Never Split the Difference

Notes taken chapter by chapter while reading, written to help me recall the core aspects of each chapter rather than to summarize the whole book. Built to be crammed before a negotiation, not read once: each technique states the move first, then why it works, then the case it came from.

Quick reference#

Technique	Move	Chapter
FM DJ voice	Deep, soft, slow, downward-inflecting statements — not questions	2
Mirroring	Repeat the last 1–3 critical words, inquisitive tone	2
Confront without confronting	FM DJ voice → "I'm sorry..." → mirror → silence ≥4s → repeat	2
Tactical empathy	Name what they're feeling before you address content	3
Labelling	"It seems/sounds/looks like ___," then go silent	3
Accusation audit	List every bad thing they could say about you, before they say it	3, 6
Calibrated "No" question	Ask "Is now a bad time to talk?" instead of "got a minute?"	4
Force a "No"	Mislabel on purpose, or ask what they don't want	4
Revive a dead thread	"Have you given up on this project?"	4
Summarize to "that's right"	Label + paraphrase their situation back precisely	5
Give a range, not a number	Cite a comparable range instead of your own target figure	6

Chapter 1: The New Rules#

Buy time under a shock question#

• Move: pause → blink → breathe the heart rate down → a long, slow stare → answer with a question, not compliance.
• Why it works: a sudden, brutal demand is designed to make you react instantly. Slowing your own physiology first buys the half-second you need to respond instead of comply.
• Case: in a classroom role-play, "We've got your son, Voss. Give us $1 million or he dies." Voss ran the sequence, then answered with "How am I supposed to do that?" — followed by using the "kidnapper's" first name and an apology ("I'm sorry, Robert, how do I even know he's alive?") to seed warmth into an exchange built to bulldoze him.

Ordinary, not alpha#

• Move: don't posture as the smartest or toughest person in the room.
• Why it works: "I've always thought of myself as just a regular guy — hardworking and willing to learn, not particularly talented" (Voss on himself). Overconfidence closes you to new information; the point of a negotiation is discovery, and you can't discover anything you're sure you already know.

Listen instead of argue#

• Move: replace logic-and-argument tactics with intense listening.
• Why it works: human beings are fundamentally irrational, which is why Getting to Yes-style argument doesn't reliably work. People want to be understood and accepted; being listened to makes them listen to themselves more carefully, get less defensive, and become more willing to hear other views — the state you actually need them in to solve anything.

Chapter summary#

• Negotiation isn't a fight to avoid — it's the mechanism society runs emotional agreement through. Not liking it doesn't exempt you from it.
• You get what you ask for; you have to ask correctly, so claim the right to ask.
• The first move under pressure is always physiological (pause, breathe, slow down) before it's verbal.
• Ordinary self-image beats alpha posturing — it keeps you curious instead of certain.

Chapter 2: Be a Mirror#

Hold hypotheses, not assumptions#

• Move: keep several hypotheses live at once; let each new piece of information kill one and promote another.
• Why it works: assumptions calcify into blind spots — you decide what someone is before meeting them, then bend your own perceptions to fit. Really smart people often negotiate badly because they're sure there's nothing left to discover.
• Case: a bank-robbery ringleader kept deflecting to "the other guys" — a personal-pronoun tell (we/they, me/I: the less important he makes himself sound, the more important he probably is). There turned out to be one other robber, tricked into it, and a getaway driver who'd already fled — not the crew "we" implied. Until you know what you're dealing with, you don't know what you're dealing with.

Active listening#

• Move: stop planning what you'll say next; make the other person and their exact words your entire focus.
• Why it works: conscious working memory holds about seven items at once (Miller). In an argument-focused exchange both sides are stuck rehearsing their own case while the other talks — what looks like two people is really four, each listening only to the voice in their own head. True active listening quiets that voice and makes the other person feel safe enough to reveal what they actually need (monetary, emotional, or otherwise) — not just what they want. Wants are the easy, aspirational ask; needs are survival, and admitting them is vulnerable. Neither is the starting point — listening is.

Slow it down#

• Move: deliberately slow your own pace.
• Why it works: rushing reads as not-listening and burns rapport; someone talking isn't someone shooting.
• Case: a bank teller stalled robbers just long enough for the team to run plate numbers on nearby cars and identify a suspect, Chris Watts.

The FM DJ voice#

• Move: deep, soft, slow, downward-inflecting statements — not questions. "Joe's gone. This is Chris. You're talking to me now."
• Why it works: tone is read almost telepathically, before conscious thought — it's the cheapest, most effective lever in any exchange. A downward inflection signals "I've got this," letting you be flatly direct ("We don't do work-for-hire") without inviting a fight, because the tone already says I'm okay, you're okay, let's figure this out. Warmth and a smile are reflexive — they trigger the same back, and a positive frame makes both sides think faster and collaborate rather than resist.

