Transcription is one of the few remote jobs with no barrier at all — no degree, no portfolio, no interview, just a test you can take today. That is exactly why it is crowded, and why the advertised rates and the real ones are so far apart.
The thing that explains every disappointed review
Transcription is paid per minute of audio, not per hour of work. A platform offering “$0.60 per audio minute” pays $36 for a one-hour recording. That sounds fine until you know how long an hour of audio actually takes.
| Experience level | Work time per 1 min of audio | Effective hourly rate at $0.60/min |
|---|---|---|
| First month | 5 – 6 min | $6 – $7 |
| A few months in | 3 – 4 min | $9 – $12 |
| Experienced, clean audio | 2 – 3 min | $12 – $18 |
| Experienced, specialised | 1.5 – 2 min | $20 – $30+ |
That ratio is the whole job. Everything that raises your pay works by lowering it — typing speed, shortcut software, familiarity with the subject, and above all refusing bad audio.
What beginners realistically earn
| Type of work | Typical effective hourly (USD) |
|---|---|
| General transcription, beginner | $8 – $15 |
| General transcription, experienced | $18 – $25 |
| Legal transcription | $25 – $40 |
| Medical transcription | $25 – $40 |
| Captioning and subtitling | $15 – $30 |
The top two rows are the same skill. The difference is vocabulary and formatting conventions you can learn, and a smaller pool of people who have.
What actually raises your rate
Refuse bad audio. Three overlapping speakers on a phone recording can take eight minutes of work per audio minute. The job pays the same as clean audio. Learning to decline is worth more than typing faster.
Learn the shortcuts. Foot pedal control, text expansion for repeated phrases, and playback-speed control are the difference between the second and third rows above.
Specialise. Legal and medical are where the money is, and both are learnable without a degree — the barrier is terminology and format, not credentials. Choosing one skill and going deep applies here more than almost anywhere, because the general tier is permanently crowded.
Spotting the bad offers
Transcription attracts fake listings because it advertises to people who need money now and have no experience to judge with.
- Never pay to apply, test, or access work. Every legitimate platform tests you free
- Be suspicious of rates far above the table above. $2 per audio minute for general work is not a discovery, it is bait
- Do not buy special software to start. Free tools cover everything a beginner needs
- Check payout terms in writing — per audio minute or per audio hour, and when it is paid
These are the same patterns as every other fake remote listing, covered in how to spot online job scams.
Is it worth doing?
Honestly: as a first paid remote job, yes. It requires nothing, it starts immediately, and it builds a payment history and a review record you can use elsewhere.
As a career, only if you specialise. General transcription competes globally on price and automatic speech recognition keeps improving at exactly the tier beginners work in. The people still earning well are the ones doing what the software gets wrong — accented multi-speaker audio, legal formatting, medical vocabulary.
Treat the general tier as the bridge, not the destination, and pick the specialisation early.
The bottom line
Expect $8–$15 an hour starting out, not the headline per-minute rate. The ratio of work time to audio time is the entire economics of the job. Refuse bad audio, learn the shortcuts, and specialise into legal or medical if you want it to pay properly.