was paying the $1,300 ICANN dispute resolution costs). That framing quickly brought his asking price to something below $1,300. At that point, we had done a little dance, and I felt it was time to explain the pie. I used the pie to explain why his proposed division was unfair. Then I made an ultimatum—take a fair deal or no deal—thought I was more polite. "I am willing to split the savings evenly with you, $650:$650, but that is as far as I will go." He tried to present his counteroffer as an ultimatum, but it was an arbitrary number and thus didn't stick. After I didn't respond to his counter, he thought things over and took the fair deal.
P118. The solution is to split the pie on a bottle-by-bottle basis. Recall that Coke's cost was 11 cents, and Honest Tea was paying 19 cents/bottle. The total savings are 8 cents/bottle. Instead of splitting an uncertain $20M pie, Honest Tea buys bottles from Coke at 15 cents each. That way, Coke makes 4 cents/bottle, and Honest Tea saves 4 cents/bottle. The pie is split evenly on each bottle, and this split evenly no matter how many bottles are sold.
P119. Much better is to make a contingent counteroffer. Anal says to the dealer: I understand you are expecting to spend some money on restoration. You might even be planning on spending money on authentication of this is by a famous artist. Here is what I propose. I'll sell you the painting for your offer of $1,000, and you share with me 50 percent of the proceeds for anything you get above $10,000. Anal is not protected even without knowing the value. The offer is fair to the dealer as well.
P129. If you don't know the size of the pie, agree to split it 50:50 later.
P132. When you're in such a situation, I advise you to ask for things others might not want or where exceptions can be justified. In my case, when I moved from Princeton to Yale, I asked Yale to provide a second mortgage in the event my wife had trouble finding a new job in New Haven. We could afford the mortgage on our desired house with our two salaries, but it would be dicey with just one. Even if Yale had to provide this offer to others, few would value it. Doing this once didn't mean they'd have to do it ten times. In other circumstances, you can get around the equal treatment rule by asking for something where exceptions are justified. A sales rep who spends time on the road might ask the company to pay for a cell phone, while the accounting person can't make the same argument.
P170. When someone asks for something in a negotiation, people say no because they think it is about dividing the pie. When it is simply asking for more money, that's true. When you can give them something they want other than cash, that's how you create pie.
P179. Find out what the other side really wants. Then, look for options to close the gap. Throw in the living room furniture, or move the closing date if you are buying a house. Or free oil change and tire rotation if you are buying a car.
P202. When Daylian is teaching executives how to negotiate, he has them think about negotiating with a young child who doesn't want to get out of a swimming pool. The parent says it is time to go home, and the kid starts crying. The kid is crying for two reasons. First, the kid doesn't want to leave the pool. Second, and the point typically missed, the kid doesn't have the vocabulary or language skills to make his or her case. The kid is frustrated. The kid is thinking: If only I could explain why the pool is so great, I'm sure I could persuade my parent to let me stay another ten minutes. What comes out is Waaah, waaah, sob. Daylian's solution is for the parent to make the kid's argument. "I know that being in the pool is great. You can do somersaults. You can enjoy feeling weightless. I'd like to spend my entire life underwater. But we have to eat at some point. You might not be hungry now, though you will be by the time we get home. And you aren't the only one coming home for dinner. That's why it's time to leave." Here is a big lesson: We can't always get our way. But we can always be understood. What works for children also works for adults.
P216. Think of the third option. Is there room for contingencies?
P223. Preparation work goes beyond planning to sell the pie. You also have to consider how you will expand the pie. What questions will you ask, and what information will you volunteer? How can you make an offer that gets heard? When it comes time to make the presentation, be flexible. Be sure you understand their objections, not just the objections you would have made. Confirm you have it right. To overcome objections, make the objections yourself. This demonstrates your understanding and allows them to hear your counterarguments—end by presenting their victory speech.
P240. My advice is backed up by experimental evidence. Professors Francesca Gino and Don Moore find that hiding the shorter deadline increased the chance of an impasse by more than half, from 23 percent up to 37 percent. Among those who reached a deal, the party with the shorter deadline got half the pie on average when the deadline was revealed but only a 43 percent share when it was kept hidden. Unless you frame things using symmetry, this isn't intuitive. Indeed, when participants with the shorter deadline were asked beforehand if revelation would help, almost 60 percent thought it would put them in a weaker position. They were wrong.
P256. Look out for anchoring numbers, and remember that precise numbers are sticky.