Mirroring#

• Move: repeat the last one to three critical words the other person just said, in an inquisitive, downward tone.
• Why it works: mirroring (isopraxism) triggers a subconscious "we're alike" signal — we fear difference and trust similarity — so the other person elaborates instead of getting defensive. It lets you disagree without being disagreeable, useful against forceful, authority-through-aggression types where meeting force with force just produces bruised feelings.
• Case: "We chased your driver away?" mirrored back to Chris Watts kept him talking through a series of damaging admissions. Richard Wiseman's waiter study: mirroring a customer's order back earned 70% more in tips than positive-reinforcement phrases like "great" or "no problem."

Confront without confronting#

• Move: FM DJ voice → "I'm sorry..." → mirror, intending "please, help me understand" → silence, at least four seconds → repeat.
• Why it works: every mirror forces a reword, never a verbatim repeat — that's what gets you real clarity. ("What do you mean by that?" just triggers defensiveness.)
• Case: an assistant mirrors her boss — "I'm sorry, two copies?" ... "Anywhere?" — until his vague, irritated instructions become specific and composed, ending with him settling for two digital backups instead of printed, filed copies with nowhere to store them.

Chapter summary#

• Hold hypotheses, not conclusions — enter as discovery, not defence.
• Active listening beats arguing because it quiets the "schizophrenia" of both sides rehearsing their own case at once.
• Slowing the pace calms the process; rushing signals not-listening.
• The FM DJ voice — calm, downward, slow — is the cheapest, most effective tool in the room.
• Mirroring gets elaboration and disarms confrontation without you ever disagreeing out loud.

Chapter 3: Don't Feel Their Pain, Label It#

Tactical empathy#

• Move: name what someone is feeling before you address the content of what they're saying.
• Why it works: empathy is paying attention, asking what they're feeling, and committing to understanding their world — none of which requires agreeing with them.
• Case: a Harlem standoff with no phone line into the apartment — Voss talked through the door for six hours in the FM DJ voice, naming their likely state ("It looks like you don't want to go back to jail") instead of giving orders. All three fugitives came out silently once a sniper saw a curtain move.

Labelling#

• Move: "It seems like...", "It sounds like...", or "It looks like..." — then go silent.
• Why it works: the instinct is to tack on a follow-up question ("Where did you get that shirt?") — don't. Silence is what pulls the other person into revealing themselves; a label's job is to invite that, not to interrogate.

Neutralize the negative, reinforce the positive#

• Move: label a negative emotion to defuse it; label a positive one to lock it in.
• Why it works: emotion has two layers — the presenting behavior on the surface, and the underlying feeling driving it (a grumpy grandfather is often just lonely, not actually cranky). Naming the negative outright ("Look, I'm an asshole") clears it faster than smoothing over it.
• Case: having angered a top Canadian FBI official by skipping "country clearance," Voss opened the apology call with "Bless me, Father, for I have sinned" instead of an excuse — going straight at the mistake instead of around it.

Clear the road before advertising the destination#

• Move: label the fear or hesitation before you pitch your ask.
• Why it works: naming a fear drags it into the open and proves you already understand it — that alone drains its power.
• Case: a donor with "no interest in donating" was met with "I'm sensing some hesitation with these projects" — she immediately explained she wanted her gift tied to specific programming, then signed a check with no project even chosen: "You understand me... I trust you'll find the right project."

Accusation audit#

• Move: list every terrible thing your counterpart could say about you, and say it yourself, before they do.
• Why it works: hedging a negative ("I don't want this to sound harsh...") gives it more credence, not less. Defense lawyers do this properly — stating every weakness in the opening statement, "taking the sting out."
• Case: Anna and her negotiating partner, facing a soured relationship with a government client, opened by listing the accusations themselves: "You are the typical prime contractor trying to force out the small guy."

Chapter summary#

• Empathy doesn't require agreement — it requires proving you're listening, which is what earns you something you can use.
• The reasons a counterpart won't agree are usually stronger than the reasons they will — clear those barriers first; naming a negative removes its power, denying it adds power.
• After a label or a mirror, pause — the other party fills the silence.
• People want to be appreciated and understood — labels reinforce whichever perception you name.

Chapter 4: Beware "Yes" — Master "No"#

Treat "No" as information, not rejection#

• Move: hear "No" as one of: I'm not ready to agree / you're making me uncomfortable / I don't understand / I can't afford it / I want something else / I need more information / I want to check with someone else — never as rejection. Get it said early rather than hoping to avoid it.
• Why it works: saying "No" preserves the other side's autonomy — the deep, universal need to feel in control. Once they've been allowed to say it, their emotions calm and they can actually engage with your proposal. After a "No," pause, then ask a solution-based question: "What about this doesn't work for you?"

