Interview Transcription: Preparing Audio for Publication

Between the audio recording of an interview and the published text lies a multi-step process. Transcription is the first step — and the one that used to consume the most time. Here's the complete workflow from recording to publication-ready text.

Stage 1: Recording Quality Matters

Transcription accuracy is only as good as your audio. A few principles:

Location. Quiet room without echo. Avoid coffee shops, open-plan offices, and outdoor settings. A conference room or home office works well.

Microphone. For in-person interviews: a lavalier mic on the interviewee (15–20 cm from mouth). For remote: ask the subject to use a headset, not their laptop's built-in microphone.

Format. MP3 at 128+ kbps or M4A. WAV gives slightly better quality but files are 10x larger — the accuracy difference for speech recognition is negligible.

Backup recording. Record on two devices. Phone + dedicated recorder, or Zoom cloud + local recording. Losing an interview recording is unprofessional and painful.

Stage 2: Transcription

Upload the recording to Nagovori. A one-hour interview processes in 2–5 minutes. Copy the text into your editor (Google Docs, Notion, Word).

Handling Recognition Errors

Every model makes mistakes. Common issues:

• Names and proper nouns — "Tim Ferriss" might become "Tim Ferris" or "Tim ferris." Check capitalization and spelling.
• Numbers — the model may write "fifteen" instead of "15" or vice versa. Standardize.
• Domain jargon — specialized terminology may be garbled. Cross-reference with context.
• Homophones — "their/there/they're," "affect/effect" — check these manually.

Stage 3: Editorial Processing

Spoken language differs from written language. A verbatim transcript contains:

• Filler words: "um," "uh," "you know," "like," "so basically"
• Repetition: "we decided, we made the decision, we agreed"
• False starts: "and then we — well actually, what happened was —"
• Grammatical irregularities: people don't speak in perfectly formed sentences

The editor's job is to remove noise while preserving the speaker's voice and meaning. Guidelines:

1. Remove fillers, keep personality. If your subject says "that's a game-changer" — keep it, it's their style. But "um, so, like" — remove.
2. Consolidate repetition. "We decided, we made the decision to pivot" becomes "We decided to pivot."
3. Complete fragments. If a thought trails off but the meaning is clear from context, complete it. If not, cut it.
4. Verify edited quotes. If you significantly altered a statement, send it to the subject for approval. This is both ethical and protective.

Stage 4: Structure

Interviews rarely follow a clean narrative arc. Subjects jump between topics, circle back, and digress. For publication, the text needs structure:

• Group answers by theme, not chronological order
• Add subheadings that tell the story
• Pull the strongest quotes and feature them prominently
• Cut sections that don't add value for the reader

Stage 5: Format and Publish

Q&A Format

Classic interview format: journalist's question, subject's answer. Best for expert interviews and profile pieces. Clean, easy to follow, lets the subject's voice dominate.

Narrative With Quotes

Author-driven text with embedded quotes from the subject. Best for feature stories, profiles, and investigative pieces. Gives the writer more control over pacing and context.

Summary Format

Brief overview with key quotes. Best for news articles and digests. Fastest to produce, least editorial investment.

Pre-Publication Checklist

• All names, titles, and company names verified
• Numbers and dates cross-checked with sources
• Quotes approved by subject (if required by your publication's policy)
• Text proofread for typos and grammatical errors
• Headline and meta description written
• Featured image selected and captioned
• Internal and external links added where relevant

Time Comparison

Stage	Manual workflow	With transcription
Transcription (1 hour)	3–4 hours	5 minutes
Editorial processing	1–2 hours	1–2 hours
Structuring	30–60 minutes	30–60 minutes
Formatting	30 minutes	30 minutes
Total	5–7.5 hours	2–3.5 hours

The time saved on transcription alone — 3+ hours per interview — is the biggest single productivity gain in the workflow.

Conclusion

Automated transcription eliminates the most tedious stage of interview publishing. The creative work — editing, structuring, polishing — still requires human judgment and skill. But now you have the time and energy to do it well, instead of arriving at the editing stage already exhausted from hours of manual transcription.