Transcription Pay Math: How Typing Speed Turns Into an Hourly Rate
$66 is the top of a real platform pay range — per audio hour. Work that audio hour in four hours, which is a normal pace, and the wage is $16.50 an hour. Work it in six, and it is $11. The rate was never the whole number; it is half of a division, and the other half is how long the file takes you. This piece does that division openly, at published rates, so the decision about transcription work rests on arithmetic instead of a headline.
The two numbers that decide the wage: rate and ratio
Transcription pay is usually quoted per audio hour, often shortened to PAH: sixty minutes of recorded content, not sixty minutes of the transcriber's working time. The second number is the ratio — working hours per audio hour. One audio hour takes a skilled transcriber approximately three to six working hours, depending heavily on audio quality and the number of speakers. Clean, single-speaker audio sits at the fast end. Several speakers, crosstalk, and poor recording push toward the slow end, and the platform does not adjust your clock when that happens.
The rates themselves, checked in October 2026, come in two different units, which is why they are so rarely compared honestly. TranscribeMe's own freelancer page states starting pay of $15–22 per audio hour, average monthly earnings of $250, and Special Teams at $60–70 or more per audio hour. A dated check of Rev's freelancer rates on 26 October 2026 listed $0.40–1.10 per audio minute for transcription — multiply by 60 and that is $24–66 per audio hour. Same work, two units, and a top rate nearly three times the other platform's starting band. None of those figures is an hourly wage until the ratio is applied.
Worked scenarios: the effective hourly wage
Every figure below is a scenario — a published rate divided by a stated ratio — not a promise of what any individual will earn. Read across a row to see what the ratio does to the same rate.
| Rate (per audio hour) | At 3:1 ratio | At 4:1 ratio | At 6:1 ratio |
|---|---|---|---|
| TranscribeMe starting, low end: $15 | $5.00/hr | $3.75/hr | $2.50/hr |
| TranscribeMe starting, high end: $22 | $7.33/hr | $5.50/hr | $3.67/hr |
| Rev, low end: $24 ($0.40/audio min) | $8.00/hr | $6.00/hr | $4.00/hr |
| Rev, high end: $66 ($1.10/audio min) | $22.00/hr | $16.50/hr | $11.00/hr |
| Special Teams range: $60–70 | $20.00–23.33/hr | $15.00–17.50/hr | $10.00–11.67/hr |
Two lessons fall out of the table. First, at starting rates the ratio dominates: the difference between a 3:1 and a 6:1 worker on the same $22 rate is the difference between $7.33 and $3.67 an hour — a doubling earned entirely in speed and file selection. Second, the rate ceiling matters just as much at the top: even a fast 3:1 worker on a $15 rate earns $5 an hour, while a slower 6:1 worker on the $66 rate earns $11. Getting faster fixes the ratio. It cannot fix a rate that is too low for the target.
The break-even form of the same math is the useful planning tool: the maximum ratio you can afford is the rate divided by your target hourly wage. At $24 per audio hour and a $12 target, the ratio must be 2:1 or better — faster than the three-to-six range a skilled transcriber typically needs, which tells you that combination does not work before you start. At $66 and the same $12 target, the ceiling is 5.5:1, which sits inside the realistic range. Run your own two numbers before accepting a queue of files.
A four-step runbook for measuring your own ratio
- Time ten audio minutes honestly. Take a short clip — TranscribeMe's work, for example, arrives as short 2–4 minute clips — and record your total working time, including rewinding, checking, and formatting. Divide working minutes by audio minutes. That is your current ratio.
- Convert your platform's rate. If the rate is per audio minute, multiply by 60. Write it as dollars per audio hour next to your ratio.
- Do the division. Rate per audio hour ÷ your ratio = your effective hourly wage today. Compare it with the table, not with the platform's headline.
- Log it per file for two weeks. Note speakers and audio quality beside each ratio. Averages hide the pattern; the log shows whether difficult files, not slow typing, are what is costing you.
What success looks like: after two weeks you can state three numbers without guessing — your ratio on clean audio, your ratio on difficult audio, and the effective hourly wage each produces at your current rate. Decisions made from those numbers are business decisions. Decisions made from the per-audio-minute headline are not.
What moves the ratio — and what does not
Typing speed is one lever, and it is a real one: many companies prefer candidates who can maintain 50 to 75 WPM with exceptional accuracy, and speed at that accuracy is what lets a transcriber hold a 3:1 pace on clean audio instead of rewinding. Audio selection is the second lever — the three-to-six-hour range is driven heavily by quality and speaker count, so a worker who can choose files is choosing ratios. The third lever is the rate itself, reached by qualification rather than speed: beginning transcribers on large platforms may see $5–25 per audio hour, while those who pass advanced skill tests and move into specialised teams can reach $60–70 or more per audio hour — the same band TranscribeMe publishes for its Special Teams, which include medical and specialty styles.
What does not move the ratio is working longer hours at the same pace. Volume multiplies a wage; it does not improve it. Ten hours at $3.67 is a longer day at $3.67.
Getting started, for reference, is cheap in cash terms: TranscribeMe asks for a computer and an internet connection, a PayPal account connected during registration, and a passed entrance exam. The real startup cost is the unpaid practice time spent bringing the ratio down — which is exactly why the kill criterion below is stated in hours, not dollars.
The kill criterion: when to stop, specialise, or move on
Set the line before the first file. A workable version: after 20 logged audio hours, if your effective hourly wage is still below the floor you set for your own time — and your log shows the ratio is not improving week over week — platform general transcription is not a slow start. It is the product working as priced, and the honest options are to qualify for the higher-rate teams, take the typing speed to work that pays by the hour instead of by the audio minute, or stop.
Distinguish the two failures, because they have different fixes. A rate problem — good ratio, low rate — is solved by moving: higher bands, special teams, direct clients. A ratio problem — decent rate, ratio stuck at 6:1 on clean audio — is solved by training accuracy and speed, or by accepting that transcription specifically is the wrong vehicle for this skill. Confusing the two wastes months getting faster for a rate that could never have paid, or chasing premium platforms with a ratio that halves every rate they offer.
General information only, not financial advice. Rates cited were published by the platforms or checked on the dates stated and can change; verify current rates before relying on them.
— Sterling Cole, Money Writer — Earning Desk