Settlement Negotiation Tactics: Principled Negotiation, Anchoring, and Handling Hardball in Litigation

Summary. This article explains how litigation settlements actually get negotiated and how lawyers can do it better, combining the interest-based method of Getting to Yes with the distributive realities of the dance of numbers. It covers BATNA and the zone of possible agreement in a lawsuit, anchoring and first offers, reactive devaluation and other cognitive barriers, and decision-tree risk analysis. It then turns to mediator techniques such as brackets and the mediator's proposal, the common hardball tactics and how to answer each one, and the ethical limits on what a negotiator may say under Model Rule 4.1 and ABA Formal Opinion 06-439. Separate sections address confidentiality under Federal Rule of Evidence 408 and the ADR Act, offers of judgment under Rule 68 and fee-shifting leverage, settlement dynamics in patent, trademark, and copyright cases, and enforcement after Kokkonen. An extended worked example follows a trademark dispute from opening demand to signed term sheet, and the article closes with common mistakes, an FAQ, and practical takeaways.

It is 11:40 on a Thursday morning in a mediator's conference suite, and the two sides of a software licensing dispute have not been in the same room since 9:15. The plaintiff, a logistics company called Corvane Freight, opened at $4.8 million. The defendant, a mid-sized developer called Lattice Works, answered at $75,000 and a mutual release. The mediator has carried each number down the hall and watched it land like an insult. Corvane's general counsel now says the defendant "is not serious" and wants to leave. Lattice's CEO says the same thing about Corvane, in nearly the same words.

Both of them are wrong, and both of them are behaving exactly as the research predicts. Each side has anchored on its own number, discounted the other side's number because of who offered it, and read the gap as a statement about the other side's character rather than as the opening move of a ritual that both of their lawyers have performed dozens of times. By 6:30 that evening the case will settle for $1.15 million, a source-code escrow, and a two-year license extension that Corvane wanted more than it wanted the money. Whether it gets there depends less on the merits, which have not changed since breakfast, than on how the people in the two rooms manage the next seven hours.

This article is about those seven hours, and about the weeks of preparation that make them productive. It is not about drafting the release; MC Law covers that in Drafting a Settlement Agreement and Release: A Practical Guide and Settlement Agreements, Releases, and Rule 68: Ending a Case on Terms That Hold. It is about the negotiation itself: the method, the psychology, the numbers, the tactics, and the rules that limit what a lawyer may say while doing it.

The short answer

Most cases settle, so the negotiation is the main event. Trial is the alternative that gives settlement its shape. Lawyers bargain "in the shadow of the law," and the better they can describe that shadow, the better they bargain.
Principled negotiation works in litigation, with adjustments. Separating people from the problem, focusing on interests, inventing options, and insisting on objective criteria all apply. But lawsuits also involve a fixed sum of money that one side pays and the other receives, and pretending otherwise gets a lawyer outmaneuvered.
Your power is your alternative. In a lawsuit, the best alternative to a negotiated agreement is not "trial." It is a probability-weighted range of outcomes, net of fees, time, and collection risk. Lawyers who cannot state that range in numbers cannot tell whether an offer is good.
First numbers matter more than they should. Anchoring is one of the most robust findings in decision research, and it works on judges, juries, and experienced negotiators. A credible, well-justified first offer tends to pull the outcome toward it. An incredible one can end the conversation.
People devalue proposals because of who makes them. Reactive devaluation means a good offer from the other side looks worse than the same offer from a mediator. That is one reason mediators, brackets, and mediator's proposals work.
Hardball has answers. Exploding offers, last-minute demands, limited-authority games, and threats each have a standard response. The general rule, from William Ury's Getting Past No, is to step back, name the tactic, and redirect to the problem.
Puffery is allowed; lies about material facts are not. Under Model Rule 4.1 and ABA Formal Opinion 06-439, statements about value and a client's willingness to settle are generally not "statements of fact," but misrepresenting insurance limits, evidence, or a client's death is misconduct and can unwind the deal.
Settlement talk is protected, but not completely. Fed. R. Evid. 408 limits use of offers and statements to prove liability or amount, but it is a rule of admissibility, not a discovery privilege, and it has exceptions.
Procedural tools change the math. A Fed. R. Civ. P. 68 offer of judgment, a fee-shifting statute, or an inter partes review petition can move the other side's alternative more than any argument can.
A deal is only as good as its enforcement. After Kokkonen v. Guardian Life Insurance Co. of America, 511 U.S. 375 (1994), a federal court does not automatically keep jurisdiction to enforce a settlement. Plan for enforcement at the table.

Why tactics matter: bargaining in the shadow of the law

Only a small fraction of federal civil cases reach trial; figures compiled from Administrative Office data have hovered around one percent for years. The rest are dismissed, decided on motions, abandoned, or settled, and settlement is the largest category. In 1979 Robert Mnookin and Lewis Kornhauser described divorcing couples as "bargaining in the shadow of the law," meaning that the legal rule a court would apply sets the endowments each side brings to the negotiation even when no court ever applies it. The phrase now describes all civil litigation. The motion to dismiss, the summary judgment ruling, the damages expert, and the jury instructions matter mostly for what they do to the settlement number.

Two consequences follow. First, litigation and negotiation are not separate phases. Every pleading, discovery request, and motion is also a negotiation move, because each changes one side's estimate of the trial outcome or the cost of getting there. A lawyer who files a strong summary judgment motion has made an offer, of a kind: "this is what your case looks like after the court rules." Second, the quality of a settlement depends on the quality of each side's prediction. George Priest and Benjamin Klein's selection hypothesis (1984) suggests that cases which go to trial tend to be the close ones, because lopsided cases settle once both sides see the likely result. When cases fail to settle, it is often because at least one side's prediction is wrong.

Empirical work suggests that predictions are wrong often. A study by Randall Kiser, Martin Asher, and Blakeley McShane, "Let's Not Make a Deal," published in the Journal of Empirical Legal Studies in 2008, compared the last settlement positions in California cases that went to trial with the eventual verdicts. Plaintiffs made the wrong call (rejecting an offer that beat the verdict) in about 61 percent of the cases studied, and defendants in about 24 percent. But the defendants' mistakes were far more expensive: the average plaintiff error cost about $43,000, while the average defendant error cost about $1.1 million. The data are California cases spanning four decades, but the lesson is durable. Both sides misjudge, and the side that pays the judgment tends to misjudge at larger scale.

Negotiation tactics matter because they are how lawyers correct, exploit, or are victimized by these misjudgments. For the front-end analysis of whether a case is worth bringing at all, see Evaluating and Assessing a Civil Case.

Principled negotiation, translated for lawsuits

Roger Fisher and William Ury's Getting to Yes (1981, later editions with Bruce Patton) proposed "principled negotiation" as an alternative to the haggling of positional bargaining. The method has four elements. Each translates into litigation, but each needs adjustment.

Separate the people from the problem

Fisher and Ury's advice is to be "hard on the problem, soft on the people." In a lawsuit this is harder than in a business deal, because the parties are usually angry, often feel wronged, and are paying lawyers to fight. The client who was fired, the founder whose company was sued for infringement, and the insurer's adjuster who thinks the claim is inflated all bring feelings that will block a rational deal if no one addresses them.

In practice, separating people from problems means three things. First, give the other side's principal a chance to be heard. A plaintiff who has never had the experience of telling the story to someone with power may need an opening session or a direct conversation with the defendant's decision maker before any number will feel acceptable. Second, keep the lawyers' relationship professional even when the clients' relationship is broken. Opposing counsel who trust each other's statements about logistics, deadlines, and authority settle cases faster. Third, attack positions with evidence, not adjectives. "Your damages expert used the wrong royalty base, and here is the Federal Circuit case that says so" moves a negotiation. "Your expert is a hired gun" does not.

