Principled negotiation, tactical empathy, and the science of making deals. From Fisher and Ury to Chris Voss to the contemporary research on what actually works at the table.
Negotiation is the conversation by which two or more parties with overlapping but not identical interests reach a joint decision. Most adult coordination — buying a house, accepting a job, making peace, sharing housework — is some form of it.
The discipline as a research subject is younger than most assume. The Harvard Program on Negotiation was founded in 1979. The first edition of Getting to Yes appeared in 1981. The empirical experimental literature on negotiation is largely a product of the last forty-five years.
This deck covers the major frameworks (positional vs principled, BATNA, ZOPA, anchoring, framing), the modern operationalised approaches (Voss, the Diamond model), the research findings that have held up, and the practical applications.
The fundamental taxonomy. Distributive (zero-sum) negotiation: dividing a fixed pie. Most price negotiations look this way. One party's gain is the other's loss. Tactics: anchoring, claim-the-value, hide your reservation point.
Integrative (variable-sum) negotiation: enlarging the pie. Most relationship-rich and multi-issue negotiations have this potential. Different parties value different things differently; trades that exploit those differences create joint value.
Most real negotiations are mixed. A salary negotiation has a distributive component (today's pay) and an integrative one (signing bonus, equity, vacation, project assignment, growth path). The skill is recognising which moves create value vs claim it.
Best Alternative to a Negotiated Agreement. Coined by Roger Fisher and William Ury in Getting to Yes (1981). Your BATNA is what you'll do if this negotiation fails — and it determines what you should accept.
The principle: never accept an agreement worse than your BATNA. The corollary: improve your BATNA before negotiating. A second job offer changes salary negotiations more than any tactic.
BATNAs are often badly estimated. People underestimate their BATNA before a negotiation (which makes them too willing to settle) and overestimate it during difficult moments (which makes them walk away from acceptable deals).
Two related concepts: WATNA (Worst Alternative to a Negotiated Agreement — useful for risk-assessment) and reservation point (the worst deal you'll accept; ideally just better than your BATNA).
Zone of Possible Agreement. The overlap between your reservation point and the other party's. Any deal in the ZOPA leaves both parties better off than walking away.
If your minimum is $80,000 and the employer's maximum is $90,000, the ZOPA is $80–90K and any salary in that range is a possible deal. If the employer's max is $75K, there is no ZOPA — no deal benefits both parties.
Most of distributive negotiation is about claiming as much of the ZOPA as possible. Anchoring, time pressure, and information asymmetry are the major tools. Most of integrative negotiation is about creating a ZOPA where none initially appeared, by trading across multiple issues with different valuations.
The 1981 book that founded the academic-popular negotiation literature. Roger Fisher (Harvard Law) and William Ury (Harvard Negotiation Project) proposed principled negotiation against the prevailing positional-bargaining tradition.
Soft on people, hard on the issue. Don't attack the other person; attack the problem.
Positions ("I want $90K") are claims; interests ("I want financial security and career growth") are the underlying needs. Multiple positions can satisfy the same interest.
Generate multiple possible deals before evaluating them. Brainstorming is undervalued in negotiation prep.
Anchor proposals in independent standards — market rates, expert opinions, precedent — not in your or their stubbornness.
The first concrete number put on the table in a negotiation has disproportionate influence on the final outcome. Tversky and Kahneman's anchoring research (1974) showed that even arbitrary numbers — a randomly spun roulette wheel — bias subsequent quantitative judgements.
In negotiation, this means: whoever speaks first, with a credible-but-aggressive number, sets the reference point around which the eventual settlement clusters. Galinsky and Mussweiler's research (2001+) shows that aggressive opening offers correlate with better outcomes for the offerer, in ranges where the offer remains within the ZOPA.
The traditional advice — "never give the first number" — is empirically wrong when you have decent market information. Make the first move with an aggressive but defensible anchor; expect counter-pressure; be prepared to defend the basis.
Kahneman and Tversky's prospect theory (1979) showed that losses are felt about 2× as strongly as equivalent gains. This shapes negotiation. The same outcome framed as a loss vs as a foregone gain elicits very different responses.
Practical applications: present concessions as already-made (so they feel like baseline, not gifts); present what the other party stands to lose by not agreeing, not just what they gain by agreeing; in mediation, frame the alternatives to settlement (litigation cost, time, uncertainty) in loss terms.
The endowment effect — people overvalue what they already possess — is a related cognitive bias. House sellers consistently overprice; employees underestimate the cost of the job offer they didn't take. Recognising this in yourself improves your negotiation; recognising it in the other party improves your strategy.
Chris Voss, former FBI lead international hostage negotiator. Never Split the Difference (2016). Combines hostage-negotiation field experience with academic behavioural-economics findings.
