Summary. This guide takes a litigator from the decision to explore settlement through a signed term sheet, in thirteen steps that work for plaintiffs and defendants alike. It shows how to value a case as a range of outcomes net of fees, time, interest, taxes, liens, and collection risk, and how to turn that valuation into a reservation price and a target. It covers confirming client authority and interests, improving the alternative to settlement before talks begin, choosing between direct talks, private mediation, and a judicial settlement conference, and building an opening strategy and a written concession plan. It then walks through mediation preparation, including the mediation statement and the decision whether to hold a joint opening session, and running the day itself, from first caucus to brackets and mediator's proposals. The final steps cover closing the gap, writing a binding term sheet at the table that anticipates enforcement under Kokkonen, and documenting the result, followed by a timeline and cost table, a worked example in a software dispute, common mistakes, an FAQ, and takeaways.

The scheduling order says mediation must be completed by March 14. It is January 20. The partner on the case asks the senior associate a simple question: "What is this case worth, and what do we want out of mediation?" The associate has a 40-page summary judgment brief in draft, a damages expert's preliminary numbers, and a client who has told three different people three different things about what a "fair" result would be. None of that is an answer.

The answer has to be built, and it has to be built before mediation, not during it. A negotiation that starts without a valuation, without confirmed authority, and without a written plan for how the numbers will move is a negotiation the other side will run. This guide is the build sequence. It is written for counsel on either side of a civil case, and it flags where the steps differ for plaintiffs and defendants.

The analytical background (why anchoring works, what reactive devaluation is, the ethics of puffery, Rule 68 doctrine) is in the companion article, Settlement Negotiation Tactics: Principled Negotiation, Anchoring, and Handling Hardball in Litigation. The worksheets referred to below are in Settlement Negotiation Toolkit: Case Valuation Worksheets, Concession Plans, and Offer Letters, and the task list is in Settlement Negotiation Preparation Checklist: A Practical Checklist.

The short answer

  • Value first, talk second. Build a decision tree with explicit probabilities, net out remaining fees, interest, taxes, liens, and collection risk, and express the result as a range. That range, not the pleadings, is what you are negotiating against.
  • Separate three numbers. The reservation price (walk-away), the target (realistic best), and the opening (anchor) serve different purposes. Write all three down, with the client's sign-off, before the first exchange.
  • Confirm authority and interests in writing. The client decides whether to settle. Find out who that is for an entity or insured defendant, what they need besides money, and what you are authorized to offer or accept.
  • Improve your alternative before you negotiate. A pending dispositive motion, a Rule 68 offer, an expert report, or a coverage position can move the other side further than any argument at the table.
  • Pick the right format. Direct talks are cheap and fast. Private mediation adds a neutral who can reality-test and carry proposals. A judicial settlement conference adds authority and pressure. Each has costs and confidentiality rules.
  • Plan the moves. A written concession plan, with increments, conditions, and non-monetary trades, keeps the client from negotiating against fatigue or the clock.
  • Prepare the mediation statement as advocacy and the client as a participant. Decide whether to hold a joint session, what the mediator may share, and how the client will be involved.
  • Close with paper. A signed, binding term sheet that addresses payment, releases, dismissal, and retained jurisdiction should exist before anyone leaves.

Before you start: five decision points

1. Is now the right time? Settlement value moves with the case. Early talks save fees but happen before discovery reduces uncertainty. Talks after a key ruling (a motion to dismiss, class certification, claim construction, a Daubert ruling, summary judgment) often succeed because the ruling resolves the variable the parties disagreed about. Talks on the eve of trial benefit from maximum information and maximum cost pressure. Ask which upcoming event most changes the value, and whether it is better to negotiate before or after it.

2. Who has to agree? For an individual, the client. For a company, a named executive, sometimes a board or committee. For an insured defendant, the insurer, often with a coverage counsel and a reservation of rights. For a class, the court under Fed. R. Civ. P. 23(e). For a minor or incompetent person, typically a court, under state procedure and Fed. R. Civ. P. 17(c). Identify every approval and its timeline before you plan the session.

3. Is the other side able to pay? A defendant without insurance or assets cannot fund a settlement, however good your arguments. Check insurance (initial disclosures under Fed. R. Civ. P. 26(a)(1)(A)(iv) require production of insurance agreements that may satisfy a judgment), public financial information, liens, and bankruptcy risk.

4. What does the client actually want? Money is usually part of it. So are timing, confidentiality, a continuing relationship, a change in conduct, an apology, tax treatment, or the end of a distraction. The interests determine which deals are possible.

5. Are there structural limits? Fee-shifting statutes, Rule 68 offers already served, insurance policy limits that erode with defense costs, court approval requirements, lien holders (Medicare, ERISA plans, hospitals), and multiple plaintiffs or defendants (aggregate settlement rules under Model Rule 1.8(g)) all constrain what can be agreed. List them now.

Step 1: Decide whether and when to negotiate

What to do. Identify the case events over the next six to twelve months and estimate how each will change value. Decide whether to propose talks now, wait for a ruling, or respond to a court-ordered process. Check the scheduling order, the local ADR rule, and any judge-specific requirements.

Why. Under 28 U.S.C. § 652, litigants in civil cases in federal district court must consider using ADR, and many courts require mediation or a settlement conference. Fed. R. Civ. P. 16(c)(2)(I) lists "settling the case and using special procedures to assist in resolving the dispute" among the matters a court may take up at a pretrial conference. Even without a court order, the Rule 26(f) conference is a natural point to raise ADR.