Focus on interests, not positions

A position is a demand: "$4.8 million." An interest is the reason behind it: Corvane needs to show its board a recovery, needs its warehouse software to keep running, and fears that the developer will go out of business and leave it with unsupported code. Positions in litigation are almost always expressed in dollars, because damages are what the court can award. Interests are much wider.

The litigator's job is to ask, before and during the negotiation, what each side actually needs. Common interests hidden behind dollar positions include:

Timing. A defendant may be able to pay more over 24 months than today. A plaintiff may need money before a fiscal year closes or before a bank covenant is tested.
Precedent and reputation. A repeat-player defendant (an insurer, a franchisor, a platform) may pay a premium for confidentiality or a non-precedential structure. A plaintiff may care more about a public statement than a check.
Continuing relationships. Suppliers, licensees, franchisees, and co-founders often have to keep doing business after the case ends. A contract extension or revised pricing can be worth more than damages.
Tax and accounting treatment. The character of a payment can change its after-tax value to both sides. See 26 U.S.C. § 104 for the personal physical injury exclusion and 26 U.S.C. § 162, including subsection (q) on deductions for certain confidential sexual harassment settlements.
Insurance. A defendant's real interest may be staying within policy limits or getting its carrier to fund more of the deal.
Vindication. Some clients want an apology, a corrected record, or a change in practice. These cost the other side little and are often worth a great deal.

Invent options for mutual gain

Fisher and Ury recommend separating brainstorming from evaluation and looking for trades in which each side gives up something it values less than the other side does. In litigation these trades are the difference between a deal and an impasse. Payment plans, non-monetary relief, licenses, supply agreements, phase-out periods for an infringing product, mutual non-disparagement, neutral references, structured payments, and carve-outs from releases all expand the set of possible deals.

The limit is that a lawsuit always has a distributive core. Somebody pays and somebody receives, and the size of the check is a zero-sum question. Robert Mnookin, Scott Peppet, and Andrew Tulumello call this "the tension between creating and distributing value" in Beyond Winning (2000). Creating value requires sharing information about interests and priorities. Distributing value rewards concealing information about reservation prices. Good negotiators do both, often in the same session, and they are deliberate about which information they share.

Insist on objective criteria

Fisher and Ury's fourth element is to base agreements on standards that are independent of either side's will: market value, precedent, expert opinion, or what a court would do. This is where litigation has a real advantage over other negotiations. There is an external standard, the law, and there are data about how courts apply it: jury verdict reports, published damages awards, royalty rates in comparable licenses, statutory damages ranges such as those in 17 U.S.C. § 504(c), and the fee-shifting standards discussed below.

Objective criteria do two jobs. They give the other side a reason to move that is not surrender ("we are not paying you more because you yelled; we are paying you more because the comparable license rate supports it"). And they give the client a defensible explanation for the deal to a board, a spouse, or an insurer. When a client says "I need to be able to explain this number," that is an interest in an objective criterion.

The critique, and why it matters

James J. White's 1984 review of Getting to Yes argued that the book underestimated how much of negotiation is distributive and how often one side gains by concealment and firmness. Litigators recognize the point. Being "soft on the people" does not mean disclosing a client's walk-away number, and "objective criteria" can be gamed by the side with the better expert. The workable synthesis is to use the interest-based method to find value and build credibility, while managing the dollar exchange with the discipline of a distributive bargainer. The rest of this article covers both halves.

BATNA, WATNA, and the zone of possible agreement

Fisher and Ury's most influential idea is the BATNA, the Best Alternative to a Negotiated Agreement. A party should accept a deal that beats its BATNA and reject one that does not. The concept forces a negotiator to measure an offer against something real rather than against a hoped-for number.

In litigation, the alternative to settlement is continued litigation, ending in a judgment that may or may not be collectible, years from now, after fees that may or may not be recovered. That alternative is not a single number. It is a distribution of outcomes. Lawyers therefore talk about three reference points:

Concept	Meaning in a lawsuit	Example (plaintiff)
BATNA	The realistic value of continuing, as a probability-weighted expected value net of costs	40% chance of a $1.5M verdict, net of $450K remaining fees: about $150K
WATNA (worst alternative)	The bad but plausible outcome	Summary judgment loss plus a fee award against the client
Reservation price	The walk-away point, derived from the BATNA plus non-monetary interests and risk tolerance	"We will not take less than $300K"

Howard Raiffa (The Art and Science of Negotiation, 1982) and David Lax and James Sebenius (The Manager as Negotiator, 1986) developed the idea of the zone of possible agreement, or ZOPA: the range between the plaintiff's reservation price and the defendant's. If the plaintiff will take $300,000 and the defendant will pay up to $700,000, any number between them leaves both better off than litigating. The negotiation determines where in that range the deal lands. If the plaintiff's floor is above the defendant's ceiling, there is no ZOPA, and no amount of tactical skill will produce a deal until something (a ruling, a cost, new information) moves one of the reservation prices.

Three features of litigation BATNAs deserve attention.

They change over time. Each ruling, deposition, and expert report shifts the probabilities, and each month of litigation spends money that is not coming back. A defendant's BATNA often looks best before discovery and worst on the eve of trial; a plaintiff's may be the reverse. Timing the negotiation to the moment when the other side's BATNA is weakest is a large part of the art.

They are asymmetric. A plaintiff on a contingent fee and a defendant paying hourly face very different cost curves. An insured defendant may have a carrier paying defense costs, which flattens its cost curve but adds a third party with its own interests. A plaintiff in a case with a fee-shifting statute may expect fees as part of the recovery, which raises its BATNA.

They can be changed. Fisher and Ury urge negotiators to improve their own BATNA and, where legitimate, to worsen the other side's perception of its own. In litigation the tools are procedural: a strong dispositive motion, a Fed. R. Civ. P. 68 offer, a petition for inter partes review under 35 U.S.C. § 311, a counterclaim, a motion for fees. Each makes the alternative to settlement look different.

For the worksheet versions of these calculations, see Settlement Negotiation Toolkit: Case Valuation Worksheets, Concession Plans, and Offer Letters.

Positional bargaining and the dance of numbers

Whatever principled negotiation recommends, most litigation settlements involve an exchange of numbers. The plaintiff demands, the defendant offers, and each side moves in decreasing increments until the numbers meet or the parties stop. Lawyers call it "the dance." It looks irrational from the outside. It is a signaling system.

Opening numbers set the frame. The first demand and first offer establish the range within which everyone expects the deal to land. Experienced mediators often predict the settlement as something near the midpoint of the first credible numbers, adjusted for who moves faster. That is why parties open aggressively: a demand of $4.8 million and an offer of $75,000 frame a midpoint near $2.4 million, which favors whoever is more credible about the outer number.

Concession size communicates. A move from $4.8 million to $4.2 million says "we have room." A move from $1.4 million to $1.35 million says "we are near the end." Shrinking concessions are the normal way to signal an approaching reservation price. Large late concessions signal desperation or an earlier bluff.

Reciprocity governs. Negotiators expect concessions to be matched roughly in size and timing. A side that moves $500,000 and receives a $10,000 response will often stop, or will retract. Mediators spend much of their time translating moves: "they went down $500,000; I need something meaningful from you or they will walk."

Patterns can be gamed, and read. Some negotiators deliberately make a few large concessions early to create momentum, then slow sharply. Others start with small moves to test the other side's patience. Tracking the sequence of numbers on a whiteboard, with the time each move was made, helps both the lawyer and the client see the pattern instead of reacting to the last move.