Read the three kinds of "Yes"#

• Move: don't trust a "Yes" until you know which kind it is — Counterfeit (an easier exit than "no," or a stall for more information), Confirmation (a reflexive answer to a black-or-white question, no promise of action), or Commitment (the real one, ending in a signature).
• Why it works: a negotiator who walks in all "me, me, me, ego, ego, ego" gets a counterfeit "Yes" purely to make him go away — he'll be told later about budget issues, changing conditions, the weather. Everyone, regardless of intensity, is driven by two primal urges: to feel safe and secure, and to feel in control. Satisfy those and you're in the door.

Force or invite a "No"#

• Move: ask "Is now a bad time to talk?" instead of "Do you have a few minutes?" — or force a "No" by deliberately mislabeling an emotion ("So it seems you're eager to leave your job" when they want to stay), or by asking directly what they don't want.
• Why it works: a question that prompts "No" gives the other person the feeling that they're in the driver's seat, which is what actually gets them engaged rather than stonewalling.
• Case: facing dismissal by a jealous supervisor, Marti didn't argue her case — she asked, "Do you want the FBI to be embarrassed?" He said "No." She followed with "What do you want me to do?" and kept her position, with him feeling entirely in control of that decision.
• If the other party won't say "No" no matter what, that's a warning sign — indecision, confusion, or a hidden agenda. No "No" means no go; walk away.

Revive a dead thread#

• Move: one line — "Have you given up on this project?"
• Why it works: it implies you're about to walk away on your own terms, which pushes the other side to reply fast and disagree, to protect their position: "No, our priorities haven't changed..."

Chapter summary#

• "Be nice" has hollowed politeness into a lubricant — death for a negotiator, whose power comes from getting the other side to disagree, draw boundaries, and define wants by naming what they don't want.
• Stop chasing "Yes" — pushing for it makes people defensive, and our own love of hearing it blinds us to that defensiveness.
• "No" usually means "wait," not rejection — it's the start of the negotiation, not the end.
• "Yes" is the destination, not the opener — asking for it too early raises guard.
• Force a "No" on purpose when stuck: mislabel an emotion, or ask a question that can only be answered "No."

Chapter 5: Get to "That's Right"#

Summarize to trigger "that's right"#

• Move: label plus paraphrase their situation back to them — precisely enough that they feel they've judged it correct of their own free will.
• Why it works: "that's right" means they own the conclusion; "yes" is just compliance.
• Case: getting kidnapper Sabaya to say "that's right" repeatedly drew out the talks long enough for a rescue operation, without pushing him to hurt his hostage.

Watch for "you're right" as a red flag#

• Move: if you hear "you're right," you haven't landed it — relabel the belief driving the behavior instead of repeating the same argument.
• Why it works: "you're right" is what people say to make you go away — agreement in theory, with no ownership, so the behavior doesn't change.
• Case: a lineman who loved flattening blockers, moved to linebacker (where the job is to avoid them), kept saying "You're right" to direct coaching with no change. Relabeling the belief — "You seem to think it's unmanly to dodge a block... cowardly to get out of someone's way" — got "That's right," and only then did the behavior shift.

Chapter summary#

• "Sleeping in the same bed and dreaming different dreams" — outward intimacy without shared understanding is the recipe for bad marriages and bad negotiations alike.
• The reflexive "yes"es and "you're right"s thrown out early in any conversation are social lubricant, not real alignment.
• A summary that earns "that's right" = a label + a paraphrase: identify, rearticulate, and emotionally affirm "the world according to them."

Chapter 6: Bend Their Reality#

Don't compromise#

• Move: treat "no deal" as better than a bad deal — refuse to "meet halfway" by default.
• Why it works: we compromise because it's easy and face-saving, not because it's right — most negotiators are driven by fear of pain, not by their actual goals. Creative solutions come from risk and friction, not from accommodation.
• Case: a husband and wife who can't agree on shoe colour "meet halfway" — one black shoe, one brown. Not a fair middle ground; the worst possible outcome for both.

Deadlines as leverage#

• Move: treat deadlines — yours and theirs — as almost never as ironclad as they feel, and find the other side's real deadline.
• Why it works: a deadline can trick you into valuing "a deal now" over "a good deal" — the pressure comes from an imagined future consequence, not the deadline itself. Letting that fear drive you is taking yourself hostage. Revealing your own deadline, contrary to old-school advice, gets better deals (Moore): it lowers the risk of impasse and speeds real concessions, because when the negotiation ends for one side, it ends for both.
• Case: Haitian kidnapping ransom demands spiked on Mondays and urgency spiked toward the weekend — the kidnappers were ordinary thugs racing to get paid by Friday, not political actors. Knowing that self-imposed deadline let the team stall until Thursday or Friday and offer far less than the $150,000 asked. (Herb Cohen's counter-case: hiding his own deadline in a Japan negotiation, he was stalled with a week of parties and lost leverage in the final-hour rush to the airport.)