Plaintiff's view. Early talks can be valuable when liability is clear, the defendant is facing fee exposure, or the plaintiff needs money quickly. They are risky when the plaintiff has not yet obtained the evidence that proves damages. Raising settlement early is not a sign of weakness if it is framed around efficiency.

Defendant's view. Early talks limit fee spend and, in fee-shifting cases, cap the plaintiff's growing fee claim. Defendants often prefer to negotiate after a motion to dismiss narrows the case or after the plaintiff's damages theory has been tested. A defendant considering a Rule 68 offer should time it so that the post-offer period captures the expensive phase of the case.

Pitfalls. Waiting for "more information" indefinitely; negotiating immediately before a ruling that is likely to favor you; letting the mediation deadline in the scheduling order dictate timing when a different date would be better (courts will often extend it on a joint request).

Timing and cost. This decision takes hours, not days. Revisit it after every significant ruling.

Step 2: Build the valuation

What to do. Build a decision tree for the case as it stands, then convert it into a net present expected value and a range. The tree should include each major contingency (dispositive motions, liability at trial, damages outcomes, fee awards, appeal), with an explicit probability on each branch.

Why. Without numbers, the other side's anchor becomes the reference point. With numbers, every offer can be compared to something. For background on case evaluation generally, see Evaluating and Assessing a Civil Case.

The components.

Component Questions to answer Sources and authority
Liability probability What is the chance of surviving each dispositive stage and winning at trial? Pleadings, evidence, rulings, judge's record, jury pool
Damages range What are the low, middle, and high plausible awards, and how likely is each? Expert reports; statutory ranges such as 17 U.S.C. § 504(c); enhancement under 35 U.S.C. § 284; verdict reports
Punitive or enhanced damages Is there a realistic path, and what is the constitutional ceiling? State Farm Mutual Automobile Insurance Co. v. Campbell, 538 U.S. 408 (2003) and BMW of North America, Inc. v. Gore, 517 U.S. 559 (1996) on ratios
Fees to date and to come What will each side spend through trial and appeal? Budgets, billing records
Fee-shifting Can either side recover fees if it wins? 35 U.S.C. § 285, 15 U.S.C. § 1117(a), 17 U.S.C. § 505, 42 U.S.C. § 1988, contract fee clauses
Taxable costs What costs will the prevailing party recover? 28 U.S.C. § 1920; Fed. R. Civ. P. 54(d)
Time and interest When would a judgment be paid, and what interest accrues? Post-judgment interest under 28 U.S.C. § 1961; prejudgment interest under the governing substantive law
Collectability Can the defendant pay a judgment? What does insurance cover? Insurance disclosures, financial information; bankruptcy risk under 11 U.S.C. § 547 and § 362
Taxes What is the after-tax value of a payment? 26 U.S.C. § 104; 26 U.S.C. § 61; Commissioner v. Banks, 543 U.S. 426 (2005) (contingent fees generally included in plaintiff's gross income)
Liens Who else has a claim on the proceeds? Medicare Secondary Payer provisions, 42 U.S.C. § 1395y(b); ERISA plan terms; state hospital lien statutes
Non-monetary value What is an injunction, a license, a reference, or a confidentiality term worth? Client's business data

How to run the numbers. Multiply through the tree to get a probability-weighted gross outcome. Subtract the client's remaining fees and costs (plaintiff) or add them (defendant). Discount for time if payment is years away. Adjust for collectability. Then run the tree twice more with pessimistic and optimistic estimates. The three results give a range.

Plaintiff's view. In a contingent-fee case, show the client the net to the client after fees and costs, not the gross. A $500,000 settlement and a $700,000 verdict two years from now may be close in net terms once the fee percentage, costs, and interest are considered, and the verdict carries risk.

Defendant's view. Include the full cost of losing: the judgment, the plaintiff's fees under any fee-shifting statute, interest, and the cost of an appeal bond. Include the cost of winning, too: fees through trial and appeal are spent either way.

Pitfalls. Using point estimates instead of ranges; letting the trial team's optimism set the probabilities (ask a colleague not on the case to review them); ignoring collection risk; leaving out fees on the other side of a fee-shifting statute; double-counting by treating the plaintiff's demand as the "high" damages scenario.

Timing and cost. A first-pass tree takes a few hours. A careful one, with expert input on damages, may take one to three days of lawyer time. Refresh it after each major event.

For damages modeling in depth, see Building a Damages Case: A Practical Guide from Pleading to Verdict.

Step 3: Map the client's interests and confirm authority

What to do. Meet with the decision maker, not just the day-to-day contact. Ask open questions about what a good outcome looks like, what the client fears, and what matters besides money. Then confirm, in writing, who will make the settlement decision, what the lawyer is authorized to offer or accept, and how authority can be expanded during the session.

Why. Model Rule 1.2(a) requires a lawyer to abide by a client's decision whether to settle. Model Rule 1.4 requires the lawyer to keep the client informed and to promptly communicate settlement offers unless the client has given prior instructions. Settlements made without authority may not bind the client, depending on state agency law, and the lawyer may face discipline or a malpractice claim. Courts can also require that a representative with authority attend: Fed. R. Civ. P. 16(c)(1) allows a court to require that a party or its representative be present or reasonably available by other means, and G. Heileman Brewing Co. v. Joseph Oat Corp., 871 F.2d 648 (7th Cir. 1989) (en banc) upheld an order requiring a corporate representative with authority to settle.

