The recruiter names a number, you name a higher one, and somebody says, “Shall we meet in the middle?” Chris Voss’s book tells you to refuse. This never split the difference summary gives his method in one line: listen first, name what the other side feels, ask “how” questions, and let them solve your problem for you.

Below are the techniques chapter by chapter, the ones that carry over to a salary talk or a client call, and what the research behind them says. We also read 24 other summaries, and one statistic they pass on does not mean what they say it means.

What We Found

Of 24 summaries of Never Split the Difference we read on 14 and 15 November 2026, 11 repeat the 7-38-55 rule as fact, and none mentions that Albert Mehrabian, the psychologist behind it, limits it to feelings and attitudes.

The rule says 7 percent of a message comes from the words, 38 percent from tone of voice and 55 percent from body language. Mehrabian’s own page says the numbers came from experiments on feelings and attitudes, meaning like and dislike, and that outside that case “these equations are not applicable.”

Voss still teaches it. In a post under his name on his firm’s site, dated 9 May 2022, he calls 7:38:55 the ratio his method teaches and gives no study for it.

The rest of the book travels better. All 24 summaries cover mirroring, 23 cover labelling and 23 cover calibrated questions. The harder tools drop out: 12 cover the Ackerman model for haggling over price and 11 cover the accusation audit.

What Is Never Split the Difference About?

Never Split the Difference is a 2016 negotiation book by Chris Voss, a former FBI hostage negotiator, written with the journalist Tahl Raz. It takes the listening tactics of crisis talks and applies them to business and everyday deals, on the view that feelings, more than logic, decide a negotiation.

Publishers Weekly’s listing gives the publisher as Harper Business and the length as 288 pages, and its review of 8 February 2016 called the book “an intriguing approach to business and personal negotiations.”

That puts it on the same shelf as an older book. Our How to Win Friends and Influence People summary argued for keeping Carnegie’s listening rules, and Voss applies the same habit under pressure. For a different view of what you owe the other side, see our piece tracing Adam Grant’s Give and Take to its 15 studies.

Never Split the Difference Summary: The Techniques by Chapter

Graham Mann’s chapter notes, the never split the difference summary we used for chapter placement, list ten chapters. The tools start in chapter 2, and we have followed those notes for where each one sits.

Chapter 2, Be a Mirror. Repeat the last one to three words the other person said, then go quiet so they keep explaining. Voss pairs this with a deep, slow, reassuring voice he calls the late-night FM DJ voice.

Chapter 3, Don’t Feel Their Pain, Label It. Name the emotion you hear, as an observation, and stop talking. The accusation audit belongs here too: before the meeting, list the worst things the other side could say about you, and say them first.

Chapters 4 and 5, “No” and “That’s Right.” Voss says to stop pushing for an early yes. A no usually means the person wants to keep things as they are, and it opens the conversation. The reply to work for is “that’s right,” which follows an accurate summary of their position. “You’re right” is often a polite way of ending the discussion.

Chapter 6, Bend Their Reality. Deadlines can trick you into valuing a deal now over a good deal. Voss leans towards letting the other side name the first price, though not as a fixed rule, warns that anyone offered a range will take its low end, and gives his salary advice, which we come to below.

Chapter 7, Create the Illusion of Control. Calibrated questions are open questions that start with “what” or “how” and cannot be answered with a yes or a no. The best known is “How am I supposed to do that?”, which gets the other side working on your problem.

Chapter 8, Guarantee Execution. The rule of three is getting the same commitment three times in one conversation: the agreement, a summary they confirm, and a question about how it will be carried out. The 7-38-55 rule sits in this chapter, beside the idea that liars use more words than people telling the truth.

Chapter 9, Bargain Hard. This is the price chapter, with the Ackerman model and three negotiator types: Analyst, Accommodator and Assertive. For the harder school of influence, our 48 Laws of Power summary sorts those laws into 17 to use at work, 17 to handle with care and 14 a manager should ignore.

Chapter 10, Find the Black Swan. Black Swans are unknown unknowns, facts that would change the negotiation if you had them.

Technique Chapter Summaries covering it (of 24) Research we could trace
Mirroring 2 24 van Baaren, 2003: a waitress who repeated orders back got larger tips
Labelling 3 23 Lieberman, 2007: naming an emotion on a face lowered the amygdala response in 30 adults
Accusation audit 3 11 None found by us
Calibrated questions 7 23 None found by us
7-38-55 rule 8 11 Mehrabian: applies to feelings and attitudes only
Ackerman model 9 12 None found by us; Harvard’s Program on Negotiation reports that first offers anchor in the opener’s favour
Black Swans 10 18 None found by us
A consultant listening to a client in a members' club lounge, the tone-of-voice side of Chris Voss's negotiation method

What Is the 7-38-55 Rule, and Does It Hold Up?