No universal "fair"#

• Move: never assume the other side reasons like you do — that's projection, not empathy.
• Why it works: in the Ultimatum Game (a proposer splits $10 with an accepter; a refusal means both get nothing), no single split is ever the group consensus on "fair," and the rational move — offer $1, which should never rationally be refused — is almost never made, and is often refused when it is. Every choice on both sides turns out to be emotional, not rational.

The F-word: "fair"#

• Move: know which of three uses you're facing before you react to the word "fair."
  • As an attack — "We just want what's fair" — a judo move meant to destabilize you into a defensive concession. Don't concede on reflex; apologize and offer to rewind instead.
  • As a deflection (e.g. "we've given the players a fair offer," instead of opening the books) — mirror it back, "Fair?", then label it: "It seems like you're ready to provide the evidence that supports that."
  • As a pre-empt — say it yourself, early: "I want you to feel like you're being treated fairly at all times — stop me any time you feel I'm being unfair."
• Why it works: people comply when they feel respected and lash out when they don't — the sting of unfairness outweighs the rational value of money. In the Ultimatum Game, accepters reject almost anything under half, and effectively anything under a quarter.
• Case: Robin Williams voiced the genie in Aladdin for a cut-rate $75,000 and didn't get angry until the film made $504 million — the anger was about perceived unfairness, not the money itself.

Sell the emotional driver, not the service#

• Move: identify and speak to the feeling behind the purchase, not the literal service.
• Why it works: a babysitter sells a relaxed evening, a furnace salesperson sells cozy family time, a locksmith sells a feeling of security — know the driver and any deal can be framed in language that resonates.

Framing changes value#

• Move: control the frame, not just the number.
• Why it works: decisions are irrational but not random — they follow consistent, learnable patterns. Loss aversion is the strongest one: it's not enough to show what they'd gain — they need to feel there's something concrete they'd lose if the deal falls through.
• Case: a used Swiss-flag coffee mug is worth ~$3.50 to a buyer and $5–7 to its owner asked to sell it — same mug, different frame. $20 for a three-minute coffee errand feels like a $400/hour win — until you learn the requester made a million dollars off it, and the same $20 feels like a rip-off.

Anchor their emotions with an accusation audit#

• Move: stack an accusation audit before delivering bad news, so the eventual ask lands as relief, not insult.
• Why it works: pre-loading every fear reframes the real ask around loss aversion — not "you're losing money" but "don't lose this to someone else."
• Case: needing to cut contractors from $2,000/day to $500/day, Voss opened each call with "you're going to think I'm a lousy businessman... he might even have lied to me" before framing the new rate as an opportunity "I wanted to bring to you first." Every contractor took the deal.

Let them anchor first#

• Move: let the other side name a number first when you lack market information.
• Why it works: neither side ever has perfect information — opening blind risks the winner's-curse feeling of having agreed to too little, while letting them go first risks nothing you didn't already risk by not knowing the market.
• Case: Jerry opened a salary negotiation at $110,000 (a 30% personal raise) without knowing the firm was already starting everyone else in the program at $125,000. (Not an absolute rule: against a rookie counterpart, or an equally informed pro, opening first can still work — but a reputation for always beating your counterpart eventually means nobody wants to deal with you.)

Give a range, not a number#

• Move: cite a comparable "ballpark" instead of your own target figure — "at places like X Corp, people in this role get $130,000 to $170,000" instead of "I'm worth $110,000."
• Why it works: an extreme anchor unconsciously drifts the other side's expectations toward it — some go straight to their limit. Columbia Business School research found job applicants who named a range got higher overall offers than those who named a single number, especially with a "bolstering range" (the low end is what they actually wanted). Expect the other side to anchor to the low end of whatever range you give.

Chapter summary#

• Never split the difference — compromise is usually the worst outcome dressed up as fairness.
• Deadlines pressure both sides; revealing your own often gets a better deal than hiding it.
• "Fair" is the most emotionally loaded word in a negotiation — know which of its three uses you're facing.
• Frame controls value more than the underlying number does; loss aversion is the strongest frame available.
• Let the other side anchor first when you lack market information; give a range instead of a number when you have to go first.

Chapter 7: Create the Illusion of Control#

Notes pending.

Chapter 8: Guarantee Execution#

Notes pending.

Chapter 9: Bargain Hard#

Notes pending.

Chapter 10: Find the Black Swan#

Notes pending.

Acknowledgments#

Notes pending.

Appendix: Negotiation One Sheet#

Notes pending.