Start hereYou are about to ask for a raise, or haggle over a car. Should you be tough, or nice? The studies also point to quieter moves, such as knowing your alternative, discussing issues together and pausing before you answer.
In this guide, a negotiation is any talk where two sides try to agree on terms, from a salary to a used car. Below are 15 tactics in three groups. Each card says what the tactic is, what the studies found, when it tends to help and what to watch for.
Some well-known tactics already have a page of their own. Mirroring, labelling, calibrated questions, first offers and anchors are in our Never Split the Difference summary.
Our reading: most studies below used role-plays, scenarios or simulations, so a real deal may play out differently. Nothing here promises a result.
Before you talk
Four tactics are about preparing: knowing your options, your limits and your aim before anyone makes an offer.
1. Know your best alternative (BATNA)
Work out what you will do if this deal falls through.
Your BATNA, short for “best alternative to a negotiated agreement”, is the course of action you would take if you do not reach agreement. The Program on Negotiation at Harvard Law School calls it the true measure by which to judge any proposed agreement.
Lab studies point the same way. In a 1994 study, having an alternative raised a negotiator’s own outcome and the joint outcome. The more attractive the alternative, the greater those benefits. A 1995 follow-up found the person with the alternative gained only a little more for themselves, and only when both sides knew about it.
For example, a written job offer from another company is a BATNA when you talk about pay with your current employer. Our reading: a good alternative makes it easier to say no to a poor deal.
How sure: Moderate lab studies support it, though the personal gain was small in the 1995 study.
2. Set a walk-away point
Decide before you start the worst deal you will accept, and stop there.
Working out your BATNA helps you set your reservation point, also called your walk-away point. The same program describes an “agreement trap”: the tendency to agree to a deal that is worse than your BATNA. The 1995 study found that having an attractive alternative changes the walk-away points of both negotiators.
For example, a seller who fixes a lowest price before an auction can turn down a lower bid calmly. The same article warns of the opposite mistake too: walking away from a deal that could have been mutually beneficial.
How sure: Limited working out a walk-away point is advice the Program on Negotiation gives. The studies we found tested alternatives rather than walk-away points on their own.
3. Set a specific, ambitious goal
Aim for a specific, challenging result rather than “as much as I can get”.
A 2002 meta-analysis of 22 research reports found that negotiators with specific, challenging goals consistently earned higher profits than those with less specific or less challenging goals, or none. Harder goals had a strong, positive influence on profits. The effect was larger when the other side had no strong goal, when performance was rewarded, when talks were not face to face and when the negotiator had experience.
For example, “a set raise plus one extra week of leave” is more specific than asking for “more”. The studies counted profit. Our suggestion: weigh a hard goal against the relationship if you will deal with this person again.
How sure: Moderate one meta-analysis, with a consistent result across the reports it pooled.
4. Estimate the zone of possible agreement (ZOPA)
Work out where both sides’ limits overlap, if they overlap at all.
If there is a set of deals both sides would prefer to no deal, a ZOPA, or zone of possible agreement, exists. In the program’s salary example, a candidate who will accept no less than $70,000 and an employer who will pay no more than $65,000 have no ZOPA. The same article says issues other than price, such as vacation time and work assignments, can and should be included in the ZOPA too.
You can rarely see the other side’s limit, so this is an estimate. A 2011 meta-analysis found that negotiators who bargained hard did best when they knew the bargaining zone.
For example, if pay alone cannot meet, other terms such as extra leave or an earlier pay review might. Our reading: an honest estimate shows early when a deal on these terms cannot happen.
How sure: Limited analysing the ZOPA is part of the careful preparation the Program on Negotiation advises. The meta-analysis tested it only as one factor among several.
Making the deal bigger
Six tactics look for deals that are better for both sides, rather than only splitting what is on the table.
5. Look for the interests behind positions
Ask why the other side wants what it wants, and say why you want what you want.
Positions are surface-level statements, while interests are the needs, motivations and concerns behind them. Focusing on interests, not positions, is one of the main ideas of Getting to Yes by Roger Fisher, Bill Ury and Bruce Patton, first published in 1981.
In a 1991 study, both giving and asking for information led to more mutually beneficial agreements. Joint results improved even when only one person in a pair shared or asked. Those who shared information did not lose out on their own profit.
In the program’s example, two siblings arguing over where to hold a party chose an affordable restaurant once they talked about time and cost. Our suggestion: share what matters to you, but keep your walk-away point to yourself.
How sure: Moderate lab studies support asking and sharing, and focusing on interests instead of positions is advice the Program on Negotiation gives.