Interest-mapping questions.

  • If this case ended tomorrow, what would you do with the time and money it is consuming?
  • What would a result have to include for you to explain it to your board, investors, family, or customers?
  • What are you most worried about if the case continues?
  • Is there any business relationship with the other side worth preserving?
  • Are there terms (timing, confidentiality, a statement, a change in practice) that matter as much as money?
  • What would make you feel the other side had taken this seriously?

Insured defendants. The insurer usually controls settlement within limits, but the insured has interests too: uninsured claims (such as punitive damages in some states), deductibles or self-insured retentions, reputational harm, and exposure above limits. When defense costs erode the policy limits, every month of litigation reduces the money available for settlement, which is an interest the plaintiff should understand too. Coordinate with coverage counsel; see Business Insurance and Coverage Disputes: CGL, E&O, Cyber, and D&O.

Funded plaintiffs. If a litigation funder has a stake, its consent rights and economics may affect what the plaintiff can accept. See Third-Party Litigation Funding: Structures, Privilege, and Disclosure Obligations.

Pitfalls. Taking authority from someone who does not have it; failing to explain the lawyer's own fee incentives; accepting "do the best you can" as authority; not planning for how to reach the decision maker at 7 p.m.

Timing and cost. One or two meetings, plus a short written authority memo. Expect corporate approvals to take a week or more.

Step 4: Set the BATNA, WATNA, reservation price, and target

What to do. From the valuation, derive four numbers and write them on one page, signed off by the client:

  • BATNA value. The net expected value of continuing to litigate (Step 2's middle estimate).
  • WATNA. The plausible bad outcome: a dispositive loss, a fee award against the client, or a judgment at the high end.
  • Reservation price. The point at which the client would rather litigate than settle. For a plaintiff, the lowest acceptable number; for a defendant, the highest. It is derived from the BATNA, then adjusted for the client's risk tolerance, cash needs, and non-monetary interests.
  • Target. The realistic best outcome, typically somewhere between the reservation price and the opening.

Then estimate the other side's numbers. What is their likely BATNA? What does their fee spend look like? What pressures (insurance limits, a fiscal year, a board meeting, a product launch, a funder) affect them? The overlap between your reservation price and your estimate of theirs is the zone of possible agreement.

Why. The reservation price prevents two errors: accepting less than litigation is worth, and walking away from a deal that beats it. The target keeps the negotiator from settling for the first number above the reservation price. Writing them down before the session reduces the effect of anchors, fatigue, and the "we've come this far" bias on the day.

Pitfalls. Treating the reservation price as a number to announce (never disclose it as such); setting it by reference to the fees already spent (sunk costs are irrelevant); adjusting it during the session without a reason grounded in new information. A reservation price can move if new information arrives (a document, a candid concession about a weakness), but it should not move because the other side was persistent.

Timing and cost. An hour with the client after the valuation is done. Revisit before the session.

Step 5: Improve the alternative before you negotiate

What to do. Identify moves that would strengthen your BATNA or weaken the other side's assessment of theirs, and decide whether to make them before talks start.

Options for either side.

  • File or win a dispositive motion. A pending summary judgment motion forces the other side to price the risk of losing. See Summary Judgment Under Rule 56: Winning and Surviving Before Trial.
  • Serve an expert report. A well-supported damages report is both a litigation step and an anchor.
  • Take the key deposition. A strong admission changes the probability on a branch of the other side's tree.
  • Fix your own weaknesses. Get the missing document, retain the needed expert, cure the pleading defect, or correct the damages theory.

Defendant-specific tools. A Fed. R. Civ. P. 68 offer of judgment shifts post-offer costs to a plaintiff who recovers less. Where the underlying statute defines fees as costs, the effect can be large: Marek v. Chesny, 473 U.S. 1 (1985) held that a civil rights plaintiff who rejected an offer and recovered less could not recover post-offer fees. The rule does not apply if the defendant wins outright, Delta Air Lines, Inc. v. August, 450 U.S. 346 (1981). In patent cases, an inter partes review petition under 35 U.S.C. § 311 can reduce the patentee's leverage, but the petition must be filed within the one-year window in 35 U.S.C. § 315(b).

Plaintiff-specific tools. Plaintiffs cannot use federal Rule 68, but some states have two-way offer statutes, such as California Code of Civil Procedure § 998. Plaintiffs in insured cases sometimes use a time-limited demand within policy limits to put the insurer at risk of liability above limits if it unreasonably refuses; the rules on such demands vary by state, and several states, Georgia (O.C.G.A. § 9-11-67.1) and Missouri (Mo. Rev. Stat. § 537.058) among them, now prescribe their content and timing by statute, so read the governing statute before sending or answering one. In fee-shifting cases, a documented record of reasonable early offers supports a later fee petition.

Pitfalls. Filing a motion purely for leverage (it must be well grounded under Fed. R. Civ. P. 11); serving a Rule 68 offer that is ambiguous about fees and costs; waiting so long to improve the alternative that the negotiation window closes.

Timing and cost. Varies widely. A Rule 68 offer costs little to prepare. A summary judgment motion may cost tens of thousands of dollars, which should be weighed against its effect on settlement value.

Step 6: Choose the format and the neutral

What to do. Decide among direct negotiation, private mediation, a court-annexed mediation program, or a judicial settlement conference. If using a neutral, select one and agree on the terms of engagement.

