For complex, multi-party conflict — competing departments, public policy, joint ventures — "win-win" is a slogan, not a process. The MIT-Harvard Mutual Gains Approach is the four-phase roadmap that finds value no one saw at the outset, then builds an agreement that survives contact with reality.
By Jared Ohman7 min readLast updated June 2026Source: Difficult Conversations, Ch. 8
Value Creation: "inventing without committing." Floating hypothetical "what if" scenarios and bundling options across issues to discover synergies invisible at the outset.
— Difficult Conversations, Ch. 8 (the Mutual Gains Approach)
SHORT ANSWER
The Mutual Gains Approach (MGA) is a four-phase process, developed at MIT and Harvard, for complex multi-stakeholder conflicts — organizational change, public policy, joint ventures. Phase 1, Preparation: align internally, analyze every party's BATNA, and map each stakeholder's underlying interests (not their stated positions). Phase 2, Value Creation: "invent without committing" by floating hypothetical "what if" options and bundling issues to find synergies. Phase 3, Value Distribution: allocate the gains using objective criteria everyone can justify to their constituencies. Phase 4, Follow-Through: imagine what could derail the deal and embed dispute-resolution mechanisms, communication protocols, and metrics directly into the agreement.
GET THE FREE PRINTABLE ↓One page, wallet-card layout. Free. One email below, no spam, unsubscribe in a click.
The problem
It's not a two-person argument. It's four departments with competing needs, or a policy that touches a dozen constituencies, or a partnership where everyone's protecting something different. You call it "win-win" in the kickoff meeting, and then it stalls — because "win-win" is a hope, not a method, and nobody knows the steps.
Complex disputes fail in predictable ways: people negotiate stated positions instead of underlying interests, they treat every idea as a binding offer so no one explores, they divide a fixed pie instead of growing it, and they sign an agreement that collapses the first time reality tests it.
The Mutual Gains Approach is the four-phase process that addresses each of those failure modes in order.
The mechanism
Developed at MIT and Harvard, the Mutual Gains Approach is a procedural roadmap for exactly the multi-stakeholder, multi-issue conflicts where simple negotiation advice breaks down. Four phases:
Phase 1 — Preparation. Get internally aligned, rigorously analyze every party's BATNA (their best alternative if there's no deal), and map each stakeholder's underlying interests, not their stated positions. The interests are where the deal lives; the positions are just the opening costume.
Phase 2 — Value Creation. "Inventing without committing." Float hypothetical "what if" scenarios and bundle options across issues to discover synergies that were invisible at the start. Because one party may value cheaply what another needs dearly, trading across issues can grow the pie before anyone divides it.
Phase 3 — Value Distribution. Allocate the gains using objective criteria and agreed principles, so every representative can justify the outcome to their own constituency. A deal that one side can't defend back home isn't a deal; it's a future grievance.
Phase 4 — Follow-Through. Proactive risk management. Imagine the future scenarios that could derail the agreement, and embed dispute-resolution mechanisms, communication protocols, and performance metrics directly into the document. The deal is engineered to survive the problems you can foresee.
The operating system
Five steps to run a complex dispute toward an agreement that holds.
STEP 01
Map interests, not positions
Before anyone proposes anything, map what each stakeholder actually needs underneath what they're demanding. "We need the bigger budget" is a position; "we need to not be blamed when the project misses" is the interest. The interests are negotiable in ways the positions never are.
Ask "why does that matter to you?" until you hit something that isn't a number. That's the interest.
STEP 02
Know every BATNA, including yours
For each party, understand what happens if there's no deal — their best alternative. A stakeholder with a strong walk-away has leverage; one with a weak alternative has reason to deal. Knowing the BATNAs tells you who needs this agreement and how hard each side can realistically push.
Your own BATNA is your floor: never accept a deal worse than your best no-deal alternative. Knowing theirs tells you their floor.
STEP 03
Invent without committing
Open a phase where ideas are explicitly non-binding. Float "what if we bundled X with Y," explore wild options, trade across issues — with everyone clear that nothing said here is an offer. Separating invention from commitment is what unlocks the synergies; the moment every idea is treated as a promise, creativity dies.
Say it out loud: "This is brainstorming, nothing's committed." That permission is what lets people propose the option that becomes the deal.
STEP 04
Distribute by principle, not by power
When it's time to divide the value, use objective criteria and agreed principles rather than raw leverage. The test: can every representative go back to their constituency and justify the outcome as fair? A distribution that fails that test will unravel the moment people get home and explain it.
Agree on the principle for dividing before you divide. "Let's split by usage" decided in advance beats fighting over the number.
STEP 05
Engineer the follow-through into the deal
Before signing, run the failure scenarios: what could derail this in six months? Then build the answers into the document — how disputes get resolved, who communicates what to whom, what metrics signal trouble early. You're not just closing a deal; you're building one that can absorb the predictable shocks.
Write three "what if it goes wrong" scenarios and embed a mechanism for each. Most agreements die from problems someone could have named in advance.
The printable: the four phases
Print it. Run a complex dispute through the phases in order.
THE MUTUAL GAINS APPROACH
For many stakeholders, many issues.
1 · PREPARATION
Map each party's interests (not positions) and BATNA.
"Why does that matter?" until it's not a number.
2 · VALUE CREATION
Invent without committing. Bundle issues. Find synergies.
Grow the pie before dividing it.
3 · VALUE DISTRIBUTION
Divide by objective criteria everyone can justify back home.
A deal they can't defend will unravel.
4 · FOLLOW-THROUGH
Embed dispute mechanisms, protocols, and metrics in the document.
Engineer it to survive the predictable shocks.
THE HUMAN FREQUENCY · FIND COMMON GROUND
Go deeper
This page is the surface. Each layer below goes further.
MGA is an advanced negotiation method, developed at MIT and Harvard, for complex multi-stakeholder disputes. It's a four-phase procedural roadmap — preparation, value creation, value distribution, and follow-through — that moves beyond simple 'win-win' rhetoric to a repeatable process for finding and dividing value among many parties, then making the agreement durable.
How is MGA different from regular negotiation?
Most negotiation advice assumes two parties and a single issue. MGA is built for many stakeholders and many interlocking issues, where the real opportunity is bundling issues to create value none of the parties saw alone. It also takes follow-through seriously as its own phase — most agreements fail not at signing but afterward.
What does 'inventing without committing' mean?
It's the heart of the value-creation phase: you float hypothetical 'what if' scenarios and explore options together without anyone being bound by them. Separating invention from commitment lets parties brainstorm freely and discover synergies across issues that would be invisible if every idea were treated as a firm offer.
Why is follow-through a separate phase?
Because complex agreements usually fail after they're signed, not during negotiation. The follow-through phase is proactive risk management: you imagine the future scenarios that could derail the deal and build dispute-resolution mechanisms, communication protocols, and performance metrics directly into the final document, so the agreement can survive the problems you can predict.
Did this help?
One signal helps the next page get sharper. Optional note if you want to add detail.
Thanks. I read every one. — Jared
Something went sideways. Try again in a minute.
Continue the wiki
More operating systems most readers of this page also need.