The dance is not a substitute for analysis. It is the surface on which analysis gets expressed. The lawyer who knows the client's BATNA can make each move with a purpose: to signal room, to signal limits, or to induce a move. The lawyer who does not will make concessions in response to fatigue, pressure, or the clock.

For a step-by-step plan for building a concession sequence before the session, see Planning and Running a Litigation Settlement Negotiation: A Practical Guide.

Anchoring and the first offer

In 1974, Amos Tversky and Daniel Kahneman published a study in Science in which participants spun a wheel rigged to land on either 10 or 65 and were then asked to estimate the percentage of African countries in the United Nations. People who saw the higher number gave higher estimates, even though the number was obviously random. The effect, anchoring, has been replicated many times in many settings, including legal ones.

Chris Guthrie, Jeffrey Rachlinski, and Andrew Wistrich tested sitting federal magistrate judges in a 2001 Cornell Law Review study, "Inside the Judicial Mind." In one experiment, judges evaluated a personal injury damages question. Some judges first saw a meritless motion to dismiss arguing that the claim did not meet the $75,000 amount-in-controversy requirement of 28 U.S.C. § 1332. Those judges, on average, awarded substantially less than judges who did not see the motion. A low number in a procedural motion had pulled down a damages estimate made by professionals trained to ignore irrelevant information.

The implications for settlement are direct.

Going first can help. Research by Adam Galinsky and Thomas Mussweiler (2001) and others found that the party who makes the first offer often gets a better outcome, because the first number anchors the counterpart's sense of the range. The conventional advice to "never make the first offer" is weaker than its popularity suggests. In litigation, the plaintiff usually goes first by convention, which gives the plaintiff the first anchor.

The anchor must be credible. An offer that is so extreme that the other side cannot justify a response may end the negotiation or prompt an equally extreme counter, which cancels the effect. The best anchors are aggressive but supported: "Here is the damages model, here are three comparable verdicts, and here is why our number sits where it does." A number with reasons is harder to dismiss than a number alone.

Anchors can be resisted. The best defense against an anchor is to have done the valuation before seeing it. Lawyers who walk into mediation with their own decision tree, and who have shown it to the client, are much less susceptible to the other side's first number. A second defense is to refuse to treat an extreme anchor as an opening: "That number is not connected to this case, and we are not going to negotiate off it. When you have a number based on the evidence, we will respond." A third is to re-anchor with a well-justified counter rather than splitting the difference from the other side's opening.

Pre-mediation briefs and demand letters are anchors too. The number in a demand letter, a mediation statement, or a damages expert report frames the negotiation well before the session. Drafting them is a negotiation decision, not just a litigation one. MC Law's Drafting a Demand Letter: A Practical Checklist covers the mechanics.

Reactive devaluation and other barriers to settlement

Rational models of settlement predict that cases settle whenever the ZOPA is positive. Many cases with positive ZOPAs fail to settle anyway. Psychologists and legal scholars have identified several reasons.

Reactive devaluation. Lee Ross and Constance Stillinger described a pattern in which people value a proposal less once they learn the other side has offered it. In one widely discussed study involving a university's response to student demands on divestment, students rated a compromise less favorably when they believed the university had adopted it than when it was described as a hypothetical. In litigation, an offer from opposing counsel is suspect because it comes from opposing counsel: "If they are willing to pay that, the case must be worth more." Mediators counter this by presenting proposals as their own, by using hypotheticals ("if I could get them to X, would you consider Y?"), and through the mediator's proposal, discussed below.

Self-serving bias. Linda Babcock and George Loewenstein ran experiments in which participants read identical case materials but were assigned to play the plaintiff or the defendant. Participants' predictions of what a judge would award shifted in the direction of their assigned role, and the size of the shift predicted whether pairs reached agreement. When roles were assigned only after participants read the materials, the bias shrank and settlement rates improved. Lawyers absorb the same bias from their clients. A useful corrective is to have a colleague who has not worked on the case argue the other side's position before mediation.

Loss aversion and framing. Kahneman and Tversky's prospect theory holds that people feel losses more strongly than equivalent gains and become risk-seeking when choosing between losses. Jeffrey Rachlinski's work on litigation framing ("Gains, Losses, and the Psychology of Litigation," 1996) applied this to settlement: plaintiffs, who see settlement as a sure gain, tend to be risk-averse and accept; defendants, who see settlement as a sure loss, tend to gamble on trial. That pattern fits the Kiser study's finding that defendant errors are expensive. A defense lawyer can counter it by reframing: "Paying $600,000 today avoids a 35 percent chance of a $2.2 million judgment plus fees."

Sunk costs and endowment. Clients who have spent $400,000 on fees want a settlement that "justifies" the spend. That is irrelevant to the decision, since the money is gone whatever happens. Russell Korobkin and Chris Guthrie's experimental work ("Psychological Barriers to Litigation Settlement," 1994) found that framing and reference points can change whether identical offers are accepted. Lawyers should reset the client's reference point explicitly: the question is whether this offer beats the future, not whether it repays the past.

Agency problems. The lawyer and client may have different incentives. A contingent-fee lawyer bears the cost of trial without sharing equally in the upside; an hourly lawyer may earn more if the case continues; an in-house lawyer may be judged against a reserve number set months ago. Model Rule 1.2(a) gives the client, not the lawyer, the authority to decide whether to settle, and Model Rule 1.4 requires the lawyer to keep the client informed, including about offers. Disclosing the incentive structure to the client is both an ethical requirement and a practical one.

Risk analysis and decision trees

A decision tree is a diagram of the uncertain events between now and judgment, with a probability on each branch and a value on each outcome. Multiplying through gives an expected value. Lawyers have used them since at least the 1980s; Marjorie Corman Aaron's work on risk analysis for mediators and the Harvard Program on Negotiation's teaching materials made them standard in mediation practice.

A simple tree for a plaintiff in a breach of contract case might look like this.

Node	Branch	Probability	Cumulative probability	Outcome
Summary judgment	Defendant wins	25%	25%	$0
	Case survives	75%
Liability at trial	Plaintiff loses	40% of 75%	30%	$0
	Plaintiff wins	60% of 75%	45%	to damages
Damages (if win)	Low ($400K)	30% of 45%	13.5%	$400,000
	Middle ($900K)	50% of 45%	22.5%	$900,000
	High ($1.6M)	20% of 45%	9%	$1,600,000

The probability-weighted gross recovery is (0.135 × $400,000) + (0.225 × $900,000) + (0.09 × $1,600,000) = $54,000 + $202,500 + $144,000 = $400,500. If the plaintiff will spend another $250,000 to get through trial, faces an appeal risk, and will wait two years for payment, the present net expected value may be well under $150,000. The client's reservation price can then be set with reference to that number, adjusted for risk tolerance and non-monetary interests.

Trees are useful because they force specificity. Saying "we have a strong case" means nothing. Saying "we think we survive summary judgment three times out of four" is a statement someone can argue with, and arguing with it is how lawyers find out what they actually believe.

Trees are dangerous because they look precise. The probabilities are judgments, and small changes compound. Running the tree with high and low estimates for each branch, and showing the client the resulting range, is more honest than presenting a single number. The tree also omits things that matter, including the chance of a fee award against the client, collection risk, publicity, and the client's tolerance for continued disruption.

Trees are persuasive when shared carefully. Some mediators ask each side to build its tree and then compare them privately. The branches where the two sides differ most are where the negotiation should focus. Sharing a tree with the other side is a strategic choice; it reveals reasoning, which can build credibility, but it can also reveal the reservation price.

