Harvard Negotiation Principle: Interests Over Positions

Two people can argue for an hour and only repeat numbers. "I need 120." "We can do 95." Each restatement feels like firmness. It is often just a position, with the real constraint (rent, headcount, risk, pride) never named.

When the person and the problem fuse, every offer sounds like an insult. The other side defends face instead of exploring trades. The meeting ends with the same two numbers and more heat.

Principled negotiation is a way to separate those layers. It does not make the other party collaborative. It asks whether you are still bargaining only over the label they put on the table.

This page's method is interest-based negotiation as taught here.

A position is the demanded term. An interest is why that term matters. "I need 120" is a position. "I cannot accept a cut that reopens last year's retention risk" is closer to an interest.

The working split:

*Getting to Yes* (Fisher, Ury, Patton) also teaches inventing options for mutual gain, using objective criteria, and knowing your BATNA (walkaway). Those ideas are part of the same family. They are not this page's full operating list, and they are not a generic last consulting step.

This method is not a numeric overlap map (ZOPA). It is not a choice among five conflict modes (Thomas-Kilmann). It is not "be nice."

A useful pass names interests from the case. It does not invent the other party's inner life.

1. Separate people from the problem

Write the issue without a villain. "The handoff missed Tuesday" is substance. "You never care" is a people attack.

2. Translate positions into interests

For each demand, ask what it protects. If you do not know, say you are guessing and check. Do not treat your guess as their confession.

3. Keep the climate usable

Name emotion without using it as evidence. Appreciation should point to a specific move they already made. Frame requests as what you observed and what you need next.

4. Refuse the react cycle

If they attack, return to the problem. Matching tone feels strong and usually restarts the same two positions.

If they will not discuss interests at all, stop pretending the method is running. You may still need a walkaway, a range test, or a different conflict mode.

When another lens fits better, or when you need a complementary view, these frameworks do different jobs. They are not interchangeable labels for the same question.

The Harvard Negotiation Principle is the lens for unsticking positions and separating people from the problem. Other methods help when the question is overlap, conflict mode, or walkaway.

• A salary or rate conversation that has become two numbers. Both sides can recite the figure. Neither has said what the figure is doing (cost of living, budget cap, internal equity).
• A contract where blame and terms are tangled. Delay is discussed as character, not as a delivery constraint.
• An internal dispute that must leave a working relationship. The parties still have to share a roadmap after the meeting.

Example: A salary talk stuck on 95 versus 120

A product designer is at 95. They asked for 120. The manager has a band ceiling at 108. Both keep restating the number.

A position-only version is a contest. An interest-based version might surface:

• Designer interests: rent, internal equity with a peer hired last month, a signal that the scope increase is recognized.
• Manager interests: band integrity, a promotion cycle in six months, not setting a one-off that Finance will unwind.

Implication: Trades might include a title timeline, a scope letter, or a mid-cycle review. None of that is guaranteed. Inventing a "win-win 120" the band cannot pay would be a method error. If the designer will only discuss 120, you are still on positions.

Example: A vendor delay blamed as bad faith

A startup wants a penalty clause. The vendor hears an accusation. Emails get shorter and colder.

• People vs problem: The missed milestone is the problem. "You sandbagged us" is a people move.
• Interests: The startup needs a launch date it can sell. The vendor needs not to accept a penalty that assumes a dependency the startup still owns (content, API keys).
• Constructive frame: "Launch is 12 October. The remaining blocker is the key delivery on our side plus your load test window."
• Escape react: Do not answer a sarcastic line with a sarcastic line. Return to the milestone and the dependency.

Implication: Appreciation might name the extra weekend the vendor already worked. It is not a substitute for the date. ZOPA would ask whether any penalty range overlaps. This method asked what each position was protecting.

What this method can help with

• Splitting people issues from the substance
• Translating a demanded term into an interest
• Reducing a react-and-counter loop in the room
• Naming when collaboration is not actually happening

What this method cannot replace

• ZOPA. The Zone of Possible Agreement tests whether acceptable ranges overlap. Interest talk does not prove an overlap exists.
• Thomas-Kilmann. Five conflict modes (compete, collaborate, compromise, avoid, accommodate) describe style. This page is interest-based problem-solving, not a mode inventory.
• BATNA as the primary method. Your walkaway is a *Getting to Yes* idea. This page's stages do not run that calculation for you.
• Option generation and objective criteria as a full process. The book teaches them. This production structure emphasizes interests and communication practices.
• A guarantee of mutual gain. One party can refuse. The method cannot force a deal.

Honest scope: this is principled negotiation as taught here. It should not be sold as win-win.