Productivity
Can you speak faster than you type? Yes — but not by as much as you think
Raw speaking speed is about three times typing speed. Once you count the read-back and the corrections, the honest advantage is closer to two. That is still an enormous amount of time.
Last updated
Where the raw numbers come from
Conversational English sits in the 120–150 words per minute range; audiobook narration is deliberately slower, around 150–160, and auctioneers and sports commentators go far higher. Typing speed is more variable: touch typists reach 70–90, most people who type for a living land between 40 and 60, and hunt-and-peck typing is often under 30.
Those figures are uncontroversial and easy to verify on yourself in two minutes. What is contested is what they mean, because nobody sends a first draft.
The number that actually matters
Call it finished words per minute: how long a piece of text takes from starting to having something you would send. For typing, that is typing plus a light edit. For dictation, it is speaking, reading back and correcting.
| Typing | Dictating | |
|---|---|---|
| Producing the words | 45 seconds at 50 wpm | 8 seconds |
| Reading it back | 5 seconds — you wrote it as you went | 10 seconds |
| Fixing what is wrong | 5 seconds of typos | 7 seconds, mostly proper nouns |
| Total | 55 seconds | 25 seconds |
An illustration of the shape, not a benchmark. Your figures depend on your typing speed, your microphone, and whether the app knows your vocabulary.
Two things are visible in that table. First, the win is real but roughly two to one, not three to one — the editing eats the difference. Second, the correction line is where the whole thing is decided, and it is the only line you can change.
Measure your own in ten minutes
- Pick three real pieces of writing Not test sentences. An email reply you owe someone, a note to yourself, a message to a colleague. Real writing has your vocabulary in it, which is the whole point.
- Type all three, timing from blank screen to would-send Include the edit. Stop the clock when you would actually press send.
- Wait a day, then dictate three equivalent pieces A day's gap matters: doing the same three immediately measures your memory of the sentences, not your speed.
- Time to would-send again Speaking, reading back, correcting. Same finish line.
- Divide words by minutes for each Two figures. If dictation is not at least 1.5× faster, do not conclude it is not for you — go to the next section, because it is almost certainly vocabulary.
The correction line is the whole game
If dictation feels slow, it is rarely the recognition. It is that the same fifteen words come out wrong every day — a colleague's name, a client, a product, a piece of field jargon. Each one costs a few seconds, and fifteen of them across forty emails is most of your saving.
This is fixable and takes about fifteen minutes, once. Teach the app how those words sound and how they are spelt, and the correction line in your own table collapses. It is the difference between dictation being a novelty and dictation being how you write.
Vocabulary in VV Say the word once, type how it should be spelt. VV biases towards your spelling from then on.Where dictation is slower than typing
A page that only argued one way would be useless. Speaking loses, clearly and repeatedly, in these cases:
- Heavy formatting. Tables, nested lists, anything where half the work is structure rather than words. You end up narrating markup.
- Code. Symbol density defeats it. Dictating prompts about code is a different matter and works well — see voice coding.
- Short bursts. A four-word reply is faster typed than the shortcut plus the speaking plus the glance.
- Anywhere you cannot speak freely. Open-plan offices, trains, a shared house with someone on a call. No software fixes this.
- Precise numbers and identifiers. Order references, postcodes, sort codes. Type them and stop fighting.
The second-order effect nobody measures
Speed is the easy argument. The one people report after a month is different: the things they were putting off get done, because the cost of starting dropped.
A reply that needs three paragraphs sits in the inbox for two days when it costs four minutes of typing. It gets answered immediately when it costs ninety seconds of speaking. The saving is not four minutes — it is the two days, and the mental weight of carrying it.
That effect does not show up in a words-per-minute table, and it is the reason most people who stick with dictation give when asked why. If you want the habit rather than the metric, read building a voice-first workflow.
Questions
How many words per minute does the average person speak?
Comfortable conversational English is usually put at 120–150 words a minute. Presenters deliberately slow to around 100–130 for clarity. You can measure yours in two minutes: read a page aloud at a natural pace and divide the word count by the time.
How many words per minute does the average person type?
Most people who type for work land between 40 and 60 words a minute. Trained touch typists reach 70–90. The variation between people is much larger than for speaking, which is why the dictation advantage is bigger for some people than others.
Does dictation get faster with practice?
Yes, and more than you would expect — but the gain comes from habit rather than skill. Speaking a whole thought before looking at the screen, and trusting the read-back to catch errors, are the two changes that matter. Most people are meaningfully faster in week three than in week one.
Is dictation faster for long documents?
For the first draft, yes, substantially. For the finished document, it depends on how much structure it needs. Many people dictate the draft and then restructure by keyboard, which plays to both. See how to dictate long-form writing.
Does the app's accuracy matter more than my typing speed?
Beyond a threshold, no. Once recognition is good enough that corrections are a handful of proper nouns rather than whole sentences, your own typing speed determines the size of the win. Someone who types at 30 words a minute gains far more from dictation than someone who types at 90.