For collection risk, see Collecting a Judgment: Discovery in Aid of Execution, Liens, Levies, and Garnishment. A plaintiff who wins a judgment against an insolvent defendant has won a piece of paper, and a defendant in financial distress may file for bankruptcy, which triggers the automatic stay under 11 U.S.C. § 362. Even a settlement payment received shortly before a bankruptcy filing can be clawed back as a preference under 11 U.S.C. § 547, which reaches certain transfers made within 90 days before the petition (one year for insiders). Plaintiffs settling with a shaky defendant should weigh security, guarantees, or a stipulated judgment against the bare promise to pay.

What mediators do, and how to use it

Most significant civil cases now settle at or after mediation. Federal law encourages it: 28 U.S.C. § 651 directs each district court to authorize alternative dispute resolution by local rule, and 28 U.S.C. § 652 requires litigants in civil cases to consider it. Fed. R. Civ. P. 16 lets the court take up settlement and "special procedures to assist in resolving the dispute" at pretrial conferences, require that a party or its representative be present or reasonably available to consider settlement, and sanction a party that does not participate in good faith. In G. Heileman Brewing Co. v. Joseph Oat Corp., 871 F.2d 648 (7th Cir. 1989) (en banc), the Seventh Circuit upheld a district court's power to order a represented corporate party to send a representative with authority to settle, while making clear that the court could not force the party to settle. The First Circuit in In re Atlantic Pipe Corp., 304 F.3d 135 (1st Cir. 2002) recognized an inherent power to order mediation in appropriate cases, subject to procedural safeguards such as limits on duration and cost.

Mediation helps negotiation in ways that go beyond carrying messages.

Caucus and controlled information. In private sessions, each side can tell the mediator things it will not tell the other side: its real concerns, its weaknesses, its flexibility. The mediator uses that information to find a ZOPA without disclosing either side's reservation price. Lawyers should decide in advance what the mediator may share and say so explicitly ("you can tell them we have authority to move, but not how far").

Reality testing. A mediator can ask a client the hard questions that the client's own lawyer cannot ask without seeming disloyal. "What happens if the judge excludes your expert?" Lawyers who want their clients reality-tested should brief the mediator on the weak points in advance.

Brackets. When numbers stall, one side proposes a range: "We will move to between $1.6 million and $1.2 million if you will move to between $400,000 and $700,000." A bracket communicates the midpoint each side is aiming for (here, $1.4 million and $550,000) without committing to a specific number. Counter-brackets narrow the range. Brackets work because they let parties make large conditional moves without appearing to cave. A lawyer receiving a bracket should calculate its midpoint immediately, because the midpoint is the real message.

The mediator's proposal. When the parties are close but stuck, the mediator proposes a specific number and terms and asks each side to accept or reject it privately, usually by a deadline. If both accept, there is a deal. If either rejects, the mediator tells neither side what the other said. This structure eliminates reactive devaluation (the number comes from the neutral), removes the fear of bidding against oneself, and lets a party say yes without the other side ever learning it would have done so if the other had refused. The tactical question for each side is whether a mediator's proposal will land closer to its own number. Mediators typically pitch it where they believe both sides will accept, which is not necessarily the midpoint.

The "final" number and the "last, best." Mediators often ask for each side's "best number" late in the day. Whether to give the real reservation price is a judgment call. Experienced lawyers give a number that is close to the real one and label it accurately: "This is our best number today. I cannot promise it will be available next month."

For the logistics of selecting a mediator and preparing the mediation statement, see Mediation and Settlement: Preparing, Negotiating, and Documenting the Deal and Arbitration, Mediation, and Choosing a Dispute-Resolution Forum: A Practical Guide.

Hardball tactics and how to answer them

Some negotiators rely on pressure rather than persuasion. Fisher and Ury call the response "negotiation jujitsu": when the other side pushes, do not push back; step aside and redirect the energy toward the problem. Ury's later book, Getting Past No (1991), gives a five-step sequence for difficult counterparts: go to the balcony (pause and get perspective rather than reacting), step to their side (listen and acknowledge), reframe (turn attacks into questions about the problem), build a golden bridge (make it easy for them to say yes), and use power to educate (show them their alternative, rather than threatening). Chris Voss, a former FBI hostage negotiator, adds the technique he calls tactical empathy in Never Split the Difference (2016): labeling the other side's emotions ("it sounds like you feel this case has been dragged out unfairly") and asking calibrated open questions ("how am I supposed to take that number back to my client?").

The common litigation hardball tactics, and the standard answers, are these.

Tactic	What it looks like	Response
Extreme anchor	A demand or offer far outside any supportable range	Do not counter from it. Ask for the basis, state that you will respond to a number tied to the evidence, and re-anchor with your own justified figure.
Exploding offer	"This offer expires at 5 p.m. today"	Test whether the deadline is real. Ask what changes at 5:01. If the offer is acceptable, accept; if not, let it expire and say you remain open to talking. Real deadlines are tied to events (a ruling, a filing, a board meeting).
Limited authority	"I would love to, but my client has capped me at $200,000"	Confirm early who has authority and insist that the decision maker attend or be available, as Fed. R. Civ. P. 16 and many local mediation rules allow. Make your own concessions conditional on their authority.
Good cop, bad cop	The lawyer is reasonable; the client (or the "committee") is unmovable	Name it lightly and address the actual decision maker. Ask the reasonable one what the unreasonable one needs to hear.
Nibbling	After agreement on money, new demands appear: confidentiality, a non-disparagement clause, a different payment date	Put all material terms on the table early ("before we talk numbers, what non-monetary terms do you need?"). Treat a late nibble as reopening the price.
Bidding against yourself	"That is not enough; you need to come up" with no counter	Refuse politely: "We moved last. We will respond to a number."
Personal attacks and theatrics	Raised voices, walkouts, insults to the client	Go to the balcony. Take a break. Let the mediator manage the other room. Do not let the client decide anything while angry.
Threats of collateral consequences	Threats of publicity, regulatory complaints, criminal referral, or bar grievances	Assess legitimacy, then address the ethics. Threats tied to a demand for money can cross legal lines (see below).
Surprise evidence	A new document, witness, or report appears at mediation	Ask for time to review. Do not adjust the number on the spot unless the new information truly changes the analysis.
Take it or leave it	"This is our final offer"	Ask whether "final" means final today or final forever. Most final offers move once the other side has demonstrated that the alternative is real.

Threats deserve particular care. Federal criminal law makes it a crime to demand or receive money "under a threat of informing, or as a consideration for not informing, against any violation of any law of the United States." See 18 U.S.C. § 873. Most states have analogous extortion statutes. The old Model Code of Professional Responsibility flatly prohibited threatening criminal charges solely to gain advantage in a civil matter (DR 7-105). The Model Rules dropped the express prohibition, and ABA Formal Opinion 92-363 concluded that a lawyer may refer to possible criminal charges in a civil negotiation only where the criminal matter is related to the civil claim, the lawyer believes both are well founded, and the lawyer does not suggest improper influence over the criminal process. Many states retain stricter rules. A lawyer tempted to say "pay up or we call the prosecutor" should not.

Threats of sanctions are another pressure tool. A lawyer who believes the other side's claims are frivolous can serve a Rule 11 motion, which under Fed. R. Civ. P. 11 must be served 21 days before filing to give the other side a chance to withdraw. That safe harbor is itself a negotiation window. Using a sanctions threat as a bargaining chip when the motion would not be well grounded is sanctionable in its own right. See Litigation Sanctions and Professional Responsibility Toolkit.

The ethics of what you say at the table

Negotiation involves concealment. A lawyer does not reveal the client's reservation price, and often states the client's position more firmly than the client holds it. Where is the line?

