The last number, twice
It's six in the evening. The mediation started at nine. The parties — a manufacturer and a former distributor in a contract dispute — have moved a long way. The gap has narrowed from several million to four hundred thousand.
Then it stops.
The distributor's last offer was the same as its previous one. The manufacturer's response was the same as its previous one. Both lawyers have told the mediator, privately, that this is their final position. One of them has mentioned a flight at nine.
The mediator can feel the day ending without a deal. And the gap, measured against what both sides have already spent on the dispute, is not large.
An impasse is a signal, not a wall
Breaking impasse in mediation requires the mediator to recognize that the stated position is rarely the real obstacle. When parties stop moving late in a negotiation, the reason is usually one of a small number of things.
Principle. A party has decided that paying — or accepting — more would be an admission it can't make.
Face. Someone has committed publicly, to a board or a client or themselves, that they won't go past a number.
Risk perception. One side sees the litigation risk very differently and hasn't been confronted with that difference.
Authority. The person in the room has reached the limit of what they're authorized to do.
Something unsaid. A non-monetary issue — an apology, a reference, confidentiality, timing, tax treatment, an ongoing relationship — that neither side has put on the table.
Fatigue. It's six o'clock and people are tired.
Each has a different remedy. A mediator who applies a money technique to a principle problem will fail.
Diagnose before you intervene
Ask each side what's in the way. "Help me understand — what makes four hundred thousand the line?" The answer often reveals whether the issue is money at all.
Ask what would need to be true. "If you were going to move from here, what would you need to see?" This frequently surfaces a non-monetary term or a concern about how the deal will be perceived.
Check authority privately. "Is this your limit, or the limit of what you can agree today?" If authority is the problem, the solution may be a call, not a concession.
Name fatigue honestly. Sometimes the most useful move is a break, food or an agreement to reconvene — not another round.
Tools for different impasses
For money gaps that are genuinely about money: Bracketing — each side proposes a range rather than a point — can restart movement without either side appearing to concede. Some mediators use conditional offers: "If they came to X, would you come to Y?"
For principle and face: Reframe what the payment represents. Structured payments, different characterizations, confidentiality terms or mutual releases can let a party move without appearing to abandon its position.
For non-monetary issues: Put them on the table explicitly. A reference letter, a statement, a transition period, a joint announcement — these often close gaps that money alone won't.
For risk-perception gaps: Return to reality testing — covered in the previous post in this series — with specific questions about the path to trial and its cost.
The mediator's proposal. In some mediations, where both parties agree, the mediator proposes a specific number or terms, and each side confidentially tells the mediator yes or no. If both say yes, there's a deal; if either says no, neither learns the other's answer. It's a powerful tool, usually reserved for the end, and it should be used only with the parties' agreement.
When it doesn't close today
Not every impasse breaks in the room. Mediations that end without agreement often settle later, particularly when the mediator stays involved. Agree on next steps: a follow-up call, a deadline for a response, information to be exchanged. A mediation that ends with a plan is not a failure.
Four ways it goes wrong
The money-mover, who treats every impasse as a gap to be split and ignores principle, face or non-monetary issues.
The mediator who pushes past fatigue, running round after round with exhausted parties.
The premature proposer, who makes a mediator's proposal too early or without the parties' agreement.
The quitter, who lets the day end without a plan for continuing.
Why this isn't trained
Impasse techniques are taught as a toolkit. Mediators learn bracketing, mediator's proposals and conditional offers — but less often how to diagnose which one fits.
Late-day dynamics are hard to reproduce. Fatigue, face, and authority limits don't emerge in a thirty-minute role play.
Non-monetary interests are easy to miss. Parties rarely volunteer them. Mediators have to ask.
And practice parties are too flexible. In role plays, parties usually move when the mediator tries something. Real ones sometimes don't — which is exactly what the mediator needs to practice.
What impasse simulation can rehearse
A ten-minute simulation can put the mediator into the late stage of a negotiation with two parties who have stopped moving for different reasons — one on principle, one on authority, or one waiting for a non-monetary term — so the mediator practices diagnosing before intervening. The AI agents in Foretell AI play each party and counsel consistently; the dispute, settlement ranges and mediation agreement stay with the program.
Four versions to build:
- The principle holdout, for whom paying more would feel like an admission.
- The authority limit, where the representative can’t go further without a call.
- The unsaid term, where a reference letter or public statement would close the gap.
- The fatigued room, where the right move is a break or an agreed plan to reconvene.
Design caution. The use of mediator's proposals, conditional offers and other impasse techniques depends on the parties' agreement, the mediator's standards and program rules, which vary. Modules must use the program's own rules and settlement parameters. Nothing here is legal advice.
Designing the module
Ten minutes, scored against an impasse-management rubric.
Pass one — diagnosis. Did the mediator ask what was blocking movement before proposing a technique?
Pass two — fit. Did the intervention match the cause — principle, face, authority, risk, non-monetary interest or fatigue?
Pass three — continuity. If agreement wasn't reached, was a concrete next step agreed?
Rubric on observable behavior: Were diagnostic questions asked? Was authority checked? Were non-monetary interests explored? Was the chosen technique appropriate to the cause? Was a mediator's proposal used only with agreement? Was fatigue recognized? Were next steps agreed if the session ended?
Diagnosis-before-intervention is the measure. The right technique applied to the wrong cause doesn't work — and the only way to know the cause is to ask.
For mediators, ADR providers and law firms
Impasse is where mediations succeed or fail. Skill at the end of the day determines outcomes.
Diagnosis generalizes. The same questions work in commercial, employment, family and community disputes.
Advocates benefit. Lawyers who understand why they're stuck can propose non-monetary terms or authority solutions themselves.
And it's rarely observed. Late-day private sessions aren't visible to trainees. Simulation makes them practicable.
For law schools, negotiation and ADR courses can add realistic late-stage impasse to exercises that often end with agreement.
Frequently asked questions
How do mediators break impasse? By diagnosing what's actually blocking agreement — principle, face, authority, risk perception, an unaddressed non-monetary issue or fatigue — and choosing a technique that fits that cause.
What is a mediator's proposal? A technique, used with the parties' agreement, in which the mediator proposes specific terms and each party confidentially accepts or rejects them. If both accept, there's a deal; if not, neither learns the other's answer.
What is bracketing in mediation? A technique in which each side proposes a range rather than a single number, signaling where it might move without committing to a specific figure.
What should happen if mediation ends without a settlement? Agree on next steps — a follow-up call, a response deadline, information to exchange. Many disputes settle after mediation, especially when the mediator stays involved.
The short version
Four hundred thousand apart at six o'clock, and both lawyers say it's final. Maybe. Or maybe it's principle, face, authority, a reference letter nobody mentioned, or the flight at nine.
Ask what's in the way. Ask what would need to be true. Check authority. Offer a break if people are exhausted. Then use the tool that fits — bracketing, a reframe, a non-monetary term, a mediator's proposal if both agree. And if it doesn't close tonight, leave with a plan.
Most impasses aren't about the last dollar. Finding out what they're about is the whole technique.
Foretell AI lets mediators and ADR programs build late-stage negotiation simulations — including principle holdouts, authority limits and hidden non-monetary interests like the ones above — with configurable AI parties and counsel, recordings and rubric-based evaluation. If your mediators only meet real impasse in live sessions, we're happy to walk through how other programs have structured it.