Transcription for NGOs Doing Field Research

Field research produces the hardest recordings and the most important ones: interviews conducted in a participant's own language, on a phone, in a compound, a market or a clinic, by staff whose job is the conversation rather than the audio.

Record for the conditions you are actually in

Field audio fails in predictable ways, and a short briefing for field staff fixes most of it.

Get close. A phone on the mat between two people beats a phone in a pocket by a wide margin. It is the highest-return instruction you can give.

Silence the phone first. Notification tones land directly on the microphone and always mid-answer.

Say the metadata aloud. Date, location, interviewer, participant reference. Eight seconds at the top of every recording, and every file afterwards identifies itself even if the filename does not.

Do not re-record. If a note comes in by WhatsApp, forward the original file rather than playing and re-capturing it.

Language is the whole job

Fieldwork happens in the language people actually speak. Reporting happens in English. That gap is where most of the effort goes, and it is the part usually done by whoever on the team is bilingual, in the evening, unpaid.

Working from a transcript makes it tractable: the original stays alongside the translation rather than being replaced by it, which also means a reviewer can check the rendering later.

Set the language honestly for each recording. Where support is partial we say so, on the languages page, and there is a longer piece on why African language transcription matters that covers the underlying difficulty.

Focus groups in the field

Community sessions are the most valuable and the hardest to transcribe: several voices, an open space, and the moments of real insight are the moments people talk over each other.

A facilitator who names people as they speak does more for the eventual transcript than any equipment. More on the limits in speaker diarization.

Consent, and what happens to the recording

Participants should know they are recorded and what happens to the audio afterwards. Our privacy policy sets out what is stored and how to delete it, and recordings are never used to train any model — worth being able to state plainly when you are asking a community for consent.

From transcript to report

The reason to transcribe at all is usually a donor report or a programme decision. Once the interviews are text you can search across a whole round of fieldwork, count how often something came up, and quote participants accurately rather than from a fieldworker's summary.

That last point is the quiet one: quoting from a transcript rather than a recollection is the difference between evidence and impression.

For a programme team

Where several field officers share one pool of hours on a single invoice, transcription becomes a programme cost rather than an individual one. See the pricing section.

A free account includes transcription minutes each month with no card. Sign up and run one real field recording through before you standardise on anything.