Format Best for Costs Confidentiality
Direct lawyer-to-lawyer Small cases; cases where counsel trust each other; early feelers Low Fed. R. Evid. 408 admissibility limits only; consider a written confidentiality agreement
Private mediation Most commercial and IP cases; cases with emotional or relationship issues Mediator fees often several thousand dollars to well over $10,000 per day, usually split Mediation agreement plus state law, such as the Uniform Mediation Act where adopted
Court-annexed mediation Cases in districts with ADR programs; parties with limited budgets Often low or free Local rules adopted under 28 U.S.C. § 652(d); neutrals under 28 U.S.C. § 653
Magistrate or judicial settlement conference Cases where a judicial voice will move a party; repeat-player defendants Low Local rules; the settlement judge typically does not preside at trial

Selecting a mediator. Consider subject-matter knowledge (a patent mediator should understand royalty bases), style (facilitative, evaluative, or directive), credibility with the other side's decision maker, availability, and willingness to make a mediator's proposal. Ask colleagues who have used the mediator recently. In the First Circuit, In re Atlantic Pipe Corp., 304 F.3d 135 (1st Cir. 2002) recognized that a district court may in some circumstances order mediation even over objection, so be prepared to participate in a process you did not choose.

Confidentiality terms. Sign a mediation agreement that addresses governing law, confidentiality, the mediator's immunity from subpoena, and whether a signed term sheet is enforceable and may be disclosed to enforce it. Do not assume confidentiality extends to discovery: Fed. R. Evid. 408 governs admissibility, and courts disagree on a discovery privilege. Compare Goodyear Tire & Rubber Co. v. Chiles Power Supply, Inc., 332 F.3d 976 (6th Cir. 2003) (recognizing a settlement privilege) with In re MSTG, Inc., 675 F.3d 1337 (Fed. Cir. 2012) (declining to recognize one). Strict state mediation statutes can cut the other way: Cassel v. Superior Court, 51 Cal. 4th 113 (2011) barred evidence of mediation communications even in a client's later malpractice case.

Pitfalls. Choosing a mediator the other side does not respect; failing to confirm the mediator's availability for a follow-up session; agreeing to a half-day session for a case that needs a full day; leaving confidentiality to defaults.

Timing and cost. Book good mediators four to eight weeks out, sometimes longer. For the mechanics of mediation more broadly, see Mediation and Settlement: Preparing, Negotiating, and Documenting the Deal.

Step 7: Set the opening strategy

What to do. Decide who will make the first number, what it will be, how it will be justified, and when and how it will be delivered (demand letter, mediation statement, or at the session).

Why. First numbers anchor. A credible, well-supported first number pulls the range toward it; an incredible one can provoke an equally extreme response or end talks. The plaintiff conventionally goes first in litigation, which gives the plaintiff the first anchor and the defendant the first chance to re-anchor.

Choosing the number.

  • Start from the top (plaintiff) or bottom (defendant) of the defensible range from Step 2, not from the pleading's prayer for relief.
  • Attach reasons: the damages model, comparable verdicts or licenses, the statutory range, the fee exposure.
  • Leave room for the concession plan in Step 8. If the target is $600,000 and you expect four rounds of concessions, an opening of $650,000 leaves no room to move.
  • Consider an opening that includes non-monetary terms, so that later trades can be made without moving the dollar figure as far.

Responding to the other side's first number. If it is within reason, counter from your own analysis, not by splitting the difference. If it is extreme, decline to treat it as an opening and ask for the basis. Do not let a lowball or sky-high number pull your counter toward it. Template letters for both situations are in the companion toolkit.

Plaintiff's view. A demand letter that sets out the damages model with exhibits often does more anchoring work than a number delivered at mediation. See Drafting a Demand Letter: A Practical Checklist.

Defendant's view. A defendant who wants to anchor low should explain why, preferably with the plaintiff's own documents. An offer accompanied by a Rule 68 offer of judgment carries a cost consequence if it is reasonable.

Pitfalls. Opening at the reservation price; opening without reasons; opening so far out that the mediator has to spend the morning just getting the other side to stay; negotiating the opening number with yourself before the other side has said anything.

Timing and cost. Decide the opening at least a week before the session, with the client.

Step 8: Write the concession plan

What to do. Write down a planned sequence of moves from the opening to the target, with the reservation price as the floor (plaintiff) or ceiling (defendant). For each round, record the planned number, the condition for making it (for example, "only if they move at least $100,000"), and any non-monetary trade that could be offered instead of money.

Why. Concession patterns communicate. Shrinking increments signal that you are nearing your limit; large late moves signal that earlier positions were bluffs. A written plan helps the client see in advance that a series of moves is expected, which reduces the shock of the first exchange, and it prevents moves made from fatigue.

Elements of a concession plan.

Round Planned number Increment Condition Non-monetary alternatives
Opening $1.4M Delivered with damages model Injunction, public statement
2 $1.15M $250K They move to at least $250K Payment over 12 months
3 $975K $175K They move at least $150K Drop public statement
4 $850K $125K They are above $400K Mutual non-disparagement
5 $775K $75K Bracket or mediator's proposal Reference letter
Floor $650K Reservation price

Trades. List the non-monetary items each side values and what each is worth to the client. Typical items include payment timing, installment security, confidentiality of amount, non-disparagement, neutral references, a license or supply agreement, product phase-out periods, abandonment of trademark applications, dismissal with or without prejudice, and the scope of the release. A term that is cheap for you and valuable to them is worth trading for money.

Brackets and conditional moves. Plan in advance whether you will propose a bracket ("we will move to between X and Y if you move to between A and B") and what midpoint you want to signal. Calculate the midpoint of any bracket you receive immediately; it is the real message.

