Begin with a focus you can hold in one sentence

A short oral history project is not a life story. It is a slice of a life, and the sharper the slice, the more likely you are to finish it. Before you contact anyone, write your focus as a single sentence and pin it above your desk. Something like: "I am recording six women who worked at the same Leicester hosiery factory between 1965 and 1980." Or: "I am recording four residents of a Norfolk village about the winter of 1963." If you cannot fit your idea into one sentence, it is not yet a project.

That sentence decides everything downstream — who you approach, how many sessions you book, what you ask, and what you do with the material afterwards. It also gives you a clean answer when someone offers you a wonderful story that sits just outside your frame. You can say, "That belongs in a different project," and mean it warmly.

A workable scope for a first project is three to six narrators, each recorded once or twice, for forty to sixty minutes a session. Ten narrators with one session each is harder than it sounds: the editing will swamp you before you have listened properly to anyone.

Sort out consent before you switch anything on

Consent is not paperwork that gets in the way of the good bit. It is the thing that lets your narrator speak freely, because they know exactly where their words are going.

• Explain the project in writing, in plain language — its purpose, who is running it, how long it will take, and what will happen to the recordings.
• Separate permission to record from permission to publish. Someone may be happy for their voice to sit in a local archive but not on a website. Both answers are legitimate.
• Cover archiving and copyright explicitly. Say where the files will live, who can listen, and for how long. Give a named contact for questions.
• Allow withdrawal. Offer a clear window — thirty days, say — in which a narrator can ask for a recording to be deleted or restricted.
• Get it signed in person, before recording, with two copies so your narrator keeps one. If someone cannot sign, read the form aloud and note that you did.

Also ask about photographs, letters and objects. People often remember far more with an object in their hands, but you need permission to photograph or scan it.

Test your equipment — and the room — before the day

The most common failure in small oral history projects is not a bad interview. It is a recording that cannot be heard. Do a full test a week in advance, not ten minutes before your narrator arrives.

• Record on two devices if you can. A primary recorder and a phone as backup costs you nothing and has saved many a project.
• Use an external microphone rather than the built-in one. A small clip-on microphone placed a hand's width below the collarbone gives you a close, clear voice and far less room echo.
• Put phones into airplane mode so a call cannot interrupt mid-sentence, and set devices to record in a plain, uncompressed format if they offer it.
• Check levels before you begin and aim for peaks around two-thirds of the meter. Never let the signal hit the top.
• Listen through headphones during the test. You will hear hums and buzzes that your ears ignore in the room but cannot ignore on playback.
• Choose the room deliberately. Soft furnishings, curtains and rugs absorb echo. Kitchens, conservatories and rooms with a fridge on the other side of the wall do not.
• Record thirty seconds of silence at the start and end. That "room tone" is useful later if you or an editor needs to tidy the audio.

Finally, note the date, place, names and project title at the top of every recording. In two years' time you will not remember which file is which, and an unnamed recording is an unusable one.

Ask questions that open memory rather than close it

Good oral history questions are invitiations, not interrogations. Aim for open beginnings — "Tell me about…", "What was it like…", "How did that come about…" — and avoid anything that can be answered with a single word.

Rather than asking, "Did you enjoy working there?", try, "What do you remember about a typical morning there?" Instead of, "Were you frightened?", try, "What was going through your mind at the time?" Specificity helps too: "What could you see from your desk?" or "What was the first thing you did when you got home?" often unlocks a whole scene.

Prepare eight to ten prompts but expect to use only four or five. Keep a small set of follow-ups ready — "What happened next?", "Who else was there?", "Can you describe that for me?" — and let them carry the session. Resist correcting dates or details; a misremembered year often sits next to a perfectly remembered feeling, and the feeling is the point. When your narrator stops, count to five before speaking. The most valuable sentences frequently arrive in that silence.

Structure the session around your narrator's energy

Start with easy, factual ground: name, where they were born, what their working days looked like. This settles nerves and gives you usable context. Move into the heart of your topic once the voice has relaxed, and leave the most difficult or emotional material for the final third, when trust is highest.

Watch the clock and watch the person. Offer a break at the halfway mark. If they tire, stop — a warm, complete forty minutes beats a strained ninety. Say thank you properly, and ask whether there is anything they expected you to ask and you did not. That single question has produced some of the best material in many a project.

Look after the recordings once the recorder is off

Back the files up the same day, in two places, and label them consistently. Transcribe at least the passages you intend to use, marking timecodes so you can find them again quickly. Send your narrator a short note afterwards confirming what happens next, and keep your word about the thirty-day withdrawal window. A short project done carefully and finished is worth far more than an ambitious one that stalls — and the people who trusted you with their memories will notice the difference.