How to have a difficult conversation: the one-page method

Stone, Patton and Heen's finding in Difficult Conversations is that every hard conversation is three conversations at once. There is the "what happened" conversation, about facts and blame; the feelings conversation, about what each person is carrying into the room; and the identity conversation, about what this says about who you are: competent or not, a good person or not. Most of them go wrong because we argue the first while the other two do the real damage, unspoken.

So before anything else, write down which of the three is really at stake. The missed deadline is rarely about the deadline; it is about whether you can rely on this person, and about whether they feel trusted. Naming that for yourself changes what you say first, and it stops you winning an argument about facts while losing the person.

A quick way to find it: finish the sentence "What I am actually afraid of in this conversation is…". Whatever comes after is the conversation you need to prepare for.

Fisher and Ury's most durable idea is the BATNA, the best alternative to a negotiated agreement. In a hard conversation it is the question you are least likely to have asked: if this goes nowhere, what will I actually do? Escalate, live with it, leave, change the arrangement on my side? Knowing the answer is what lets you stay calm when the other person pushes, because you are no longer negotiating from fear of the conversation failing.

Write two things: the outcome you want, in one sentence, and what you will do without it. If the second is blank, the conversation is not a negotiation yet; it is a plea, and the other person will feel that. If the second is strong, you can afford to be generous in the room.

Be honest about the want. "I want them to admit they were wrong" is a common answer and a bad goal; nobody grants it, and pursuing it turns every step below into a trap. "I want a plan that stops this happening again" is one they can say yes to.

There is your story, in which you are reasonable and they are difficult, and their story, in which the reverse is true. Neither is an opening line. The book's advice is to begin from the third story: the version a neutral observer would give, which describes the difference between the two without deciding who is right. "We seem to see the deadline differently, and I would like to understand how you see it" is a third-story opening. "You keep missing deadlines" is not.

Write the opening sentence out and read it aloud. If a fair-minded stranger could not hear it without taking a side, rewrite it. The first ten seconds decide whether the other person listens or starts defending, and you only get them once.

Then ask before you assert. A question at the start is not weakness; it is the cheapest way to learn what you have wrong, and there is always something.

Chris Voss's accusation audit is the counterintuitive step: list every negative thing the other person might think of you in this conversation, and say the strongest ones first, out loud. "You probably think I am only raising this because of my own numbers" or "This is going to sound like I am blaming you." It feels like handing them ammunition. In practice it does the opposite: named in advance, the accusation loses its charge, and what the person hears is that you already understand how this looks from their chair.

Do not soften it into a compliment or a disclaimer; the audit works because it is blunt about the worst case. Write three, pick the two that matter, and open with them right after the third-story sentence.

The audit also protects you. Once the worst version is on the table, the other person cannot surprise you with it halfway through, and you stop spending the conversation bracing for a blow that has already landed.

When the emotion arrives, and it will, label it rather than argue with it. "It sounds like this feels unfair" or "It seems like you are frustrated that this came up now." Voss's labels always start with "it seems" or "it sounds like", never "I think" or "you are", because a tentative label invites correction instead of a fight. Then stop talking. The silence after a label is where the other person tells you what is actually going on.

Labeling is not agreeing. You can name their frustration accurately and still hold your position; in fact you can only hold it well once the frustration has been acknowledged, because until then every sentence of yours is being heard through it.

The same works on your own feelings, said plainly: "I notice I am getting defensive." It is disarming, and it buys you a breath.

Most of us listen in order to reply. The method asks for something else: listen until you can summarize their position so accurately that they say "that's right", not "you're right", which is what people say to make you stop. Mirror their last few words to keep them talking, summarize what you heard, and check it. Voss treats "that's right" as the turning point in any negotiation, because it is the moment the other person feels understood and stops defending.

This is the step that takes longest and feels least productive, and it is the one that decides the outcome. Budget for it: assume the first half of the conversation is theirs.

If you catch yourself preparing your reply while they talk, that is the signal you have stopped listening. Put the reply down; it will still be there. Their next sentence often makes it unnecessary.

Kahneman's loss aversion explains much of what looks like stubbornness. A loss hurts roughly twice as much as an equal gain pleases, so a proposal framed as taking something away meets far more resistance than the same proposal framed as protecting something. "You would lose the Friday flexibility" and "we could keep the Friday flexibility if…" describe the same arrangement and land completely differently.

Before the conversation, look at what you are asking for and write down what it costs the other person, in their terms. Then find the framing in which your proposal prevents a loss they care about, rather than one that imposes it. This is not spin; it is describing the same facts in the order they will be heard.

You cannot run any of the steps above while flooded. When your heart rate climbs, the listening stops and the rehearsed argument takes over, and the other person can tell. So plan for your own reaction, not only theirs: know the line that sets you off, decide now what you will do when it comes, and give yourself a physical cue: a slower breath, a glass of water, a note to look at.

It is also allowed to pause. "I want to get this right, and I need a minute" is a stronger sentence than anything said in anger, and a conversation resumed after ten minutes usually goes better than one pushed through. Write the pause sentence on the page so you do not have to find it under pressure.

When it is time to move toward a resolution, resist the urge to present your solution. Ask calibrated questions instead: open questions that start with "how" or "what" and hand the problem to the other person. "How can we make sure this does not happen again?" "What would need to be true for this to work for you?" "How am I supposed to do that?" They cannot be answered with a no, they keep you out of the position of the accuser, and the answers often contain a better plan than yours.

Prepare three and write them down; under stress you will not invent them. Avoid "why", which sounds like an accusation in any language, and avoid questions that are really statements with a question mark.

A good calibrated question is one you genuinely do not know the answer to. If you already know what you want them to say, it is a leading question, and they will hear it as one.

A hard conversation that ends in a warm feeling and no agreement will have to be had again. Before you leave the room, say what you understood was agreed, who does what, and by when, and ask whether that matches what they heard. If you could not agree, agree on the criterion you will use to decide, or on when you will talk again. Fisher and Ury's objective criteria belong here: a standard both of you would accept from a stranger.

Then the page itself, filled in the night before: the conversation that is really at stake, what you want and your alternative, the third-story opening, two accusations to audit, the labels you expect to need, the loss you are protecting them from, your pause sentence, three calibrated questions, and the next step you will propose. It takes twenty minutes, and it is the difference between a conversation you had and one you led.