Speaking vs Typing Speed: Is Talking Really 3x Faster?

The short answer on speaking vs typing speed is yes, and the number holds up: a Stanford and Baidu study clocked dictation at 153 words per minute against 52 for an iPhone keyboard, roughly 2.93 times faster. The longer answer is that the same study found dictation left more errors in the finished text than typing did.

Almost every page quoting the "3x faster" line stops at the first half. Two things it skips decide whether talking is genuinely the quicker way to get something said.

One widely quoted figure on this topic is also simply misread, so we'll clear that up first.

Speaking vs typing speed at a glance

*Figures come from controlled or large-scale studies and describe averages, not any one person. The two error-rate rows average English and Mandarin together, which is how the study reports them. Sources at the end.

How fast people actually type on a phone

The largest measurement of mobile typing is a 2019 study from Aalto University, the University of Cambridge and ETH Zurich, which recorded more than 37,000 volunteers transcribing sentences on their own devices.

People typing with two thumbs averaged 38 words per minute. The fastest person in the entire dataset reached 85 wpm on a touchscreen.

Now the misread figure. Comparisons like this one routinely quote "most phone users type 35 to 65 wpm." That range isn't phone typing at all. The Aalto researchers were describing physical keyboards, where most people land between 35 and 65 wpm and the shrinking minority who reach 100 wpm are the exception. Phone thumbs average 38.

The correction runs in both directions, and the second half is the more interesting bit. If two-thumb touchscreen typing averages 38 wpm and physical keyboards mostly run 35 to 65, then the phone is only about 25% slower than a desktop keyboard, which is exactly what the researchers concluded. The phone keyboard isn't the bottleneck people assume it is.

Two other findings stand out. The single biggest predictor of speed was whether someone used one finger or two thumbs, and over 74% used two thumbs. And there's a clear generation gap: people aged 10 to 19 typed roughly 10 wpm faster than people in their 40s, which the researchers put down to accumulated practice rather than anything about young reflexes.

So hold 38 wpm as the realistic phone-typing figure.

How fast people talk

Conversational English is usually put at about 150 words per minute, a benchmark widely attributed to the National Center for Voice and Speech. Treat it as a rule of thumb rather than a measurement: the attribution is repeated everywhere, but the underlying publication isn't available online, so you can't check it the way you can check the studies above.

The measured number is more useful anyway. When the Stanford team clocked people dictating actual messages, they got 153 wpm, which lands almost exactly on the conventional estimate.

Set 153 against 38 and the contrast is plain. Ordinary speech moves about four times faster than thumbs on glass, and you don't need a study to feel it. Dictating a two-sentence reply takes seconds; tapping it out takes noticeably longer.

That gap is a large part of why voice notes spread the way they did. A 40-second voice message carries far more content than you could comfortably type in the same window, and it asks almost nothing of your hands.

The head-to-head study

The cleanest direct comparison is a 2016 study by Sherry Ruan, Jacob Wobbrock, Kenny Liou, Andrew Ng and James Landay, published in 2017. The team tested Baidu's Deep Speech 2 recognizer against the built-in iPhone 6 Plus keyboard, in English and Mandarin Chinese, under lab conditions.

For short message transcription:

• English: speech ran at 153 wpm against 52 wpm for the keyboard, about 2.93 times faster.
• Mandarin Chinese: speech ran at 123 wpm against 43 wpm, about 2.87 times faster.

Speed wasn't the only win. Speech also produced fewer corrections during entry, with a corrected error rate of 5.30% against the keyboard's 11.22%, averaged across both languages. People stopped to fix their own mistakes less often when talking than when typing.

Two caveats keep this honest. The study measured upper-bound performance in a quiet lab with cooperative participants, and the authors say so plainly, calling for more work to quantify performance outside the lab, where noise, distraction and being on the move all degrade recognition. And the task was transcribing short phrases people were given, not composing their own text, and certainly not long-form drafting.

Note too that the keyboard figure here, 52 wpm, sits well above the 38 wpm average from the 37,000-person study. Lab participants copying out given sentences do better than people composing their own text on a bus. If anything, the everyday speech advantage is wider than 2.93x, not narrower.

The first catch: the errors that survive

Here's the part that gets dropped when people quote the 3x line. Although speech produced fewer corrections during entry, it left slightly more errors in the final text: an uncorrected error rate of 1.30% for speech against 0.79% for the keyboard.

Those numbers are small, but the direction matters. Typing shows you each character as it lands, so you catch problems immediately. Dictation hands you a block of text the system thinks it heard.