Model Rule 4.1 and puffery

ABA Model Rule 4.1(a) provides that in representing a client, a lawyer shall not knowingly "make a false statement of material fact or law to a third person." Rule 4.1(b) prohibits failing to disclose a material fact when disclosure is necessary to avoid assisting a client's crime or fraud, unless Rule 1.6 prohibits disclosure. Rule 8.4(c) separately prohibits conduct involving dishonesty, fraud, deceit, or misrepresentation.

Comment [2] to Rule 4.1 carves out the conventions of negotiation. Under generally accepted conventions, "certain types of statements ordinarily are not taken as statements of material fact," including "estimates of price or value placed on the subject of a transaction and a party's intentions as to an acceptable settlement of a claim." So a lawyer may say "my client will not take less than $1 million" when the client would take $800,000, and may say "this case is worth $3 million" as advocacy.

ABA Formal Opinion 06-439 (2006) applied the rule to negotiations, including caucused mediation. The opinion concluded that the same standards apply whether the lawyer is talking to opposing counsel or to a mediator in caucus: posturing about value and settlement intentions is generally permitted, while false statements of material fact are not. The opinion gives examples on each side of the line. Statements about the value of the client's claim, or the client's unwillingness to settle below a figure, are ordinarily puffery. Misrepresenting facts, such as the existence of witnesses, the content of documents, or a client's medical condition, is not. Statements to a court in a judicial settlement conference are subject to the higher duty of candor in Model Rule 3.3; an earlier opinion, ABA Formal Opinion 93-370, addressed judicial settlement conferences and is generally read to allow a lawyer to decline to disclose settlement authority to a judge but not to misrepresent it.

The general guidance is easy to state and harder to apply. Opinions about value, statements of negotiating intent, and silence about the reservation price are allowed. Affirmative false statements about facts the other side would rely on are not. "We have three witnesses who saw the contract signed," when there is one, is a lie. "My client is prepared to try this case," when the client is wavering, is usually posturing. Where a statement sits in between, the safest course is to restate it as opinion or to say nothing.

Misrepresentation cases

Courts have unwound settlements and disciplined lawyers for misstatements during negotiation.

In Fire Insurance Exchange v. Bell, 643 N.E.2d 310 (Ind. 1994), the insurer's counsel told the plaintiff's lawyer that the policy limit was $100,000, when it was $300,000. The plaintiff settled for the lower figure and then sued for fraud. The Indiana Supreme Court held that the plaintiff's lawyer was entitled to rely on opposing counsel's representation; the insurer could not defend by arguing that the plaintiff should have independently verified the limits. The case is a reminder that lying about insurance is lying about a material fact, and that "you should have checked" is not a defense.

In Kentucky Bar Ass'n v. Geisler, 938 S.W.2d 578 (Ky. 1997), a lawyer negotiated a settlement of a personal injury claim without disclosing that her client had died. The Kentucky Supreme Court publicly reprimanded her. ABA Formal Opinion 95-397 had reached the same conclusion: a lawyer must disclose the client's death to opposing counsel and the court, because the lawyer's authority ends with the client's death and continuing to negotiate implies a living client.

The classic case on silence is Spaulding v. Zimmerman, 263 Minn. 346, 116 N.W.2d 704 (1962). Defense counsel learned through their own examining physician that the minor plaintiff had a potentially fatal aortic aneurysm, likely caused by the accident, that the plaintiff's doctors had not found. They said nothing, and the case settled. Because a settlement for a minor required court approval, the Minnesota Supreme Court affirmed vacating it. The case is taught for the ethical question (whether silence was permissible) and for the procedural answer (court-approved settlements can be reopened when a material fact was concealed).

Client authority and communication

Model Rule 1.2(a) provides that "a lawyer shall abide by a client's decision whether to settle a matter." Model Rule 1.4(a) requires the lawyer to keep the client reasonably informed, and the comment states that a lawyer who receives an offer of settlement in a civil controversy must promptly inform the client unless the client has previously indicated that the proposal is acceptable or unacceptable or has authorized the lawyer to accept or reject it. These rules have tactical consequences. A lawyer cannot accept a deal the client has not authorized, and a settlement entered without authority may not bind the client, depending on state agency law. Lawyers should get written authority before mediation, define its limits, and confirm any expansion in writing (an email from the client during the session will do).

Two other rules shape multi-party settlements. Model Rule 1.8(g) prohibits a lawyer who represents two or more clients from making an aggregate settlement of their claims unless each client gives informed consent in a signed writing after disclosure of all the claims and the participation of each person. ABA Formal Opinion 06-438 explains what counts as an aggregate settlement. Model Rule 5.6(b) prohibits a lawyer from offering or making an agreement in which a restriction on the lawyer's right to practice is part of a settlement; a defendant cannot buy the plaintiff's lawyer out of future cases. ABA Formal Opinion 00-417 applies the rule to restrictions on a lawyer's future use of information learned in the case.

For a broader treatment, see Legal Ethics in Practice: Conflicts, Confidentiality, Fees, and the Duty of Technological Competence.

Confidentiality: what protects settlement talk

Parties negotiate more freely when they believe their offers and admissions cannot be used against them. The protection exists, but it has gaps.

Federal Rule of Evidence 408

Fed. R. Evid. 408 makes two categories of evidence inadmissible "to prove or disprove the validity or amount of a disputed claim or to impeach by a prior inconsistent statement or a contradiction": offers or acceptances of valuable consideration in compromising or attempting to compromise the claim, and conduct or statements made during compromise negotiations about the claim. The rule contains an exception for criminal cases where the negotiations related to a claim by a public office in the exercise of its regulatory, investigative, or enforcement authority. And subsection (b) allows the evidence "for another purpose," such as proving a witness's bias or prejudice, negating a contention of undue delay, or proving an effort to obstruct a criminal investigation or prosecution.

Several features matter tactically.

There must be a dispute. Rule 408 protects negotiations over a "disputed claim." Business conversations before any claim has been asserted may not qualify. Labeling a letter "Rule 408 settlement communication" does not create a dispute that does not exist, although it helps show intent.
Pre-existing facts are not immunized. A document that exists independently does not become inadmissible because it was handed over at mediation.
It is a rule of admissibility, not discovery. Courts disagree on whether settlement communications are protected from discovery. The Sixth Circuit in Goodyear Tire & Rubber Co. v. Chiles Power Supply, Inc., 332 F.3d 976 (6th Cir. 2003) recognized a settlement privilege under Fed. R. Evid. 501. The Federal Circuit declined to recognize a settlement negotiation privilege in In re MSTG, Inc., 675 F.3d 1337 (Fed. Cir. 2012), in a patent case where the negotiations bore on the reasonable royalty.
Settlements with others can be evidence. In patent cases, prior licenses, including some that resulted from litigation, can be relevant to damages. The Federal Circuit has cautioned that litigation-driven licenses may not reflect the economic value of the patent, see ResQNet.com, Inc. v. Lansa, Inc., 594 F.3d 860 (Fed. Cir. 2010) and LaserDynamics, Inc. v. Quanta Computer, Inc., 694 F.3d 51 (Fed. Cir. 2012), but a patentee who settles cheaply with an early defendant may see that number used against it later.

Mediation confidentiality

28 U.S.C. § 652(d) requires each district court, until a federal rule addresses the subject, to provide by local rule for the confidentiality of court-annexed ADR processes and to prohibit disclosure of confidential dispute resolution communications. The resulting local rules vary. State law also matters, especially when a mediation is private. The Uniform Mediation Act, promulgated in 2001 and adopted in roughly a dozen states and the District of Columbia, creates a privilege for mediation communications held by parties, mediators, and nonparty participants, subject to exceptions (such as signed agreements, threats of violence, and claims of professional misconduct). California has its own statutory scheme, which the California Supreme Court applied strictly in Cassel v. Superior Court, 51 Cal. 4th 113 (2011), holding that private attorney-client communications made in connection with a mediation were inadmissible even in the client's later malpractice suit against the lawyer. Mediation privilege is a patchwork, so check the law of the state whose privilege will apply before saying anything in mediation you could not afford to see in evidence.