Central concept: tactical empathy. Demonstrating that you understand the other side's perspective, emotions, and constraints. Not agreement — comprehension. The negotiator who better articulates the other party's view gains permission to be heard on their own.
Mirroring — repeat the last 1–3 words of what they said, with rising inflection. Triggers elaboration.
Labelling — name the emotion you observe ("It seems like you're frustrated about the timeline"). Diffuses negative emotion; encourages corrections.
Calibrated questions — "How am I supposed to do that?" Forces the other party to solve your problem. The "no-oriented question" — "Is now a bad time?" — produces more honest engagement than yes-oriented questions.
Voss's tools have substantial empirical support and are now standard in business negotiation training.
Albert Mehrabian's 1967 research is widely (mis)cited as showing that communication is "7% words, 38% tone, 55% body language." The finding applied only to the specific case of communicating feelings when verbal and nonverbal channels conflict.
For negotiation specifically: yes, nonverbal communication matters. Tone, body language, and facial expression carry information about commitment, ambivalence, and confidence. But the specific 7-38-55 numbers are not the right takeaway and Mehrabian himself has repeatedly warned against the misapplication.
What is robust: tone matters more in conflict than in cooperation; video calls give better outcomes than audio-only or email for difficult negotiations (Drolet & Morris 2000); face-to-face negotiation builds rapport and produces more integrative outcomes than electronic media — especially for first interactions.
The Thomas-Kilmann Conflict Mode Instrument (1974) plots five styles on assertiveness × cooperativeness:
Competing (high assertive, low cooperative) — distributive bargaining; pushing for one's position. Useful in time-constrained, low-relationship contexts.
Accommodating (low assertive, high cooperative) — yielding; preserving relationship at cost to outcome. Useful when issue matters more to other party.
Avoiding (low both) — postponing; declining engagement. Useful when negotiation is premature or BATNA is strong.
Compromising (mid both) — meeting in the middle. The default fallback; rarely produces optimal value.
Collaborating (high both) — integrative; working through differences. Time-intensive; produces best outcomes in repeated games and complex multi-issue contexts.
Most professionals have a default style and a less-developed second. Strong negotiators move flexibly across styles based on context.
Single-issue negotiations are inherently distributive. Multi-issue negotiations open the door to value-creation. The key technique: logrolling — trading low-priority concessions for high-priority gains.
If you value salary highly and the company values flexibility highly, the deal might be: company gives you the salary, you accept on-call rotation. Both parties end up better than splitting the difference on each issue separately. This is integrative bargaining in its simplest form.
Present multiple proposed offers simultaneously rather than sequentially. "Would you prefer Option A (high salary, weekly travel) or Option B (lower salary, no travel)?" Forces the other party to reveal their priorities through their choice. Leigh Thompson's research has shown this is one of the highest-leverage moves available in complex negotiations.
Negotiation outcomes correlate strongly with the quality of information each party has — about the other's BATNA, reservation point, time pressure, alternative options, internal politics. The party with better information consistently captures more value.
Two information-gathering moves:
Pre-negotiation research. Comparable transactions, market rates, the other party's recent deals, the decision-makers and their incentives, public statements about strategy.
In-negotiation question-asking. Voss's calibrated questions; Fisher and Ury's "why" and "why not"; Stuart Diamond's Getting More framework explicitly emphasises asking before claiming.
The corollary: protect your own information. Reveal what helps the deal; don't reveal your reservation point until necessary; reveal your interests but think carefully before revealing the relative weights.
Negotiation power is largely about alternatives. The party with better alternatives outside the negotiation has the leverage. This is the same insight as BATNA, applied at a higher level.
Power can also be situational: time pressure (whoever can wait longer has power), scarcity (the supplier with unique capability has power), commitment (the party who has publicly committed has power, in some contexts), social proof (the party with apparent support of others has power).
The Galinsky-Magee research programme on power and negotiation shows that even feeling powerful (a primed sense of agency, control) improves negotiation outcomes — high-power individuals make more aggressive opening offers, claim more value, and emerge from negotiations more satisfied. Whether the priming-effect findings will hold up to full replication is uncertain; the underlying claim that perceived power matters has held up in multiple traditions.
Negotiation styles vary substantially across cultures. The major dimensions (Hofstede; Erin Meyer's The Culture Map):
Communication style — direct (US, Germany, Israel) vs indirect (Japan, China, Indonesia). What's stated explicitly vs what's understood.
Trust-building — task-based (US, Northern Europe — trust grows through delivered work) vs relationship-based (Brazil, China, much of Africa — trust requires extended personal acquaintance).
Decision-making — top-down (most of Asia, much of Latin America) vs consensual (Japan, Germany, Sweden). Affects who needs to be present at the table.