The 7-38-55 rule is the claim that 7 percent of a message is carried by words, 38 percent by tone of voice and 55 percent by body language. It holds up only in the narrow case it was measured in, someone expressing like or dislike. As a rule for negotiation in general, we found no study behind it.

The numbers come from two papers Mehrabian co-wrote in 1967, one with Morton Wiener and one with Susan Ferris, both listed on his publications page. Wikipedia’s account says both studies dealt with positive or negative emotion carried by single spoken words, that the 55 came from black-and-white photographs of faces, and that the famous figures were made by combining the two results. It also notes that only women took part.

MasterClass’s article on the rule credits it to Mehrabian’s 1971 book Silent Messages and says Voss applied the research to negotiation. None of the 11 summaries that state the rule says where the numbers stop applying.

The book itself is more careful than its summaries. In chapter 8, Voss credits the rule to two Mehrabian studies on what makes people like or dislike someone, then uses it to watch for words and tone that do not match. The summaries kept the percentages and dropped that context.

Summary Page date Accusation audit Ackerman model Black Swans States 7-38-55 Limit on the rule
Science of People Updated 16 Sep 2025 No Yes Yes No
Books That Slay 27 Mar 2025, updated 27 Apr 2025 Yes No No No
Tal Gur (elevatesociety.com) 15 Dec 2018, modified 23 Apr 2023 Yes No Yes No
LifeClub (Rob Liu’s Blog) Undated (footer 2019) No No No No
Shortform Undated Yes Named only Yes No
Oberlo 10 Dec 2020 No No Yes Yes None
Runn 9 Jun 2022 No No Yes Yes None
The Investor’s Podcast Network, executive summary 15 Apr 2023 No No Yes No
Graham Mann book notes 30 Jun 2026, updated Jul 2026 Yes Yes Yes Heading only
Philipp Hauer’s Blog 2 Jan 2023, updated 6 Feb 2023 Yes Yes No No
Freshsales blog Undated No Yes Yes Yes None
Develop Good Habits 8 Sep 2023 Yes Yes No Yes None
Nabil Murad (Substack) 22 Dec 2025 No Yes Yes No
Professional Leadership Institute Undated No No No Yes None
South (hireinsouth.com) 28 Jul 2025 No Yes No No
Paul Minors 26 Jul 2019 Yes Yes Yes Yes None
Sumizeit Undated Yes No Yes No
ClickUp blog Undated No No Yes Yes None
Fluently Undated Yes No Yes No
Success Summaries Undated No Named only Yes No
The Investor’s Podcast Network, blog summary 15 May 2019, modified 30 Jul 2021 Yes No Yes Yes None
Daniel Miessler (Goodreads author blog) 20 Aug 2019 No No Yes Yes None
Anant Jain (Medium, Low Pass Filter) 16 Dec 2018 No Yes Yes Yes None
URSummary 20 Jul 2025, updated 25 Oct 2026 Yes Yes Yes Yes None

A number that sounds precise gets copied because nobody goes back to the page it came from. We found the same habit when we read 31 pages on Mark Cuban’s net worth, when we traced Shaq’s net worth figure to one site, and when 11 of 32 pages cut or reworded Theodore Roosevelt’s Man in the Arena.

Popular books do it as well as web pages. Our Outliers summary checked 14 of Gladwell’s claims against the studies and six did not hold up, and our reading of The Courage to Be Disliked set its core ideas beside Adler’s own books.

What survives of the rule is small and useful. When someone’s words say yes and their voice says something else, ask. The chapter notes give Voss’s instruction for that moment as a label that names the mismatch, and it works without any percentage.

What the Research Says About Mirroring and Labelling

Mirroring has one well-known field test. In a 2003 paper in the Journal of Experimental Social Psychology, Rick van Baaren and colleagues had a waitress repeat customers’ orders back to half of them. She received significantly larger tips from those she mimicked.

The setting was a restaurant and the stake was a tip. The paper’s conclusion is that mimicry makes people more generous. Nobody has shown us that it moves a procurement manager, so treat mirroring as a way of getting more information.

Labelling rests on brain imaging. UCLA’s 2007 release on a study in Psychological Science describes 30 adults who looked at faces showing emotions while in a scanner. When they picked the word for the emotion, the response in the amygdala fell.

The participants were naming an emotion on a stranger’s photograph. Voss extends that to naming your counterpart’s feeling out loud, to their face, in the middle of a dispute. The step is reasonable, and the laboratory finding is narrower than the claim built on it.

Summaries tend to keep the name of a technique and lose what stands behind it. Our 7 Habits of Highly Effective People summary found the same pattern across 18 summaries, which kept the habit names and dropped the tools.

A manager and a younger colleague talking on a bench in a headquarters atrium, the salary conversation Chris Voss's techniques are used for

Which Techniques Work in a Salary Negotiation?

