What does a good transcript look like?

A good transcript has four visible features. Each speaker has a label, time marks let a reader find any passage in the recording, the words stay in the order they were spoken, and every part that cannot be heard is marked. The wording itself depends on the style. A meeting transcript in clean verbatim drops the "um" and the false starts, while a research transcript in full verbatim keeps them. Both can be good, because each matches its job.

A good transcript puts one speaker on each line

Each change of speaker starts a new line, with the speaker's name first. GoTranscript, a transcription company, advises formatting each line as "Speaker Name: Text of speech." ¹ When names are unknown, use a role such as Interviewer, or a generic label such as Speaker 1. Decide the label before you start and keep it the same on every line. A transcript that calls one person "Dan," "Daniel" and "D." makes a reader think there are three people in the room.

Timestamps let a reader find the moment in the audio

A timestamp connects a line of text to a point in the recording, so a reviewer can check a passage in seconds. There are two common methods. Interval timestamps mark the recording at a fixed rate, such as every one or two minutes. ² Speaker change timestamps mark each new turn, which GoTranscript says is inserted "each time a new speaker talks." ³ Interval marks are quick to add to a draft. Speaker change marks take more listening, but they are easier to use in an interview, where the turns matter more than the clock.

Clean verbatim suits most meeting transcripts

A meeting transcript usually keeps the speaker's wording and drops the filler words, false starts and stutters. Rev, a transcription company, describes its non-verbatim style, which removes these, as "the most common" type of transcript. ⁴ The same meeting in full verbatim would keep every "um." That is useful only when the hesitation is part of the record. Our guide on verbatim vs clean verbatim transcription compares the styles side by side.

An example of a meeting transcript

The example below was written for this page to show the layout. It is not from a real recording. The audio was a short planning call.

[00:00] Priya: Okay, the release is locked for March.

[00:09] Sam: Can we confirm the vendor date before we tell the client?

[00:16] Priya: Yes, it's the twelfth. Let's tell them the twelfth.

[00:21] Sam: The vendor said [inaudible 00:23] the second week, I think.

[00:27] Priya: Then we check with them today and send the update tomorrow.

Four things in this example are the conventions described above. Each turn starts with a time and a name. The wording stays as spoken, and the filler words are gone. The one passage that could not be heard is tagged with its time, so a reviewer can go straight to it. The last point matters most. A reader who never hears the recording can see where the transcriber was unsure, and can decide whether the gap matters for the decision being made.

Legal and formal transcripts follow a fixed layout

When a transcript is for a court or a formal hearing, the layout comes from the format, not from taste. The practical exam for the Certified Electronic Transcriber, run by AAERT, asks for a transcript of about 10 to 20 pages in the federal format, with a title page, an index page and a certification page ⁵. A meeting transcript does not need those pages. The lesson still applies: when a layout is required, write it into the order before anyone transcribes.

It depends: research and captions need different features

Research interviews that code hesitation need full verbatim, with pauses and fillers kept, as covered in the guide to transcribing an interview. Captions are not transcripts. They use short cues with start and end times, and the rules are in the captions and subtitles guide. If you need a transcript for a legal or research file, ask for its format in writing. If you need a readable record of a meeting, clean verbatim with speaker labels and interval timestamps is the usual choice.

What to check before you call a transcript good

Before you share a transcript, check these four things against the audio:

1. Every turn has the right speaker label.
2. The timestamps fall where the speaker changes, or at the interval you set.
3. Every tag, such as [inaudible] or [crosstalk], has a time and sits where the words were lost.
4. Names, numbers and technical terms match the audio, not the guess.

The question on transcription QA explains how a second person does this check, and why the names and numbers need the most attention. If your recording is poor, the question on transcribing low-quality audio covers what to expect. For the layout of a meeting record, the question on turning a meeting transcript into minutes shows the next step, and the question on how often to add timestamps gives more detail on the time marks. For the full set of terms, see our transcription overview, and the glossary entries for clean verbatim and timestamp.