Time orientation — sequential (US, Germany — time is linear; deadlines are real) vs synchronic (much of Latin America, Middle East — time is flexible; relationships matter more than schedule).
The risks of cross-cultural negotiation are mostly about misreading these signals. The strong fix is preparation and humility — ask, don't assume.
The negotiation most readers will face most often. The empirical findings:
Negotiating raises starting salary. Studies (Babcock, Laschever) consistently show employees who negotiate make 7–11% more than those who don't, with effects compounding across careers.
Women negotiate less often and face more social backlash when they do. The gap has narrowed but not closed. Bowles, Babcock, Lai (2007) — the canonical study on the social-cost asymmetry.
Specific anchors outperform ranges. "$87,500" elicits better counters than "$85–90K." Mason et al. 2013.
Negotiate the package, not just salary. Signing bonus, equity, vacation, remote work, professional development budget — different items have different values to different employers and different costs to fund. Logrolling territory.
The negotiation literature on M&A, complex commercial deals, and large procurement is denser than the personal-negotiation literature but less popular.
Key concepts: deal value vs accounting value (most M&A creates accounting goodwill that doesn't reflect strategic value); walk-away thresholds set early and held; deal protection mechanisms (break-up fees, exclusivity periods, no-shop clauses); structural creativity (earnouts, contingent value rights, escrows).
Bruce Wasserstein's Big Deal (1998) and the Harvard Business School M&A literature are the operational references. The empirical post-deal literature consistently finds that ~50% of large M&A deals destroy shareholder value — usually because the negotiated price exceeded the integrated synergy, or because negotiation pressure produced concessions in deal terms that disadvantaged the acquirer post-close.
The longest-running serious negotiation tradition is diplomacy. Key cases study guides: Camp David Accords (1978; Carter, Sadat, Begin) — multi-day single-text negotiation in isolation. Iran nuclear deal / JCPOA (2015) — multi-party, technically dense, multiple-round structure with deadline brinkmanship. Northern Ireland / Good Friday Agreement (1998) — multi-party, identity-based, with creative ambiguity.
Major contributors: Henry Kissinger's White House Years on shuttle diplomacy. James Sebenius, Lawrence Susskind's Negotiating Gridlock. William Ury's post-Getting to Yes work on Iran, Venezuela, Korea.
Lessons from diplomatic negotiation that transfer: the value of single-text procedure (one shared draft, both parties propose edits); the importance of back-channel communications; the centrality of internal coalition management on each side; the use of deadlines as forcing devices.
The pandemic-accelerated shift to remote negotiation has substantially changed the landscape. The empirical findings:
Email negotiations produce more impasses than face-to-face for matched problems (Morris et al. 2002). Lack of synchronous tone and rapport drives more positional bargaining.
Video is closer to face-to-face than audio-only, especially in first interactions. Video reduces stereotype activation and supports rapport-building (Bohannon et al. 2013).
Pre-negotiation rapport calls (5–10 minutes of small talk before the substantive video call) substantially improve outcomes — more so than the equivalent rapport time in person (Drolet & Morris).
Asynchronous negotiation (mostly email/Slack) favours the party with more time to draft. The opposite of face-to-face's bias toward the more verbally facile party.
Stone, Patton, and Heen's Difficult Conversations (1999, revised 2010) — Harvard Negotiation Project's framework for emotionally-charged interpersonal negotiation. Three "conversations" happen in every difficult exchange: the What Happened conversation (factual disagreement), the Feelings conversation (the emotional content), and the Identity conversation (what the situation says about each party).
Most failed difficult conversations get stuck on one layer. Strong difficult-conversation handling involves explicitly surfacing all three.
The Crucial Conversations (Patterson, Grenny, McMillan, Switzler, 2002) framework offers parallel tools: state your facts, then your story (your interpretation), then ask for the other's view. The technique sounds simple; it requires practice to use under emotional pressure.
Mediation introduces a neutral third party who assists the negotiation without imposing a decision. Distinct from arbitration (where the third party decides) and litigation (where a court decides).
Mediation works because: it changes the audience (parties speak more carefully when a neutral observer is present); it provides confidential reality-testing (the mediator can probe each side's BATNA privately); and it maintains the relationship (parties don't directly attack each other).
The major mediation organisations: JAMS (Judicial Arbitration and Mediation Services, US), AAA (American Arbitration Association), CEDR (Centre for Effective Dispute Resolution, UK). The Harvard Mediation Program's caucus model — mediator alternates between separate sessions with each party — is the dominant US workflow. The transformative mediation tradition (Bush & Folger, 1994) emphasises relational repair over settlement.
Tactics with documented effectiveness, and their counters.
Anchoring — first aggressive offer. Counter: re-anchor with a defensible counteroffer; do not negotiate against the original anchor.