Practical rules follow. Sign a mediation confidentiality agreement that specifies governing law. Do not assume that what is said in caucus will stay out of a later fee dispute or malpractice case. Do not use mediation as a free discovery session, and do not reveal anything there that would be catastrophic if repeated. For protective orders that govern documents exchanged in settlement talks, see Confidentiality Toolkit: Protective Orders, Designation Challenges, and Sealing Motions.

Offers of judgment and fee-shifting leverage

Some of the most effective negotiation moves are made with procedural tools that change what happens if the offer is refused.

Rule 68

Under Fed. R. Civ. P. 68, a party defending against a claim may, at least 14 days before trial, serve an offer to allow judgment on specified terms, with costs then accrued. The offeree has 14 days to accept in writing; if it does, the clerk enters judgment. An unaccepted offer is considered withdrawn and is inadmissible except in a proceeding to determine costs. The consequence is in Rule 68(d): if the judgment the offeree finally obtains "is not more favorable than the unaccepted offer, the offeree must pay the costs incurred after the offer was made."

Read alone, that sanction is modest. Taxable costs under 28 U.S.C. § 1920 (filing fees, transcripts, copying, witness fees) are usually small compared with attorney's fees. Rule 68 becomes powerful in three situations.

First, when the underlying statute defines attorney's fees as part of costs. In Marek v. Chesny, 473 U.S. 1 (1985), the Supreme Court held that "costs" in Rule 68 includes all costs properly awardable under the relevant substantive statute, and that because 42 U.S.C. § 1988 allows a reasonable attorney's fee "as part of the costs," a civil rights plaintiff who rejected an offer and recovered less could not recover post-offer fees. The plaintiffs in Marek had rejected a $100,000 offer and won about $60,000 at trial, and they lost their claim to roughly $139,000 in post-offer fees. That is leverage.

Second, when the statute makes fees costs and the defendant seeks its own fees. The circuits divide on whether a defendant that made a Rule 68 offer can recover its own post-offer fees from a copyright plaintiff who won less, because 17 U.S.C. § 505 awards fees "to the prevailing party as part of the costs." The Eleventh Circuit in Jordan v. Time, Inc., 111 F.3d 102 (11th Cir. 1997) included the defendant's fees; the Seventh Circuit in Harbor Motor Co. v. Arnell Chevrolet-Geo, Inc., 265 F.3d 638 (7th Cir. 2001) disagreed, reasoning that a non-prevailing defendant cannot recover fees under § 505. In civil rights cases, courts have generally held that Rule 68 does not make a plaintiff pay the defendant's post-offer fees, because § 1988 allows a prevailing defendant fees only under the demanding standard of Christiansburg Garment Co. v. EEOC, 434 U.S. 412 (1978).

Third, as a psychological anchor. A formal offer of judgment, served with a cover letter explaining the post-offer cost exposure, forces the plaintiff's lawyer to calculate the case and report the offer to the client. In a contingent-fee case, the lawyer's own fee recovery may be at stake, which concentrates attention.

Rule 68 has limits. It is available only to a party defending against a claim; plaintiffs cannot use it. It applies only when the plaintiff wins something less than the offer; in Delta Air Lines, Inc. v. August, 450 U.S. 346 (1981), the Court held that the rule does not apply when the defendant wins outright. An unaccepted offer does not moot the plaintiff's claim, as the Court held in Campbell-Ewald Co. v. Gomez, 577 U.S. 153 (2016), so it cannot be used to pick off a class representative. And an accepted offer results in a judgment, which some defendants do not want on the record. Ambiguity about whether the offer includes fees and costs is construed against the offeror, so the terms must be precise. Many states have their own offer statutes that work differently and may be available to plaintiffs, such as California Code of Civil Procedure § 998 and Florida Statutes § 768.79. Their thresholds, timing rules, and formal requirements differ, and courts enforce them strictly, so read the current text before serving one.

Fee-shifting statutes

Under the American rule each side pays its own lawyers, but many statutes shift fees, and the possibility of a fee award changes both sides' BATNAs. In patent cases, 35 U.S.C. § 285 allows fees to the prevailing party in "exceptional cases," and Octane Fitness, LLC v. ICON Health & Fitness, Inc., 572 U.S. 545 (2014) made that standard easier to meet by asking whether a case "stands out from others" in substantive strength or the manner of litigation. The Lanham Act uses the same "exceptional cases" language in 15 U.S.C. § 1117(a), and most circuits now apply Octane to it. Copyright fees under § 505 are discretionary and evenhanded between plaintiffs and defendants under Fogerty v. Fantasy, Inc., 510 U.S. 517 (1994), with substantial weight on the objective reasonableness of the losing party's position under Kirtsaeng v. John Wiley & Sons, Inc., 579 U.S. 197 (2016). In civil rights cases, a prevailing plaintiff ordinarily recovers fees under § 1988, calculated by the lodestar method of Hensley v. Eckerhart, 461 U.S. 424 (1983).

Fee exposure affects settlement in predictable ways. In a fee-shifting case, the plaintiff's demand grows with every hour its lawyers bill, so defendants have reason to settle early. Defendants can also negotiate over fees directly. In Evans v. Jeff D., 475 U.S. 717 (1986), the Court held that a court could approve a settlement conditioned on the plaintiffs' waiver of statutory fees, which means defendants may lawfully offer full merits relief in exchange for a fee waiver. Plaintiffs' lawyers respond by negotiating fees separately or by fee agreements that address waiver. And the form of the settlement matters: under Buckhannon Board & Care Home, Inc. v. West Virginia Department of Health & Human Resources, 532 U.S. 598 (2001), a plaintiff who obtains its goal through a private settlement without a judicially sanctioned change in the parties' legal relationship is not a "prevailing party" for fee purposes. A consent decree may support fees; a private agreement and voluntary dismissal may not.

For more on fee recovery, see Attorneys Fees and Costs: The American Rule and Its Many Exceptions and Lanham Act Attorney's Fees Under 15 U.S.C. § 1117(a): A Practical Guide.

Settlement in intellectual property cases

IP disputes have features that make the negotiation different from an ordinary damages case. The remedies include injunctions that can shut down a product line, the rights at issue can be challenged in parallel administrative proceedings, and the settlement often takes the form of a license or a coexistence arrangement that continues for years.

Patent cases

Injunction leverage is weaker than it used to be. After eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006), permanent injunctions require the traditional four-factor showing, and patentees that do not practice the invention often cannot get one. That pushes many patent negotiations toward royalty economics rather than shutdown risk.

Enhanced damages and fees raise the defendant's downside. 35 U.S.C. § 284 allows a court to increase damages up to three times, and Halo Electronics, Inc. v. Pulse Electronics, Inc., 579 U.S. 93 (2016) gave district courts discretion to award enhancement for egregious, typically willful, infringement. A defendant that received a notice letter and did nothing has a worse WATNA. See Willful Patent Infringement and Enhanced Damages After Halo: Knowledge, Opinions of Counsel, and Section 298.

Damages are bounded by statute. 35 U.S.C. § 286 bars recovery for infringement more than six years before the complaint, and 35 U.S.C. § 287 limits damages for patentees who sell unmarked patented products until actual notice. Both can cut the plaintiff's expected value sharply.

