The Never Split The Difference Files · #07 · ~5 min read
Give Up Control to Get It: Calibrated Questions
Based on Chris Voss's Never Split the Difference, Chapter 7, "Create the Illusion of Control."
This chapter contains the technique Voss is arguably best known for, and it's also the story from the book's own opening: an open-ended question, asked under real pressure, that turned a hostile demand into a shared logistical problem.
What a calibrated question is
A calibrated question is an open-ended question, almost always starting with "how" or "what," that has no fixed answer and can't be answered with a simple yes or no. "How am I supposed to do that?" "What about this is important to you?" "How can we solve this together?" The question hands the other person the feeling of being in control — they're the one with the answer, the one being asked for their expertise — while actually guiding the conversation toward the constraints you need surfaced.
Voss deliberately avoids "why" questions, which tend to sound accusatory even when they're not meant that way ("why do you need it by Friday?" can land as a challenge). "What" and "how" questions ask for information without putting the other person on the defensive.
Why the illusion of control works
Most resistance in a negotiation comes from a felt loss of control — the same reactance mechanism this whole content library keeps returning to. A calibrated question is, in a sense, the single cleanest antidote available: instead of you asserting your position and inviting pushback, you ask them to solve a problem, which requires them to think from inside your constraints without ever feeling told what to do. They end up doing the work of finding the answer, and answers people find themselves land with more weight than answers handed to them.
How this plays out honestly
- Ask "how," not "why." "How can I make this work on my end?" invites problem-solving. "Why do you need this?" invites defensiveness, even when it's a fair question.
- Use it when you're stuck, not when you already know the answer. A calibrated question asked as genuine information-gathering works. The same question asked as a rhetorical trap, when you've already decided what you want them to say, reads as manipulative the moment they sense it (the same failure mode covered across this whole series).
- Let their answer actually change your next move. The entire value of the technique evaporates if you ask "how am I supposed to do that?" and then ignore whatever they say in response. The question has to be a real invitation to help solve the problem, not a rhetorical stall.
Where this sits in the series
This is, in a real sense, the most complete version yet of "match, don't manipulate" — the site's own governing ethic. A calibrated question literally cannot manipulate anyone into anything on its own; it just asks a genuine question and waits. Whatever happens next depends entirely on what's actually true about the situation. That's the cleanest example in either series of a technique whose honest and dishonest versions are hardest to tell apart on the surface, and yet the dishonest version barely exists — because there's very little to fake here beyond genuinely wanting the other person's answer.
The honest caveat
Calibrated questions are slow. They require actually being willing to sit with an open answer instead of steering toward a predetermined one, which can feel inefficient under time pressure. And they don't work as a substitute for having done the earlier work in this series — mirroring, labeling, building real rapport — because a calibrated question asked cold, with no established trust, can still land as an interrogation rather than an invitation.
Previously: bending reality honestly. Next: the handshake isn't the deal.