So You Have To Conduct An Interview
You got assigned a piece, your editor told you to interview someone, and now you are sitting there wondering what the hell you are actually supposed to do with a recorder and a person who has things to say. It is less complicated than most people make it out to be and about half as exciting too. The whole process is basically just getting someone to tell you something useful, then figuring out what they actually meant by it when you play it back. Start by reading everything the person has published, or said publicly, before you ever book the call. I once spent forty-five minutes asking a subject matter expert about their own book because I had not bothered to check whether they had written one. They were not amused. You do not need to read all of it. A couple of paragraphs of their work and maybe three recent articles where they discussed the topic will give you enough to ask questions that do not sound like you pulled them from a template. The actual question list should have about eight to twelve items. More than that and you will be reading instead of listening. Fewer than that and you are hoping for a miracle. Each question should target one specific thing you need for the piece. If a question could be useful for two different reasons, keep it but make sure it serves at least one of them clearly.
I usually structure the interview around three buckets: context, concrete detail, and friction. Context means I am getting the person to explain how something works or why it matters. Concrete detail means I am asking for numbers, timelines, names, actual examples. Friction means I am asking about what went wrong, what people disagree on, what the messy part is. Beginners skip friction. They end up with articles that read like press releases. The person you are interviewing will appreciate being asked about the hard parts. Record the conversation. Use something that just works. Zoom, a voice memo app on your phone, or a cheap USB recorder. My go to has been the Voice Memos app on my iPhone for years. It is annoying, but it has never failed me in transit and the files are small enough to move quickly. One time I used an online meeting platform that encrypted the audio in a way that made it unreadable by every transcription tool I had access to. That cost me a full afternoon of trying different converters. I stopped relying on cloud meeting platforms for interviews after that. Phone or dedicated recorder now. Send the person the list ahead of time if you want thoughtful answers. If you send it, do not send the whole list verbatim. Send three or four of the questions so they know the general direction. If you send everything, people will write polished answers in their head and the interview becomes a performance instead of a conversation. I found this out the hard way on a piece about supply chain logistics. The person I interviewed had prepped written responses for every question. The result was dry and full of corporate phrasing. I called him back a week later and asked the same questions in a different order without warning. He stopped performing and started thinking. That second version had everything I needed.
During the interview, let silence do work. Most people fill dead air because it feels uncomfortable. If you pause for four or five seconds after they finish a sentence, they will often keep talking and say something useful. Do not interrupt. Do not nod loudly into the microphone. Just wait. The recording will capture it and you will have more material than you expect. Take brief notes while they talk. Not full sentences. Three or four keywords per answer so you remember which thread led where. You do not need a transcript of everything they said to write the article, but you do need to remember the shape of the conversation when you go back to it. I use a simple numbered list matching my questions. Question one gets three dots and a word or two. That is enough. When the interview ends, ask one follow up you thought of during the conversation but did not get to. Then ask them to point you at anyone else you should talk to. This is not a courtesy thing. It is how you find the people who actually know what is going on beneath the surface. Editors like it too because it shows you are not just collecting quotes from the first person you found.
Get the Full Details

After the interview, transcribe the audio. If you transcribe it yourself you will learn exactly what they said and you will catch the moments where they changed their mind mid sentence. That is valuable. If you use a tool, do it quickly and still listen to the full recording. Tools miss tone, they flatten nuance, and they sometimes confuse names. I have seen AI transcription swap a proper name for a completely different word and the writer ran with it. The piece was wrong and took three days to fix after publication. Listen to the file even if you use automated transcription. It takes longer upfront and saves you from looking careless later. Then go back and pull quotes. Not the longest quote you can find. The clearest one. A quote should do one job: make a point faster than you could write it yourself. If the quote is two sentences and you can say the same thing in one sentence, use yours and drop the quote. Readers do not need to hear the person speak in their head. They need the information. Structure the article around the answers, not the questions. I see a lot of drafts that follow the interview order exactly. That reads like a Q and A transcript disguised as prose. Rearrange everything so the piece has its own arc. Put the most interesting thing early. Put the hardest trade off in the middle so the reader understands why the topic is not simple. End with what is unresolved or what comes next. Leave the reader with something to think about instead of a tidy bow.
One thing that trips people up constantly is confirmation bias in the interview itself. You go in wanting a certain angle and you steer the person toward it without noticing. They will give you what you are asking for. If you want a story about failure, you will ask about failures and they will hand you a failure story, even if their experience is mostly about success with small bumps. The fix is simple but hard to do consistently. Write down your assumed angle before the interview. Then ask at least two questions that challenge that angle. If the person pushes back, follow the push back. That is where the article lives. Another practical detail that matters more than people admit is the environment. If the person is on a noisy call or in a place where their dog is barking or their kid is yelling, the audio quality drops and the transcriber, or your brain, will miss important words. I now ask interviewees beforehand if they have a quiet space and offer to do a short voice call instead of a video call if that helps. Video is not necessary for most writing interviews. Audio only is faster to set up and usually clearer. I lost a whole good quote last year because the person was on speakerphone in a coffee shop and the background noise made a key phrase unintelligible. That was on me for not checking conditions first. There is also the ethics side. If you are quoting someone directly, make sure you understand how they want to be represented. A small clarification call after the draft is done can prevent problems later. I had a source get upset because I paraphrased something they said in a way that shifted the meaning slightly. It was unintentional, but the fix required an awkward email chain and a retracted sentence. A quick call after the article is drafted and before you publish is worth fifteen minutes. It does not mean you let them edit the piece. It means you confirm that the quote matches what they said and that the context is fair.
If you are interviewing someone for a longform profile rather than a standard news article, the process changes a bit. You need more background reading, more casual conversation early on to build comfort, and you are looking for narrative threads instead of just facts. The same rules apply, but you spend more time listening to how the person thinks rather than just what they know. A profile interview is about rhythm and voice. A news interview is about information. Do not mix them up or you will end up with a piece that is too slow for news or too shallow for a profile. Finally, keep a simple log of every interview you do. Date, person, topic, what went well, what you would do differently, and the link to the audio file. You will forget details and you will need them when you are researching a follow up piece six months later. I use a basic spreadsheet with columns for name, organization, date, key takeaway, and file location. It takes ten seconds to update after each interview and it has saved me more times than I can count.