6. Put the issues on the table together
Discuss all the issues at once instead of settling them one by one.
In a 1993 study of four-person groups, considering the issues at the same time improved the quality of the outcomes. It also made agreement more likely. Groups that took the issues together shared more information and understood each other’s priorities better.
For example, when buying a car, talk about the price, the delivery date and the trade-in together. Our reading: agreeing the price first may leave less room to trade later.
How sure: Moderate two studies from one research team, with groups of four.
7. Trade across issues
Give way on what matters less to you in return for what matters more.
Two 1990 studies found that most negotiators start out expecting the other side’s interests to be completely opposed to their own. A substantial number failed to notice when both sides wanted exactly the same thing, and settled for worse agreements. Most of what people learned about the other side came in the first few minutes of talking. In the 1990 studies, negotiators who learned the other side’s interests early earned more than those who learned them later.
For example, a tenant who cares most about a long lease and a landlord who cares most about a quick start can each give on the other’s priority.
How sure: Limited two lab studies from 1990, in line with the information studies in card 5. They looked at what negotiators noticed and learned, not at trading itself.
8. Make several equal offers at once
Offer two or three packages you would be equally happy with, and let the other side choose.
A 2019 paper reported six experiments comparing several equal offers made at once, called MESOs, with a single offer. The person making them got better outcomes, because the other side saw the offers as a more sincere attempt to agree. Both sides together also did better. The offerer gained a more cooperative reputation, and the talks felt more cooperative.
For example, a freelancer might offer a higher fee with a later deadline, or a lower fee with a smaller job. Our suggestion: check that each package really is equal for you before you offer it.
How sure: Moderate six experiments, all from one research team.
9. Let the deal depend on what happens
When you disagree about the future, write the deal so it adjusts to what happens.
In a contingent contract, the terms are not final until the uncertain event takes place. A familiar case is a chief executive’s pay tied to the company’s share price. In a 1999 article, Bazerman and Gillespie argue that such contracts let a difference of opinion become the basis of an agreement, not an obstacle to it. They also say contingent contracts are not right for every situation.
For example, a writer sure a book will sell well might accept a smaller fee plus a share of sales. Our suggestion: write down how the outcome will be measured, so nobody argues about it later.
How sure: Limited this rests on the authors’ argument in a 1999 Harvard Business Review article, not on a controlled study we could find.
10. Agree on a fair standard
Settle what counts as fair by an outside measure both sides accept.
Principled negotiation recommends basing decisions on objective criteria: fair, independent standards both sides agree to use. Examples include market value, industry benchmarks, expert opinion and legal standards. The key is agreeing in advance which standards will apply. Getting to Yes also advises choosing standards that do not favour either side. Getting to Yes presents independent standards of fairness as a way to avoid a bitter contest of will.
For example, two neighbours sharing the cost of a new fence might agree to use the middle of three quotes.
How sure: Limited agreeing on objective criteria is advice the Program on Negotiation gives. We did not find a controlled study that tested it on its own.
Moves at the table
Five tactics are about how and when you speak once talks begin, including one worth spotting when it is used on you.
11. Talk about your own side in gains, not losses
Describe what you stand to gain, rather than what you stand to lose.
In a 1994 study, students negotiated with a computer program whose messages described its own outcomes as gains or as losses. When it talked about its gains, the students made lower demands and bigger concessions than when it talked about its losses. The program wrote, for example, that an offer implied “only small gains for me”, or in the loss version “only big losses for me”. The effect was stronger when the students thought in gains themselves. Students also made smaller concessions when the opponent was framed as facing losses.
For example, “This price leaves me too little profit” invites a different reply from “this price costs me too much”. Our page on cognitive biases looks at framing more broadly.
How sure: Limited one lab study from 1994, with students negotiating against a computer program.
12. Use silence
Pause for a few seconds before you answer.
A 2022 set of studies looked at pauses of at least 3 seconds, which the researchers call extended silence. Asking one or both sides to use extended silence led to more value being created in the deal. It worked better for creating value than asking negotiators to problem-solve. The researchers found no evidence of a link between silence and the share of value each side claimed.
In plain words, silence seemed to grow the deal rather than win a bigger slice of it. When status differences stood out, silence helped when the higher-status person used it, but not when the lower-status person did.
For example, after an offer, count to three slowly before you reply.
How sure: Moderate several studies, all from one research team.
13. Plan your concessions
Decide in advance what you could give, and when.
A 2011 meta-analysis pooled 34 studies of talks over how to split a fixed amount. Hardline tactics led to better economic results, while softline tactics led to better results in how people felt. Softline negotiators did best when they matched the other side’s concessions accurately.