The PTAB is a second front. An accused infringer can petition for inter partes review, but 35 U.S.C. § 315(b) bars a petition filed more than one year after service of an infringement complaint. The institution decision is a major settlement event: institution weakens the patentee's position, and denial strengthens it. Discretionary-denial practice at the USPTO has shifted repeatedly: in 2025 alone the Office rescinded its 2022 guidance limiting denials based on parallel litigation (February) and adopted an interim process under which discretionary-denial arguments are briefed and decided separately from, and before, the merits (March). Check the Office's current guidance before building a strategy around institution odds. See Inter Partes Review Strategy: Discretionary Denial, Estoppel, and the Parallel District Court Case.

Settlement terms carry antitrust and validity questions. A licensee may challenge the validity of a licensed patent without first repudiating the license, MedImmune, Inc. v. Genentech, Inc., 549 U.S. 118 (2007), and Lear, Inc. v. Adkins, 395 U.S. 653 (1969) abolished licensee estoppel. Patentees therefore often negotiate for no-challenge clauses, termination rights on challenge, or other protections, whose enforceability depends on context; courts have been more receptive to no-challenge terms in settlements of litigated disputes than in ordinary licenses; the Federal Circuit enforced one in a settlement agreement in Flex-Foot, Inc. v. CRP, Inc., 238 F.3d 1362 (Fed. Cir. 2001), where the challenger had the chance to litigate validity before settling. In pharmaceutical cases, FTC v. Actavis, Inc., 570 U.S. 136 (2013) held that large and unjustified "reverse payments" from a brand to a generic challenger can violate the antitrust laws under the rule of reason. See Hatch-Waxman Toolkit: Notice Letters, Complaints, and Settlement Considerations.

Trademark cases

Trademark cases usually turn on the future more than the past. The plaintiff wants the confusing use stopped; the defendant wants time and money to rebrand. Injunctions are available under 15 U.S.C. § 1116, and profits are recoverable under § 1117(a) without a threshold showing of willfulness after Romag Fasteners, Inc. v. Fossil, Inc., 590 U.S. 212 (2020), although willfulness remains an important consideration. The negotiation currency is time (a sell-off or phase-out period), scope (which products, territories, and channels), and certainty (abandonment of applications, consent to registration, a coexistence agreement with defined boundaries).

Covenants not to sue are also a tactical tool. In Already, LLC v. Nike, Inc., 568 U.S. 85 (2013), Nike's broad unconditional covenant not to sue mooted the defendant's counterclaim to cancel Nike's registration. A trademark owner facing a cancellation counterclaim it does not want to litigate can sometimes end that risk with a covenant broad enough to satisfy the voluntary cessation standard.

Copyright cases

Copyright negotiations are shaped by statutory damages and registration timing. 17 U.S.C. § 504(c) provides for statutory damages of $750 to $30,000 per work, up to $150,000 for willful infringement and down to $200 for innocent infringement. But under 17 U.S.C. § 412, statutory damages and attorney's fees are generally unavailable when infringement began before registration (subject to a three-month grace period after first publication). The first question in any copyright negotiation is therefore the registration date; a plaintiff limited to actual damages and with no fee exposure for the defendant has much less leverage. Image licensing demands illustrate this; see Image Copyright Demand Letters: How Photo Licensing Enforcement Works and What Claims Are Really Worth. Accessibility cases are another area where fee-shifting and injunctive relief drive early settlements; see Website Accessibility Lawsuits: What the ADA Requires, What Breaks in the Code, and What to Do When the Letter Arrives.

Enforcement: getting a deal that holds

A negotiation is not over when the parties shake hands. It is over when the money is paid, the conduct has stopped, and the case is closed in a way that lets the client enforce the deal if the other side does not perform.

The key case is Kokkonen v. Guardian Life Insurance Co. of America, 511 U.S. 375 (1994). The parties settled a federal case and stipulated to dismissal under Fed. R. Civ. P. 41(a)(1). When a dispute later arose over the settlement, the district court purported to enforce it. The Supreme Court held that the court lacked jurisdiction: enforcing a settlement agreement is a claim for breach of contract, which needs its own basis for federal jurisdiction. A court retains jurisdiction to enforce only if the dismissal order expressly retains it or incorporates the settlement's terms. Otherwise, enforcement belongs in state court, as a new contract action.

Negotiators should address enforcement at the table, not in the drafting stage, because enforcement terms have value and can be traded. Options include a dismissal order that retains jurisdiction, a consent judgment, a stipulated judgment held in escrow and filed only on default, security or a guaranty from a solvent affiliate, and a short cure period with an acceleration clause. A plaintiff dealing with a financially weak defendant should treat these as part of the price; a defendant should expect to pay less in cash if it provides more security. See Settlement Toolkit: Term Sheets, Releases, Offers of Judgment, and Enforcement.

The other enforcement problem is proving there was a deal at all. Oral agreements reached at mediation may or may not be enforceable, depending on the jurisdiction's mediation confidentiality rules and contract law. Many state mediation statutes require a signed writing to escape the privilege. The safe practice is a signed term sheet before anyone leaves, stating that it is binding and enforceable and that it may be disclosed to enforce it, with authority confirmed on the record.

Extended worked example: Northfield Coffee Roasters v. Ridgeline Brewing

The following example is invented. It shows how the ideas above work together in a single negotiation.

The dispute

Northfield Coffee Roasters, a 40-employee specialty roaster, has used the NORTHFIELD mark on coffee since 2011 and owns a federal registration for roasted coffee and coffee beverages. In early 2025, Ridgeline Brewing, a regional craft brewer, launched a canned nitro cold brew called "Ridgeline Northfield Nitro," named after a hiking trail near its brewery. Within six months, Northfield's customer service team had logged 37 emails and calls from consumers asking whether Northfield had partnered with Ridgeline. Northfield sued for infringement under 15 U.S.C. § 1114 and § 1125(a), seeking an injunction, Ridgeline's profits, and fees.

After 14 months of litigation, the parties face a court-ordered mediation. Discovery shows that Ridgeline sold about $1.9 million of the product. Ridgeline claims a net profit margin of 12 percent (about $228,000); Northfield's expert argues that Ridgeline's deductions are overstated and puts profits nearer $520,000. Ridgeline holds roughly 400,000 printed cans and a pending trademark application for RIDGELINE NORTHFIELD NITRO. Northfield plans to launch its own ready-to-drink cold brew in grocery channels next spring.

Interests behind the positions

Northfield's opening position is $1.2 million and an immediate injunction. Its interests, as its lawyers learn by asking the founder and CFO, are: stopping confusion before its own launch, avoiding a public fight with a well-liked local brand, recovering at least part of the $480,000 it has spent on fees, and having something to show its investors.

Ridgeline's opening position is no payment and a name change in 18 months. Its interests are: not writing off $90,000 of printed cans, keeping shelf placement at three grocery chains while it rebrands, avoiding a judgment that could be cited by other trademark owners, and staying within a litigation budget that its board capped at $250,000 beyond what it has already spent.

Valuation

Northfield's lawyers build a decision tree. They estimate a 70 percent chance of establishing likelihood of confusion (the actual-confusion evidence is strong, but the mark is arguably geographically descriptive and Ridgeline will argue weakness). If Northfield wins, they estimate profits awards of $200,000 (35 percent), $400,000 (45 percent), and $600,000 (20 percent), and a 25 percent chance of an exceptional-case fee award of about $700,000. The probability-weighted monetary recovery is roughly $260,000 in profits and $120,000 in fees, about $380,000, against $350,000 in fees still to be spent through trial. Northfield's net expected monetary value from trial is therefore near zero, but an injunction is very likely if it wins, and the injunction is what Northfield most needs. That changes the analysis: the main value of trial to Northfield is a court-ordered name change, which Ridgeline is already partly offering.