Pitfalls. Making a second move before the other side responds to the first; making concessions without asking for anything; announcing "this is our final offer" early and then moving (it destroys credibility); failing to update the plan when the other side's pattern is different from what you expected.

Timing and cost. Two to three hours, with the client. The concession plan is privileged work product and should be treated that way.

Step 9: Prepare the mediation statement and materials

What to do. Write the mediation statement (or settlement conference statement) the mediator or court requires, and decide what goes in a shared version and what goes in a confidential version for the neutral only. Assemble the exhibits, damages summary, and any visual aids you will use.

Why. The mediation statement is often the first thing the other side's decision maker reads that was written by you rather than about you. It is advocacy and an anchor. The confidential statement is your chance to brief the mediator on the real issues: what your client needs, what you are worried about, where you need the mediator's help with your own client, and what you know about the other side's pressures.

What the shared statement should do.

  • Tell the story in a few pages, with the best documents attached.
  • Explain the liability theory and confront the main weakness candidly; credibility with the mediator depends on it.
  • Present the damages model with its sources.
  • Identify the rulings and evidence that will matter at trial.
  • State the demand or offer, if you have decided to make it in writing, with its justification.

What the confidential statement should add.

  • The client's interests and any non-monetary terms that matter.
  • Settlement history, including any prior offers and Rule 68 offers.
  • Authority and decision-maker information.
  • Issues on which the mediator's reality-testing of your own client would help.
  • Any constraints on timing, payment, or structure.

Rule 408 and the statement. Label the statement as a confidential settlement communication under Fed. R. Evid. 408 and the applicable mediation rule. Do not include anything whose disclosure would be catastrophic, and do not include privileged material in the shared version; a voluntary disclosure of privileged communications can waive privilege, subject to Fed. R. Evid. 502. For privilege issues generally, see Attorney-Client Privilege and Work Product for Businesses: Upjohn, In-House Counsel, and the Common Interest.

Pitfalls. Submitting the summary judgment brief as the mediation statement; insulting the other side's principal, who will read it; overstating the case in a way the mediator will discount; submitting late.

Timing and cost. Most mediators want statements five to ten days before the session. Drafting takes one to three days of lawyer time.

Step 10: Prepare the client and decide on the opening session

What to do. Meet with the client to walk through the day: the schedule, the likely opening numbers from both sides, the concession plan, the role of the mediator, the possibility of brackets and a mediator's proposal, and what will happen if there is no deal. Decide whether to request a joint opening session and, if so, who will speak and what they will say.

Why. Clients who are not prepared for the opening exchange often react to it as an insult and want to leave. Clients who have seen the concession plan understand that the first numbers are rituals. Clients who know the mediator will test their case are less likely to feel betrayed when it happens.

The joint session decision. Many commercial mediations now skip the joint opening and go straight to caucus. A joint session can help when the other side's decision maker has never heard your client's story, when a short, respectful presentation could change the dynamics, or when the parties have a relationship worth repairing. It can hurt when emotions are raw, when a presentation would harden positions, or when the lawyers would use it to argue the case. If you hold one, keep presentations short, address the decision maker rather than opposing counsel, and avoid personal attacks. Consider whether the client should speak; a sincere statement from a plaintiff or an acknowledgment from a defendant can be worth more than any argument.

Client preparation points.

  • The first offer will be disappointing. That is normal.
  • We have a plan for each round. We will stick to it unless new information justifies a change.
  • The mediator will push you, too. That is the mediator's job.
  • We will take breaks. Do not decide anything while angry or tired.
  • You decide whether to settle. I will tell you what I think, but it is your call.
  • If there is no deal today, there may be one next week. Leaving without a deal is an option.

Pitfalls. Bringing a decision maker who has not read the valuation; letting the client speak in a joint session without preparation; promising the client a result.

Timing and cost. A two-hour preparation session a few days before mediation, plus a short call the evening before.

Step 11: Run the session

What to do. Execute the plan, while watching for information that should change it. Keep a log of every move with the time it was made, so that you and the client can see the pattern.

First caucus. Expect the mediator to ask about the case, the client's interests, and the client's view of the other side. Tell the mediator what may be shared and what may not. Present the opening (if not already made in writing) with its justification.

Information exchange. Share information that builds credibility or creates value: the document that undercuts their theory, the client's need for a payment schedule, the business reason a license term matters. Withhold the reservation price and anything that would weaken your BATNA if disclosed.

Monitoring the dance. Track the other side's increments. Shrinking moves suggest they are near their limit; a large move after a small one may signal that a ruling, a fee concern, or an internal approval has changed their view. Ask the mediator what explains a move.

Managing tactics. When the other side uses pressure (an exploding deadline, a limited-authority claim, a last-minute new demand, a walkout threat), do not react in the moment. Take a break, discuss it with the client, and respond to the substance. Ask whether a deadline is tied to an actual event. Make any concession conditional on a reciprocal move.

Ethics on the day. Statements of value and of the client's settlement intentions are generally treated as posturing, not statements of material fact, under Model Rule 4.1 comment [2] and ABA Formal Opinion 06-439. Misstatements of fact, such as insurance limits or the existence of evidence, are not; in Fire Insurance Exchange v. Bell, 643 N.E.2d 310 (Ind. 1994), the court allowed a fraud claim based on defense counsel's misstatement of policy limits during settlement talks. If authority must be expanded, get it from the client expressly, preferably in writing.

