For research teams
How Long Does It Take to Transcribe an Interview?
Far longer than the interview. Careful verbatim transcription means typing, rewinding, and re-listening — researchers routinely spend several hours of work for every hour of audio, and correcting an AI-generated draft is often no faster than transcribing fresh. On a funded study, that time is not free: it is research-assistant hours the grant already pays for. (The AI tools promise “up to 99% accuracy on clear audio” — Sonix's own FAQ wording — and a cafeteria interview is not clear audio.)
In 2025 alone, 9,668 studies in PubMed Central transcribed interviews verbatim. Only 344 mention using a professional transcription service. The rest ran on researcher time — mostly graduate students typing from scratch or cleaning up machine drafts.
Check these numbers yourself
PubMed Central full-text search, August 2026: "transcribed verbatim" AND 2025[PDAT] and "professional transcription service" AND 2025[PDAT] at ncbi.nlm.nih.gov/pmc.
The calculator
What DIY Transcription Costs Your Grant
Your numbers, not ours. Be honest about the middle one — it includes correcting the AI draft. The hourly figure is what the person costs the grant, not what they take home: stipend plus tuition remission plus fringe. A half-time research assistant funded at $40,000 a year for about 1,000 hours costs $40 an hour. Postdoc and faculty time cost more.
Doing it yourself
$4,000
100 person-hours of budgeted research time — about 5 weeks of half-time work spent typing instead of analyzing.
Cash only. Not counted: the weeks of delay before analysis can start, a second-pass accuracy check, and the de-identification pass. If the time really is free or the study is three interviews, DIY can still win — that is why the inputs are yours.
Already partway through fixing AI drafts? Stop — send us the audio instead.
Landmark, from $1.99/audio minute
$2,388
99%+ accuracy first pass, by people. HIPAA Safe Harbor de-identification included. Standard delivery 3–5 business days; rush 24–48 hours. Focus groups from $2.69/min — full transcription rates.
Sending it upward is the whole point: the person typing rarely controls the budget.
The real question
Use the Labor You Are Already Paying For — Better
The grant did not fund typing. It funded analysis. Every hour a research assistant spends transcribing — or fixing a machine draft with a pair of earbuds and a Word document — is an hour not spent coding in NVivo, writing analytic memos, recruiting participants, or running member checks. Professional transcriptionists do this work faster and more accurately because it is their profession, with HIPAA de-identification applied as they type and a second editor checking every transcript.
Honest exception: if your study is three interviews, or your methodology treats transcribing as a first pass of analysis, doing it yourself can be the right call. This page is for the other case — the 20-, 50-, 200-hour studies where transcription quietly became someone's second job.
One of the 3,500+ researchers already using Landmark? Log in — your team's account covers your next study too.
If you do it yourself
How to Transcribe an Interview for Qualitative Research
No sales pitch in this section — if DIY is the right call for your study, do it well:
- Set up like a transcriptionist. Closed-back headphones, playback at 0.75× speed, and a dedicated rewind-5-seconds shortcut. A free tool like oTranscribe puts play/pause and rewind on keys so your hands never leave the keyboard.
- Decide your conventions before the first line. Speaker labels (Interviewer/P1), a timestamp convention, and whether you need clean or full verbatim — switching mid-study means redoing early transcripts.
- Type one rough pass without stopping. Mark gaps as [inaudible 00:12:34] and keep moving; perfectionism on pass one is where the hours go.
- Do a full second listen against the text. This is the pass that catches the misheard word that changes a code. It is also the step machine drafts make people skip.
- De-identify as you type, not after. Replace identifiers with bracketed tokens — [doctor], [city], [age] — the moment you hear them, and keep pseudonyms consistent across transcripts. Our de-identification guide lists all 18 HIPAA Safe Harbor categories.
- Budget honestly. Several hours of work per audio hour is normal for careful verbatim work. If the calculator above says the study cannot afford that, the answer is not lower standards — it is not doing this part yourself.
Frequently Asked Questions
- How long does it take to transcribe an interview?
- Far longer than the interview itself — several hours of careful work per hour of audio is routine, and correcting an AI draft is often no faster. Use the calculator above with your own honest ratio.
- Is it cheaper to transcribe interviews yourself?
- Only if the time is genuinely free. On a funded study, DIY converts budgeted research-assistant hours into typing. Professional transcription starts at $1.99 per audio minute with de-identification included.
- How fast is a professional service?
- Standard delivery typically runs 3–5 business days; rush delivers in 24–48 hours. Transcripts arrive formatted for NVivo, ATLAS.ti, Dedoose, or MAXQDA.
Put the typing down
By submitting, you agree to our terms and privacy policy.