Ridgeline's lawyers run their own tree. They see a 60 percent chance of losing liability and a meaningful risk of fee-shifting because of the confusion evidence. Their WATNA is an injunction with no sell-off period (a $90,000 inventory loss plus emergency rebranding costs of perhaps $250,000), a profits award, and a fee award. Their expected cost of trying the case, including their own fees, comes to more than $700,000.

The two trees suggest a ZOPA on money between roughly $150,000 (below which Northfield would rather keep litigating to get fees) and $450,000 (above which Ridgeline's board would rather gamble), with the real value created by trading time for certainty.

Anchors and the first exchange

Northfield's mediation statement anchors at $1.2 million, but supports the number with the expert's profits figure, the fee exposure, and the actual-confusion log. Ridgeline's statement responds with a carefully reasoned anchor of its own: a $40,000 payment, a 12-month phase-out, and a mutual release, supported by an argument that the mark is weak and that Northfield's actual damages are speculative.

In the opening caucus, Ridgeline's CEO says the $1.2 million demand is "a shakedown" and wants to leave. The mediator goes to the balcony with him: what happens if he leaves? His lawyer walks him through the tree again. He stays.

The dance, brackets, and a hardball moment

By 2:00 p.m., Northfield is at $850,000 and an immediate injunction; Ridgeline is at $90,000 and a nine-month phase-out. Concessions are shrinking. The mediator suggests brackets. Northfield proposes moving to between $600,000 and $400,000 (midpoint $500,000) if Ridgeline will move to between $150,000 and $300,000 (midpoint $225,000). Ridgeline counters with a bracket of $120,000 to $250,000 (midpoint $185,000).

At 3:30 p.m., Ridgeline's lawyer tells the mediator that its offer "expires at 4:00 because the CEO has a flight." Northfield's lawyer asks the mediator whether the flight is real and whether the CEO will be reachable by phone. He is. The deadline evaporates.

The parties then turn to non-monetary terms, and the value creation happens. Northfield cares about the period before its spring launch. Ridgeline cares about its printed cans. The solution is a 120-day sell-off period limited to existing inventory, with no new production under the NORTHFIELD name, Ridgeline's express abandonment of its pending application, and a new name chosen by Ridgeline that does not include "Northfield" or any confusingly similar term. Northfield gains certainty before its launch; Ridgeline avoids the inventory loss. Both sides become more flexible on money.

The mediator's proposal

At 5:15 p.m., with Northfield at $410,000 and Ridgeline at $210,000, the mediator makes a mediator's proposal: $285,000, payable $150,000 within 30 days and the remainder in 90 days; the 120-day sell-off; abandonment of the application; a consent judgment and permanent injunction entered on the court's docket with the court retaining jurisdiction to enforce the settlement; a stipulated $5,000 per day of noncompliance after the sell-off period; mutual releases; and confidentiality of the payment amount only. Each side has until 7:00 p.m. to say yes or no privately.

Northfield's lawyers walk the founder through the numbers. The payment covers more than half the fees already spent, and the consent judgment gives Northfield what a trial win would give it (an enforceable injunction) without the 30 percent risk of losing. Ridgeline's lawyers point out that $285,000 is within the board's cap once trial fees are counted and that the consent judgment, while public, contains no findings of willfulness. Both say yes.

Closing

Before anyone leaves, the lawyers sign a two-page term sheet that states it is binding, identifies the consent judgment and retention of jurisdiction as required terms in light of Kokkonen, and provides that it may be disclosed to the court to enforce it. The lawyers confirm on the term sheet that each signatory has authority. Formal papers follow within ten days.

The example illustrates the main points. Neither side's opening number mattered as much as its valuation. Interests (timing, inventory, launch plans) created a deal that numbers alone would not. The bracket clarified midpoints. The mediator's proposal sidestepped reactive devaluation. The term sheet anticipated enforcement. For a stage-by-stage preparation list, see Settlement Negotiation Preparation Checklist: A Practical Checklist.

Common mistakes

Negotiating without a valuation. If the lawyer cannot state the client's BATNA in dollars, the other side's anchor will fill the vacuum.
Confusing the opening with the outcome. Treating a first demand or offer as an insult, rather than as the start of a ritual, ends negotiations that could have succeeded.
Ignoring interests. Cases with no apparent ZOPA in dollars often have one once timing, payment terms, non-monetary relief, and future business are considered.
Bidding against yourself. Making a second move before the other side responds to the first signals weakness and lowers the outcome.
Letting the clock decide. Late-day concessions made from fatigue are the most expensive ones. Plan the concession sequence in advance and take breaks.
Lying about facts. Puffery about value is allowed; misstatements of insurance limits, evidence, or a client's condition are not, and can unwind the deal and lead to discipline.
Assuming confidentiality. Rule 408 limits admissibility for certain purposes, not discovery or all uses, and mediation privileges vary by jurisdiction.
Misusing Rule 68. An ambiguous offer that does not address fees and costs, or that is made in a case where fees are not "costs," can misfire.
Leaving without a signed term sheet. Oral deals at mediation are fragile, and mediation confidentiality rules can make them unprovable.
Ignoring enforcement. A voluntary dismissal without retained jurisdiction sends any breach claim to state court.

Frequently asked questions

Should a plaintiff ever make an opening demand far above the expected verdict? A high anchor helps only if it is credible. A demand with no supportable basis invites an equally extreme counter or no response at all. The best practice is to anchor at the high end of the defensible range and to explain why.

Is it ethical to tell a mediator that the client will not settle below a number when that is not true? Under Model Rule 4.1 comment [2] and ABA Formal Opinion 06-439, statements about a party's intentions regarding an acceptable settlement are ordinarily not statements of material fact, so posturing about a bottom line is generally permitted. State rules and court rules may differ, and a judge conducting a settlement conference is entitled to more candor under Rule 3.3.

Can a defendant use Rule 68 to force a settlement? No. Rule 68 cannot force acceptance. It shifts post-offer costs (and fees, if the statute defines them as costs) when the plaintiff obtains a judgment that is not more favorable than the offer. It is leverage, not compulsion.

How do brackets differ from ordinary offers? A bracket is a conditional move. It communicates a midpoint and invites the other side to move by a comparable amount. Brackets are useful when parties are far apart and each fears that a large unilateral concession will be read as weakness.

What if the other side refuses to send someone with authority? Raise it with the mediator early, check the local ADR rule and any court order, and consider asking the court for an order requiring attendance of a representative with authority, which G. Heileman Brewing confirms a federal court may issue in appropriate circumstances.

Can what we say at mediation be used against us later? Possibly. Rule 408 bars use to prove liability or amount but allows other purposes, and settlement communications may be discoverable in some jurisdictions. Mediation privileges under state law and local rules vary. Treat mediation as confidential but not sealed.

Does settling an IP case with one defendant hurt us with others? It can. Settlement licenses may be relevant to reasonable royalty damages in later cases, although courts discount litigation-driven licenses. A low early settlement can become a benchmark.

Practical takeaways

Value the case in numbers, as a range, before any negotiation, and show the client the decision tree.
Find the interests behind the dollar positions; most creative deals come from timing, payment terms, and non-monetary relief.
Anchor first when you can, with a number you can justify, and do not negotiate off an unsupported anchor.
Expect reactive devaluation and use the mediator to present proposals that would be rejected if they came from you.
Plan a concession sequence, keep moves shrinking, and do not bid against yourself.
Recognize hardball tactics, name them calmly, and redirect to the problem.
Puff about value if you must; never misstate facts.
Do not assume settlement talk is confidential for every purpose.
Use Rule 68 and fee-shifting statutes deliberately, with precise terms.
Sign a binding term sheet before leaving, and build enforcement into the deal.

This document is general information, not legal advice for any particular situation; the law changes, and the facts matter.