Three transfer well to a salary negotiation: calibrated questions, labels, and Voss’s advice to negotiate the terms around the salary and the measures of success for the role. Mirroring helps you learn more. The Ackerman model and the advice to let the employer name the first number fit worst, because pay talks are rarely a simple haggle.

Voss’s salary section, as the chapter notes record it, has three parts. Be pleasantly persistent on terms other than salary. Once the number is set, agree how success will be measured and what the next raise depends on. And ask what it takes to be successful in the role.

That question is the most useful thing in the book for an employee. Voss’s point is that it gives the manager a stake in your result. The answer also gives you the list you will bring to the next review. Our guide on how to ask for a raise covers what to bring into that room.

On who speaks first, the research and the book part ways. Harvard’s Program on Negotiation, summarising work by Adam Galinsky, reports that making the first offer gives a bargaining advantage because it anchors the talk, and that the exception is when the other side knows far more than you do.

In a job offer the employer usually does know more: the band, the budget and what the last hire got. That is the case where Voss’s caution fits. If you have done the market research and know your figure, the evidence favours saying it. Our piece on how to negotiate salary walks through that preparation.

A label fits a pay talk well. “It sounds like the budget for this role was set before we spoke” is hard to take offence at, and it invites the real constraint. A calibrated question then puts it to work: “How could we close the gap by the six-month review?”

What Carries Over to Client Talks

With clients the accusation audit earns its place, and it is the tool most summaries skip. Before a renewal or a price rise, write down what the client could fairly say against you, such as a late delivery or an invoice that surprised them, and open with it.

Voss’s reasoning, in the chapter notes, is that the accusations sound exaggerated when said aloud, which takes the sting out and tends to make the other side argue the opposite. If the client answers that it was not that bad, listen closely to what they say next.

“How am I supposed to do that?” needs care. Said with an edge, it is a refusal, and clients hear it. Our guide to dealing with difficult clients argues for structure over patience, and this question works only inside one.

The rule of three is the cheapest habit to copy. A client who says yes in the meeting has agreed once. Summarise it back and wait for “that’s right.” Then ask how it will be carried out on their side, and who signs.

The test we would apply to any of these is whether the tactic rests on what the person said themselves. It is the test we ran on the Cristiano Ronaldo diet, where 7 of 25 famous claims were his own words, and on the Jennifer Aniston workout, where 15 of 20 were.

What Is the Ackerman Model?

The Ackerman model is Voss’s plan for haggling over a price. Set a target, open at 65 percent of it, then move to 85, 95 and 100 percent in shrinking steps, using calibrated questions between offers. End on a precise, non-round number and add a small non-cash item to show you have reached your limit.

We did not find a published test of the full sequence. Take it as a negotiator’s working method, and keep the preparation: a target, the steps and the non-cash item, all decided before the meeting.

It suits a one-off purchase where price is the only question: a car, a lease, a piece of equipment. It suits a salary or a retainer badly. Opening at 65 percent of your target with someone you will work with for years starts that relationship with a lowball.

How We Did This

On 14 and 15 November 2026 our newsroom searched for summaries of Never Split the Difference and kept every free page that loaded and summarised the whole book: 24 pages, from company blogs, summary sites and personal notes, including three of the US top five for the search. For each we recorded whether it covers six techniques and whether it states the 7-38-55 rule with any limit.

A page counted as stating the rule only if it gave the percentages; one gave 7 and 93 percent under the rule’s name and was counted. One page that names the rule in a heading without them was left out of the 11. A paywalled newsletter and one page that would not load were excluded, and two of the 24 are separate pages on one site. Chapter numbers in the never split the difference summary above follow Graham Mann’s notes. The full dataset, with a quote for every row, is available on request.

Frequently Asked Questions

Who wrote Never Split the Difference?

Chris Voss, a former FBI hostage negotiator, with the journalist Tahl Raz. Publishers Weekly lists it at 288 pages from Harper Business and reviewed it in February 2016. If your next read is a novel, we also counted the picks in Reese Witherspoon’s book club.

Does the 7-38-55 rule come from Chris Voss?

No. It comes from two 1967 studies co-written by Albert Mehrabian, a UCLA psychology professor, and his 1971 book Silent Messages. Voss applied it to negotiation. Mehrabian’s own page restricts the figures to messages about feelings and attitudes.

Is Never Split the Difference useful if you are not in sales?

Yes, for anyone who has to ask for something. Labels, calibrated questions and the accusation audit work in a pay review, a client renewal or a disagreement with a co-founder.

Refusing to split the difference comes down to staying in the conversation long enough to learn what the other side needs. A mirror, a label and one good “how” question will do that in most rooms, and you can leave the percentages out.

Latest From The Executive Insight