A 2004 scenario study varied when the other side made its concessions. The timing changed how much people valued the item and how satisfied they were with the deal and the partner. A second study repeated some of these findings but not all.
When the other side asks you to concede before it makes a counteroffer, the Harvard program’s advice is “Don’t bid against yourself”. Our Influence summary covers asking big and then smaller.
For example, if you can offer free delivery, hold it back until you need something in return.
How sure: Limited a 34-study meta-analysis supports the trade-off between hard and soft bargaining, and two studies looked at concession timing. We found no study that tested planning concessions in advance.
14. Use deadlines, and think about revealing yours
A deadline can move talks along, and telling the other side about it may help you.
In 2004 studies, people predicted that a final deadline would hurt their results. Moderate deadlines helped negotiators who were eager to finish because waiting cost them.
Revealing a final deadline can lead to better outcomes for the person with the deadline, because it sped up the other side’s concessions. Both inexperienced and experienced negotiators predicted the opposite.
For example, “I need to decide by Friday” can move a slow talk along. Our page on psychological triggers looks at deadlines in shopping.
How sure: Limited lab studies, both papers by one researcher. Our reading: a deadline the other side doubts may not work the same way.
15. Spot “good cop, bad cop”
One person on the other side plays tough while a partner plays reasonable.
The Harvard program lists it among hard-bargaining tactics and reminds readers that the two are working together. In a 1993 study, buyers negotiated twice with computer-simulated sellers. A tough seller in the first round followed by a friendly one in the second was the most likely way to narrow the gap and reach agreement. It also produced, in the researchers’ words, “the greatest perceived cooperativeness”.
Two 2000 studies of a hypothetical three-person work group mostly supported a model of how negotiators decide and are persuaded, which the authors used to study the tactic. One twist depended on whether a person was first inclined to accept or reject.
For example, a salesperson “checks with the manager”, returns with a firm no, then offers a small extra. Our suggestion: treat the two as one side and answer the offer, not the mood.
How sure: Limited three older studies across two papers, one with computer-simulated sellers.
Where to learn more
- First offers, anchors, mirroring, labelling and calibrated questions: our Never Split the Difference summary.
- Reciprocity and asking big then smaller: our Influence summary.
- Why sensible choices can leave both sides worse off, and the ultimatum game: our page on game theory in plain words.
- Pressure tactics in shopping, such as countdowns: our page on psychological triggers in marketing.
- Framing and loss aversion: our page on cognitive biases.
- The Program on Negotiation at Harvard Law School publishes free articles on BATNA, ZOPA and principled negotiation at pon.harvard.edu.
Quick answers
What is a BATNA in negotiation?
Your best alternative to a negotiated agreement, meaning what you would do if you do not reach a deal. The Program on Negotiation at Harvard Law School calls it the true measure by which to judge any proposed agreement.
Does staying silent help in a negotiation?
In a 2022 set of studies, pauses of at least 3 seconds led to more value being created in the deal, with no evidence of a link to the share each side claimed. The studies come from one research team.
Should you tell the other side your deadline?
In 2004 lab studies, revealing a final deadline sped up the other side's concessions and could lead to better outcomes for the person with the deadline, although negotiators predicted the opposite.
Does good cop, bad cop work?
In a 1993 study with computer-simulated sellers, a tough seller first and a friendly one second was the most likely way to reach agreement. The evidence is old and limited, and the Program on Negotiation advises remembering that the two are working together.
Use this
Our suggestion: three steps for your next negotiation, drawn from the cards above.
What will I do if we do not agree?
Quick quiz
Nothing is scored or saved. Recalling it once now is what makes it last.
1. What is a BATNA?
Answer: What you will do if you do not reach agreement.Read it again ↑
2. In a 2022 set of studies, how long did a pause need to last to count as extended silence?
Answer: At least 3 seconds.Read it again ↑
3. In 2004 studies, what happened when negotiators revealed their final deadline?
Answer: The other side made concessions faster.Read it again ↑
How we checked this · Last checked November 2026
- We read the abstracts of the studies named on each card, and articles from the Program on Negotiation at Harvard Law School for the planning ideas.
- Getting to Yes is still in copyright, so we used the Program on Negotiation's own descriptions of it rather than quoting the book.
- First offers, anchors, mirroring, labelling and calibrated questions are covered in our Never Split the Difference summary, so they are not repeated here.
- Most of the studies used role-plays, scenarios or simulations, so each card says how sure the finding is.