The mistakes that survive dictation are the sneaky kind: a wrong homophone, a name spelled the common way instead of the right way, a missing "not" that inverts the whole sentence. They slip past a quick glance precisely because the sentence still reads fluently.

So fast input doesn't automatically mean fast finishing. If a dictated message needs a careful proofread, part of the speed advantage goes straight back into the review.

The second catch: fast to send, slow to receive

Producing words by voice is quick. Consuming them is a different problem, and it's the one voice notes get criticized for most.

Silent reading of non-fiction English averages around 238 words per minute, based on a 2019 meta-analysis of 190 studies. Listening runs at whatever pace the speaker chose, somewhere near 150 wpm, and you can't skim it.

So the same message costs the sender less time by voice and the receiver more. It's the same asymmetry behind why reading tends to beat listening once a message runs past a sentence or two.

Playback speed only partly fixes it. Comprehension holds up reasonably well up to double speed for familiar material, with a rough ceiling near 275 wpm, but speeding audio up has limits that reading doesn't. You still can't jump to the third sentence, search for a name, or quote a line.

Speed for the sender isn't speed for the receiver. Any honest answer on speaking vs typing speed has to count both sides.

So when should you speak, and when should you type?

Neither method is simply better. The choice comes down to the setting and the job.

Speak when:

• You want a thought captured fast and the exact wording doesn't matter much
• Your hands are busy, or you're walking, driving or cooking
• The message is long enough that typing it would be tiring
• You're alone, or somewhere you don't mind being overheard

Type when:

• The wording has to be precise, or the text is technical
• You're on a train, in an open-plan office, or near a sleeping baby
• You need symbols, formatting, a pasted link or a surgical edit
• The recipient will want to skim, search or quote it later

The useful move is to stop treating the two as rivals. Speaking is the fastest way to get words out; reading is the fastest way to take them in.

That pairing is what Transcribbit does: forward a WhatsApp voice note and get clean text back in seconds, so a message recorded at talking speed can be read at reading speed. The free tier covers 50 transcriptions a month, and your audio is deleted straight after. Try it free.

Frequently asked questions

Is speaking really three times faster than typing?

On a smartphone, roughly yes. The Stanford and Baidu study measured dictation at 153 wpm against 52 wpm for the iPhone keyboard in English, about 2.93 times faster, and 123 against 43 wpm in Mandarin Chinese. The exact multiple depends on the device, the language and how good a typist you are.

How fast does the average person type on a phone?

About 38 words per minute with two thumbs, based on a study of more than 37,000 volunteers. The fastest person recorded hit 85 wpm. Be careful with the widely quoted "35 to 65 wpm" range: that describes physical keyboards, not phones.

Does dictation make more mistakes than typing?

Fewer while you're composing, slightly more in the finished text. Speech produced fewer corrections during entry, 5.30% against 11.22%, but left more errors behind in the final text, 1.30% against 0.79%. The survivors tend to be misheard words that still read smoothly, so dictated text rewards a proofread.

If voice is faster, why do people still type?

Because input speed isn't the only thing that matters. Typing gives you precise editing, silence in shared spaces, and text the recipient can skim and search later. People switch based on context, not on a stopwatch.

Is typing on a phone much slower than on a computer?

Less than most people assume. Two-thumb touchscreen typing averages 38 wpm, roughly 25% slower than the physical-keyboard speeds measured in a comparable large study, where most people land between 35 and 65 wpm.

Which is better for long messages, speaking or typing?

Speaking, for producing them. Reading, for receiving them. A long voice note is far quicker to record than to type, and slower to listen to than the same text is to read. A transcript gives you both ends of that.

Final Thoughts

Speaking wins on input speed, and not by a little: on a phone the gap is about 3x. It also holds up on accuracy while you're composing, since talkers correct themselves less often than typists do.

Typing wins on precision, privacy, editing, and how easy the result is to skim later. The phone keyboard is also closer to a desktop one than its reputation suggests, only about 25% behind.

The two honest caveats are that dictation leaves more errors in the final text than typing does, and that the time it saves the sender it partly charges to the reader. Speak to capture, read to consume, and you don't have to choose.

References

[1] Ruan, Wobbrock, Liou, Ng and Landay, Comparing Speech and Keyboard Text Entry for Short Messages in Two Languages on Touchscreen Phones (2016, published in PACM IMWUT 2017)

[2] Aalto University, University of Cambridge and ETH Zurich, Smartphone typing speeds catching up with keyboards (2019)

[3] Brysbaert, How many words do we read per minute? A review and meta-analysis of reading rate (2019)