Good cop / bad cop — one party tough, one soft. Counter: name the dynamic; insist on speaking with one decision-maker.
Take it or leave it — false ultimatum. Counter: leave it. Real ultimatums are rare; testing them costs you nothing if your BATNA is solid.
Limited authority — "I have to check with my boss." Counter: insist on real authority at the table, or use the same tactic.
Krunch — "You'll have to do better than that." Counter: ask what specifically is the problem; re-anchor on objective criteria.
Time pressure — artificial deadline. Counter: confirm if it's real; offer to walk if it is.
The general counter to high-pressure tactics: name them, calmly. Most lose effectiveness once recognised.
The ethical line in negotiation runs roughly along: misrepresenting facts (illegal in most jurisdictions and unethical) vs concealing information (generally allowed) vs strategic positioning (allowed and expected).
Specific issues:
Bluffing about your reservation point or BATNA — generally accepted in commercial negotiation, often defended as "norm puffery."
Misrepresenting material facts — illegal under fraud law in most jurisdictions, even if delivered with plausible deniability.
Concealing material information the other party reasonably needs — the legal lines vary; the ethical case for concealment is weaker than the case for not volunteering.
Reneging on agreements after the deal — clearly unethical and usually illegal, but unfortunately common in low-trust environments.
Robert Cialdini's research on reciprocity and Adam Grant's Give and Take argue that long-run reputation matters more than short-run claim. Most repeat-game negotiators agree.
Most negotiation outcomes are determined before anyone speaks. The strong recommendation across all serious frameworks: spend more time preparing than negotiating.
The 7-point checklist (Sebenius, Lax):
1. Your interests (the underlying needs, not positions).
2. The other side's interests (your best estimate; iterate as you learn).
3. Options — multiple possible deals.
4. Legitimate criteria — market rates, precedent, expert standards.
5. Your BATNA, and how to improve it.
6. Your reservation point (the worst deal you'll accept).
7. Your communication and relationship strategy.
For high-stakes negotiations, run the same checklist for the other side and stress-test your assumptions. Mock-negotiate with a colleague playing the counterparty.
The negotiation literature's neglected topic: implementation. Most agreements that fail in practice fail not at the handshake but in the months after, when the parties discover the deal wasn't quite what they thought.
Specific risks: ambiguous language that meant different things to each side; absent enforcement mechanisms; missing dispute-resolution procedures; underspecified obligations on either side; changed circumstances that weren't anticipated.
The mitigation: write the deal explicitly, including the unhappy paths; specify how disputes will be resolved before they arise; build in periodic review meetings on long-running agreements; document what each party's representative is committing to vs what their organisation is committing to.
The contract literature (Williamson on transaction-cost economics) and the relational-contracting tradition both deserve more attention than they get in the negotiation curriculum.
↑ Chris Voss · Never Split the Difference · TEDx
Watch · Harvard Negotiators · Getting to Yes summary
Watch · Fisher & Ury · principled negotiation core message
Negotiation is unusual among professional skills in that most people get little structured practice. Three actual practice mechanisms:
Negotiation classes. Harvard Program on Negotiation, Wharton's Negotiations course, Columbia's negotiation certificate, INSEAD's negotiation programmes, online: Coursera's Wharton course, edX's Harvard offerings. The high-quality classes spend most time on simulated negotiations followed by debrief. Most learning comes from the debrief.
Mock negotiations with feedback. Find a partner, run a structured exercise, debrief. The Harvard PON's website sells dozens of vetted exercises with role descriptions and post-negotiation analysis.
Real-stakes practice with reflection. Every salary negotiation, contract revision, vendor discussion, household disagreement is practice. The discipline is post-hoc reflection: what went well, what didn't, what would I do differently.
Across forty-five years of empirical research, a few practices have robust support.
1. Prepare more than you think necessary. The single highest-leverage activity.
2. Improve your BATNA. Real alternatives change real outcomes.
3. Make the first credible aggressive anchor, when you have decent market information.
4. Listen actively and demonstrate understanding before advocating your position.
5. Negotiate multiple issues simultaneously — package proposals.
6. Use objective criteria to anchor proposals (market rates, precedent, third-party standards).
7. Build rapport early, especially in cross-cultural and remote contexts.
The single largest gap between what works and what most people do: the time spent on preparation. Strong negotiators prepare 2–3× the time they spend at the table.
Negotiation — Volume VII, Deck 11 of The Deck Catalog. Set in IBM Plex Sans and Tiempos Text. Off-white #fafaf6; emerald and gold accents.
Twenty-eight leaves on the conversation by which two parties make a joint decision. The science is settled enough to act on.
↑ Vol. VII · Bus. · Deck 11