Brackets and mediator's proposals. When the numbers stall, consider a bracket or ask the mediator whether a mediator's proposal would help. Before responding to a mediator's proposal, compare it to the reservation price and BATNA, not to the last number the other side offered.

Pitfalls. Moving twice in a row; letting the client negotiate directly with the other side's principal without preparation; disclosing the reservation price as a "final number" too early; staying past the point of useful work (sometimes a deal needs a week of reflection).

Timing and cost. A full mediation day, often running into the evening. Lawyer time for the session and preparation commonly runs from one to several days, depending on complexity.

Step 12: Close the gap

What to do. When the parties are close but stuck, use gap-closing techniques that let both sides move without losing face.

Techniques.

  • Split the remaining difference. Simple, and often acceptable once the gap is small relative to the case.
  • Add value. Introduce a term that is cheap for one side and valuable to the other: faster payment in exchange for a lower number, a longer phase-out in exchange for a higher one, a license extension, a neutral reference.
  • Restructure payment. Installments, security, or a stipulated judgment held in escrow can make a higher number feasible for a cash-poor defendant and a lower number acceptable for a plaintiff worried about collection.
  • Use contingent terms. Where the parties disagree about a future event (product sales, a pending appeal in another case, a regulatory decision), tie part of the payment to that event.
  • Mediator's proposal. The neutral proposes terms; each side accepts or rejects privately; a deal exists only if both accept, and neither learns the other's answer otherwise.
  • The "best number" exchange. Each side gives the mediator its best number for today. If the numbers overlap, the mediator can close; if not, the mediator knows the remaining gap and can decide whether a proposal will work.
  • Leave the door open. If there is no deal, agree on a follow-up call with the mediator and leave each side's last number on the table for a defined period.

Pitfalls. Splitting the difference too early (it rewards the side that opened more extremely); agreeing on money before agreeing on material non-monetary terms, which invites a nibble; giving a "best number" that is not actually your best and then moving later.

Step 13: Write the term sheet at the table and document the result

What to do. Before anyone leaves, prepare and sign a term sheet that states all material terms, is expressly binding, and anticipates enforcement. Then follow through with the formal agreement, payment, and dismissal.

Why. Oral deals reached at mediation are fragile. Mediation confidentiality rules may make them unprovable, parties' recollections diverge overnight, and buyer's remorse is common. A signed writing avoids most of this. Enforcement must also be planned: under Kokkonen v. Guardian Life Insurance Co. of America, 511 U.S. 375 (1994), a federal court that dismisses a case without retaining jurisdiction or incorporating the settlement terms in its order lacks jurisdiction to enforce the settlement later.

Term sheet contents.

  • Parties, including affiliates to be bound or released.
  • Payment amount, timing, method, and any security, installment, or acceleration terms.
  • Non-monetary obligations, with deadlines.
  • Release scope (mutual or one-way; known and unknown claims; carve-outs).
  • Dismissal form (with or without prejudice) and mechanism, such as a stipulation under Fed. R. Civ. P. 41(a)(1)(A)(ii), and an express request that the court retain jurisdiction to enforce.
  • Fees and costs (each side bears its own, or a specific allocation); in fee-shifting cases, whether fees are waived, which is permissible under Evans v. Jeff D., 475 U.S. 717 (1986), or negotiated separately.
  • Confidentiality and non-disparagement, if any, with appropriate carve-outs.
  • Tax reporting and allocation, if relevant.
  • Lien resolution responsibility.
  • Required approvals (court approval under Rule 23(e) or for a minor, board approval, insurer consent), with deadlines.
  • A statement that the term sheet is binding and enforceable, that a formal agreement will follow, and that the term sheet may be disclosed to a court to enforce it notwithstanding mediation confidentiality.
  • Signatures by the parties (not only counsel) or by representatives whose authority is confirmed.

After the session. Confirm the deal in writing to the mediator and the other side. Notify the court if required by local rule. Draft the formal agreement promptly; see Drafting a Settlement Agreement and Release: A Practical Guide and Settlement Agreement Checklist: A Practical Checklist. Calendar every performance deadline. If a consent judgment is part of the deal, prepare it for entry under Fed. R. Civ. P. 58.

If a fee-shifting plaintiff wants fees later. A private settlement followed by voluntary dismissal may not make the plaintiff a "prevailing party" under Buckhannon Board & Care Home, Inc. v. West Virginia Department of Health & Human Resources, 532 U.S. 598 (2001). Resolve fees in the deal or use a consent decree.

If there is no deal. Send a short follow-up through the mediator within a few days. Consider whether a Rule 68 offer (defendant) or a formal demand (plaintiff) should follow, and whether the next case event will change the other side's view. Many cases settle within weeks of an unsuccessful mediation.

Pitfalls. Leaving with a handshake; signing a term sheet that says "subject to formal agreement" without saying it is binding; forgetting to address retained jurisdiction; failing to calendar performance deadlines.

