How to Interview Someone for a Book

Key Takeaways

• Prepare what only this person can answer. Anything findable elsewhere is a waste of your one hour with them.
• Ask for scenes, not summaries. "Walk me through that morning" produces usable prose; "how did you feel about it" produces adjectives.
• Silence is the technique. Most of the best material arrives after the pause you did not fill.
• Record with a backup device, always, and take timecoded notes rather than transcribing live.
• The hesitation is the signal. When someone slows down or qualifies something, that is the door.
• Get the consent conversation done properly at the start, in writing, and be specific about what "on the record" means.

Prepare What Only They Can Answer

Preparation for a book interview is not a list of questions. It is a list of things this person uniquely knows.

Do the findable research beforehand. Dates, roles, published accounts, what they have already said in other interviews. Asking a question whose answer is on the first page of a search tells your subject you have not done the work, and the tone of the entire hour changes.

Then write down what only they can answer. What they were thinking in the room. What was said before the meeting. Why they chose the thing everyone assumed was obvious. Who disagreed. What they got wrong. That list is usually shorter than people expect, and it is the whole reason for the interview.

Ten to fifteen prepared questions is plenty for ninety minutes, and you will use six of them. Preparation exists so you can stop thinking about it, not so you can work through it.

Order them by comfort, not by importance. Open with the easy and factual, put the difficult material in the middle once rapport exists, and leave time at the end for the question you would regret not asking.

Bring one specific artifact if you can. A photograph, a document, an old email. Physical or on-screen specifics unlock memory far better than questions do, and this is the single most underused preparation technique.

Ask for Scenes, Not Summaries

The difference between an interview that produces a book chapter and one that produces quotes is almost entirely in the shape of the questions.

"What was your philosophy on hiring?" produces a summary, and a summary is a paraphrase of things they have said before, polished by repetition and empty of detail.

"Walk me through the last time you had to let someone go" produces a scene, with a room, a day, a sentence somebody said, and a decision. Scenes are what prose is made of, and you cannot construct one later from an opinion.

The reliable openers: walk me through, take me back to, what happened next, who else was in the room, what did you say, what did they say.

Push for the specific. "It was difficult" is not usable. "What made it difficult" gets you closer. "What was the moment it became difficult" usually gets you the scene. Three follow-ups is normal and it is where the material lives.

Ask what happened next, repeatedly. It is the least sophisticated question in interviewing and it consistently produces the most usable material, because it moves a person forward through a sequence rather than up into abstraction.

Silence Is the Technique

Most interviewers talk too much, and the fix is uncomfortable and simple.

When they finish an answer, wait. Three or four seconds, which feels enormous. A large share of the best material in any interview arrives in the sentence someone adds because the silence was still there.

Do not fill a pause with your next question. The pause is where they decide whether to tell you the real version.

Do not signal agreement too fast. Nodding and murmuring assent teaches your subject that they have said enough, and they move on from exactly the place you wanted them to stay.

Ask one question at a time. Double-barrelled questions get the easier half answered, and the harder half quietly disappears.

Follow the hesitation. When someone slows down, qualifies, backtracks, or says "well, it is complicated", that is the door. "What is the complicated part?" is often the most valuable sentence in the hour.

Record Properly

Book interviews are long, and losing one is a category of disaster you can eliminate cheaply.

Two recorders, always. A dedicated recorder or a phone, plus a second device. Devices fail, storage fills, someone forgets to press record. The redundancy costs nothing.

Check the level and the storage before you start, not during. Sixty seconds of preparation prevents the failure that cannot be repaired.

Do not transcribe while listening. Take timecoded notes instead: a timestamp and three words whenever something lands. `00:23 the letter from his father`. That gives you a navigable index without dividing your attention, and attention is what produces good follow-up questions.

Note the non-verbal separately. Where they laughed, where they went quiet, what they would not look at. None of it is in the audio and all of it informs how you write the scene.

For remote interviews, record locally as well as in the platform, because platform recordings compress badly and occasionally do not appear.

Consent, On the Record, and Fact-Checking

Get this right at the start, in writing, and be specific, because vagueness here creates problems that surface at the worst moment.

Say what the interview is for, what form it will take, and roughly when it will be published. People agree to different things for a magazine piece and a book.

Define your terms out loud, because they are not universally understood. On the record means quotable with attribution. On background usually means usable without naming them. Off the record means not usable at all, and it only applies if agreed before the thing is said, not after.

Get written consent for a book, even from a friend. An email confirming they understand the material may appear in a published book is enough for most purposes, and its value is that it prevents a disagreement two years later when the book is real.

Agree the fact-check protocol now. Whether they see quotes before publication, and whether that is for accuracy only. Deciding this at the start avoids a fight at the end, and "accuracy only, not approval" is a normal and defensible position if you say it up front.

If they ask you to stop recording, stop, visibly. What follows is often the most important thing they say, and you may not be able to use it. That is the trade, and honouring it is what gets you the second interview.

Right After, and the Second Interview

The hour after an interview is worth more than most people realise.

Write your impressions immediately. What surprised you, what you did not believe, what you want to check, what the piece is actually about now. This takes ten minutes and it is unrecoverable a week later.

Log the standout moments with timecodes while you still remember where they are. Otherwise you will scrub ninety minutes of audio to find one sentence.

Assume there will be a second interview. First interviews produce the official version; second interviews produce the real one, because trust has been established and because your questions are now informed by the material rather than by research.

Send the follow-up questions rather than saving them. People answer specific written questions surprisingly well, and it keeps the relationship warm between sessions.

Where the Material Goes

An interview produces three different things and they need different homes: the audio, the transcript, and what you concluded from it.

The audio and transcript belong wherever your research lives, with the source and date attached so a claim can be traced to a timecode. The conclusions, the connections between interviews and the shape of the argument are a different problem, and they are the part that gets lost.

Storyflow is our product and it should be described honestly here. It does not record, does not transcribe, has no timecodes and is not a reference manager. Use a recorder, a transcription tool such as Descript or Otter, and Zotero or an equivalent for sources. The one job it does is the synthesis stage: laying quotes and themes from several interviews out spatially so the chapters and the thin sections become visible, with AI that reads the whole board. Paid-only during early access, with the Free plan landing before the end of 2026, and anyone a paid member invites to a board joins free now.

The Bottom Line

Book interviews reward preparation and then punish anyone who sticks to it.

Prepare only what this person uniquely knows, because everything else is findable and asking it wastes the hour.

Ask for scenes. Walk me through it, who else was there, what did they say, what happened next. Prose is made of scenes and cannot be made from opinions.

Wait after they finish, and follow the hesitation rather than smoothing past it.

Record twice, take timecoded notes, and write your impressions immediately afterwards while they still exist.

Settle consent and the fact-check protocol at the start, in writing, with the terms defined out loud.

Do not ask what they think. Ask what happened, and then ask what happened next.