Timeline and cost overview

Phase Typical timing Typical lawyer time Other costs
Decide whether and when (Step 1) Ongoing; revisit after rulings Hours None
Valuation (Step 2) 2 to 6 weeks before session 1 to 3 days Expert input, if needed
Interests and authority (Steps 3 and 4) 2 to 4 weeks before Half a day to a day Board or insurer approvals
Improving the alternative (Step 5) Months to weeks before Varies widely Motion and expert costs
Format and neutral (Step 6) 4 to 8 weeks before A few hours Mediator fees (often split)
Opening and concession plan (Steps 7 and 8) 1 to 2 weeks before Half a day None
Mediation statement (Step 9) Due 5 to 10 days before 1 to 3 days Exhibit preparation
Client preparation (Step 10) 2 to 5 days before 2 to 4 hours None
Session (Steps 11 to 13) Mediation day 1 day, often long Travel, room, mediator
Follow-through 1 to 4 weeks after 1 to 5 days Approvals, lien resolution

These are general ranges for a mid-sized commercial case. Actual figures depend on complexity, the court, and the parties.

Worked example: Corvane Freight v. Lattice Works

The following example is invented. It follows both sides through the steps.

The case

Corvane Freight, a regional logistics company, licensed a warehouse management system from Lattice Works, a software developer with about $30 million in annual revenue. Corvane paid $1.1 million in license and implementation fees. The go-live failed, Corvane's shipments were delayed for nine weeks, and Corvane sued for breach of contract and gross negligence, claiming $4.8 million in lost profits and remediation costs. The contract caps Lattice's liability at fees paid in the prior 12 months ($640,000) and excludes consequential damages, except for gross negligence or willful misconduct. Lattice's motion to dismiss the gross negligence claim was denied. Cross-motions for summary judgment on the cap are fully briefed. The court has ordered mediation before ruling.

Lattice has a technology errors and omissions policy with a $2 million limit, which erodes with defense costs, and a $100,000 retention. Lattice has spent $380,000 on defense; trial would cost about $600,000 more. Corvane has spent $520,000 and would spend about $700,000 more through trial. Neither the contract nor any statute shifts fees.

Steps 1 through 4: timing, valuation, interests, and numbers

Timing. Both sides decide to mediate before the summary judgment ruling. Each sees risk in the ruling: if the cap is enforced, Corvane's case is worth far less; if the gross negligence question goes to the jury, Lattice faces exposure above its remaining limits.

Corvane's tree. Corvane's lawyers estimate a 55 percent chance the court enforces the cap on summary judgment, in which case they expect to win breach at trial 80 percent of the time and recover the $640,000 cap. In the 45 percent branch where the cap question goes to trial, they estimate a 40 percent chance of a gross negligence finding (with damages of $1.5 million, $2.5 million, or $4 million, weighted 40, 40, and 20 percent), a 45 percent chance of a breach-only verdict at the cap, and a 15 percent chance of a defense verdict. The weighted gross outcome is about $843,000: $282,000 from the cap-enforced branch, $432,000 from gross negligence verdicts, and $130,000 from breach-only verdicts in the trial branch. Subtracting $700,000 in remaining fees leaves a net expected value under $150,000. And the high verdicts are not fully collectible: after another $600,000 of defense costs, Lattice's remaining policy limits would be about $1 million.

Corvane's interests. The CFO's real concern is not the lawsuit but the software. Replacing the system would cost about $900,000 and take eight months. Corvane wants the system supported, wants protection if Lattice fails as a business, and needs a recovery it can report to its lender.

Corvane's numbers. Reservation price: $750,000 in cash plus continued support, or $950,000 in cash with no continuing relationship. Target: $1.4 million plus support terms. Opening: $4.8 million, supported by the damages expert's report, to anchor high while the gross negligence claim remains alive.

Lattice's tree. Lattice's lawyers estimate a 65 percent chance the cap is enforced and a lower chance (25 percent) of a gross negligence finding if it is not. Their weighted exposure is roughly $675,000, plus $600,000 in defense costs that erode the policy. Their WATNA is a $4 million verdict, of which roughly $3 million might exceed remaining limits and fall on Lattice itself.

Lattice's interests. Lattice wants to stay within policy limits, avoid a judgment that its other enterprise customers could read about, and keep Corvane as a reference customer if possible. The carrier wants to close the file before defense costs consume more of the limit.

Lattice's numbers. Authority confirmed in writing: the carrier will pay up to $1.1 million, and Lattice's CEO will contribute the $100,000 retention plus non-cash terms. Reservation price: $1.2 million in cash plus a license extension. Target: $700,000. Opening: $75,000 and a mutual release, a low anchor justified by the cap and the consequential damages exclusion.

The ZOPA on cash alone runs from roughly $750,000 to $1.2 million. Non-cash terms can widen it.

Steps 5 through 8: alternatives, format, opening, and concessions

Improving the alternative. Lattice considers a Rule 68 offer but concludes that, with no fee-shifting provision, the post-offer cost exposure would be small; it decides the offer would add little leverage. Corvane takes the deposition of Lattice's lead engineer, whose emails show that the team knew of a database defect two weeks before go-live. That deposition is the single most important event for the gross negligence branch.

Format. The parties choose a private mediator with software-contract experience who is known for making mediator's proposals. They sign a mediation agreement selecting the governing state's law and stating that a signed term sheet is enforceable and may be disclosed to enforce it.

Concession plans. Corvane plans moves of $1.2 million, $800,000, $500,000, and $300,000, with conditions, and a list of trades: payment timing, support terms, a source-code escrow, a license extension, and confidentiality of the amount. Lattice plans moves of $175,000, $150,000, $125,000, and $100,000, with a license extension and escrow as non-cash items it can offer in place of money.

Steps 9 and 10: statements and client preparation

Corvane's shared mediation statement opens with the engineer's emails and the nine-week delay. Its confidential statement tells the mediator that software continuity matters as much as money. Lattice's shared statement leads with the cap and the consequential damages exclusion. Its confidential statement explains the eroding policy and Lattice's interest in a continuing relationship.

The parties agree to a brief joint session in which each CEO speaks for five minutes, without lawyers arguing. Corvane's CEO describes the nine weeks of missed deliveries. Lattice's CEO says, without admitting liability, that the go-live "was not what either of us wanted." Both sides are then in caucus by 9:15.

Steps 11 and 12: the session and closing the gap

Corvane opens at $4.8 million. Lattice answers at $75,000. At 11:40, Corvane's general counsel wants to leave. Corvane's lawyer reminds her of the concession plan, which predicted this exchange, and the mediator asks her what Corvane will do about its software if the case goes to trial. She stays.

By 3:00 p.m., Corvane is at $2.2 million and Lattice at $450,000. Lattice's lawyer, following the plan, introduces non-cash terms: a two-year license extension at no charge and a source-code escrow. Corvane's CFO values the extension at about $400,000 and the escrow at more than that, because it addresses the risk that Lattice fails. The numbers start to move faster.

At 4:30 p.m., Corvane proposes a bracket of $1.8 million to $1.4 million (midpoint $1.6 million) if Lattice will move to $600,000 to $900,000 (midpoint $750,000). Lattice counters at $700,000 to $1 million (midpoint $850,000). The mediator, now confident the cash gap can be closed with the non-cash terms, makes a mediator's proposal at 5:30 p.m.: $1.15 million, the two-year extension, the escrow, mutual releases, and confidentiality of the amount, with answers due by 6:30. Corvane compares the proposal to its reservation price ($750,000 plus support) and accepts. Lattice compares it to its reservation price ($1.2 million plus extension) and accepts.

Step 13: the term sheet

Before leaving, the lawyers sign a term sheet providing that the carrier will pay $1.05 million and Lattice $100,000, each within 30 days; that Lattice will execute a two-year license and support extension and a source-code escrow with a named escrow agent within 45 days; that the parties exchange mutual releases excluding obligations under the new agreements; that the case will be dismissed with prejudice under Rule 41(a)(1)(A)(ii) with a request that the court retain jurisdiction to enforce the settlement; that each side bears its own fees and costs; that the amount is confidential; and that the term sheet is binding and may be disclosed to the court to enforce it. The carrier's representative and both CEOs sign. Formal papers are executed 12 days later.

Common mistakes

  1. No valuation, or a single-number valuation. Without a range, the negotiator has nothing to compare offers against.
  2. Authority gaps. The person at the table cannot say yes, or the insurer has not been consulted, and the deal stalls at 6 p.m.
  3. Ignoring non-monetary interests. In the example, the license extension and escrow made the deal; neither appeared in either side's pleadings.
  4. Opening without reasons. Numbers without justification are easy to dismiss.
  5. No concession plan. Moves made in response to fatigue, pressure, or the last offer drift toward the other side's target.
  6. An adversarial joint session. Lawyers who use the opening session to argue summary judgment harden positions.
  7. Treating the mediation statement as a brief. It should persuade the other side's decision maker, not the judge.
  8. Leaving without a signed term sheet. Deals evaporate overnight.
  9. Forgetting enforcement. A dismissal without retained jurisdiction sends any breach claim to state court.
  10. Not following up after an impasse. Many cases settle in the days after an unsuccessful mediation.

Frequently asked questions

How precise do the probabilities in a decision tree need to be? Not very. The value of the exercise is in forcing explicit judgments and showing how the result changes with them. Run high and low versions and show the client the range.

Should we disclose our decision tree to the other side or the mediator? Sharing a tree with the mediator in confidence is common and useful. Sharing it with the other side can build credibility but may reveal the reservation price. If you share, share the structure and leave out the numbers that matter most.

What if the client's reservation price is unrealistic? Show the client the valuation and the range, explain the costs of continuing, and, if appropriate, ask the mediator to reality-test. The decision remains the client's under Model Rule 1.2(a); document your advice.

Is it better to make the first offer? Often, yes, if you have done the valuation and can justify the number. The first number anchors. If the other side moves first with an extreme anchor, re-anchor from your own analysis rather than responding to theirs.

When should a defendant serve a Rule 68 offer? When post-offer costs, including fees if the statute defines them as costs, will be significant and the offer is one the plaintiff should rationally accept. Serve it early enough to capture the expensive phase, and draft it precisely.

What if the other side shows up without authority? Raise it with the mediator immediately. If the court ordered attendance by a representative with authority, consider whether to report noncompliance. Make your own moves conditional on their authority.

What should we do if we reach a deal late at night and cannot draft a term sheet? At minimum, write the material terms by hand or in an email, state that they are binding, and have the parties and counsel sign or confirm. A short, complete term sheet is better than a long formal agreement drafted later from memory.

Practical takeaways

  • Build a decision tree, net out fees, interest, taxes, liens, and collection risk, and express the result as a range.
  • Write down the reservation price, target, and opening, with the client's sign-off.
  • Confirm authority in writing, including insurer authority and how to expand it during the session.
  • Improve your alternative before negotiating, and time talks around the events that move value.
  • Choose the format and the mediator deliberately, and settle the confidentiality terms in writing.
  • Open with a justified number, follow a written concession plan, and trade non-monetary terms for money.
  • Prepare the client for the first exchange and for the mediator's reality-testing.
  • Close with a signed, binding term sheet that addresses dismissal and retained jurisdiction.

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

Related Documents

Related Articles

Related Guides

Related Checklists

Related Toolkits