Phone Typing Speed: How Fast People Type on Phones

People type about 36 words per minute ( WPM) on a phone, about 30 percent slower than average typing on a computer keyboard ^{1 2}. In a study of 37,370 volunteers, the clearest differences come from posture and from text-entry aids: typing with two thumbs and using autocorrection go with faster speeds, while picking word suggestions goes with slower speeds. Swipe typing did not raise average speed in that study, and the authors report a gain after practice in a field study they cite ³. In a lab study on an iPhone, speech entered English text about three times faster than the built-in keyboard ^{4 5}.

This guide sets out the studies behind those numbers and what each one can and cannot show. For the wider picture, read our typing speed overview. If you want the short version of what to change, read how to type faster on a phone.

People type about 36 WPM on a phone, about 30 percent slower than on a keyboard

The main figure here comes from an online study of 37,370 volunteers who copied sentences on their own phones. Their average was 36.2 WPM, with 2.3 percent of errors left uncorrected in the final text ¹. The test measures copying speed, not composing, so it describes how fast the fingers move, not how fast ideas arrive. For how that kind of count works, see how typing speed is measured.

Each volunteer read 15 sentences, memorized each one and then typed it as quickly and accurately as possible ^{6 7}. The desktop benchmark used the same transcription method, but it comes from different people, so the two averages are not a like-for-like comparison ². Real messages also involve deciding what to say, which this test leaves out.

The desktop benchmark is about 51.56 WPM, from an online typing test that recorded 136 million keystrokes ⁸. The phone study also compared its phone data with desktop data gathered by the same method, and its authors report that average mobile typing is about 15 WPM slower ⁹. Divide 36.2 by 51.56 and you get 0.70, so phone typing runs about 30 percent slower than desktop typing ². The authors put it another way: with one or two fingers, people type about 70 percent as fast on a phone as on a full keyboard ¹⁰.

Smaller studies point the same way. A field study of 30 people who typed in their own apps for three weeks recorded a mean of 32.1 WPM ¹¹. That count leaves out suggestions and corrections, so it is not a like-for-like count with the others.

The studies differ in sample, task and device, so read their numbers as a range rather than a single benchmark. The table shows what each one measured.

Two thumbs are faster than one finger

Yes. In the study of 37,370 volunteers, typing with both thumbs was significantly faster than typing with one thumb or an index finger ¹³. Two thumbs was the fastest posture the study measured, at 38.02 WPM ¹⁴. People who reported using two fingers averaged 37.7 WPM, against 29.2 WPM for those who used only one finger ¹⁵. Most people type this way. The study's web summary puts the share of two-thumb typists at over 74 percent ¹³, and the full paper at over 82 percent ¹⁶. In the field study, two thumbs was the most popular posture, at 74.5 percent of each person's typing ¹⁷.

The authors attribute the two-thumb advantage to frequent switching between the sides of the screen, which allows preparatory movements that shorten the gaps between keystrokes ¹⁸. That explanation comes from earlier studies, and this study did not test it directly.

People also switch postures. On average, participants in the field study used 4.6 different postures at least once ¹⁹, so a single posture label per person hides several habits.

Posture was self-reported in an online test, so this shows an association, not an experiment in which people were assigned a posture. The study does not show that moving a one-finger typist to two thumbs will raise their speed by the same amount. It shows how the people who already type each way performed. The two thumbs versus one finger question covers the same evidence in more detail, and touch typing covers the desktop skill that, as the age section notes, did not raise phone speed in this study.

Autocorrect goes with faster typing, while picking word suggestions goes with slower typing

Two text-entry aids show up clearly in the study data. Autocorrection is linked with faster speeds, and word prediction, which means tapping a suggested word, is linked with slower speeds. Both are patterns in observed use, not tests in which people were given or denied a feature.

Autocorrection users averaged 43.4 WPM, faster than every other group ²⁰. Across all participants, the correlation between autocorrection use and speed was 0.237, a moderate positive link ²¹. Prediction alone averaged 32.8 WPM, a little below the 34.8 of people who used no feature ²², and its correlation with speed was negative, at -0.183 ²³.

Squared, the correlation of 0.237 ²¹ means autocorrection use accounts for about 5.6 percent of the variation in speed between people ²⁵. The rest of the difference between typists comes from other factors, so no single phone setting decides your speed.

The authors also tested the autocorrection result in their U.S. subsample of 1,475 participants ²⁶. There, the speed difference between autocorrection and normal typing was not statistically significant ²⁷. So treat autocorrection as a likely help, not a guaranteed one. The data gives no reason to turn it off for speed. For how the feature works in practice, see the guide to text expanders and autocomplete, and for the question itself, does autocorrect make you type faster.

The field study also logged autocorrection. Among the 16 people who used it, the median was 15.3 automatic corrections per person per day ²⁸. The field study points the same way for suggestions. People picked a word suggestion on average once every 63 keystrokes ²⁹. The participant who picked suggestions most often, once every 11 keystrokes, was also the slowest typist in that study, at 15.6 WPM ³⁰. That is one person, so it illustrates the pattern without proving it. The tablet study reached the same result for prediction: adding word prediction to a keyboard further decreased typing speed, without improving the error rate ³¹.

Teenagers are the fastest phone typists, and speed falls with every decade after

Participants aged 10 to 19 were the fastest group, at 39.6 WPM ³². From there the average falls with each decade: 36.5 WPM in the 20s ³³, 32.2 in the 30s ³⁴, 28.9 in the 40s ³⁵ and 26.3 in the 50s ³⁶. Participants under 10 were the slowest group, at 24.3 WPM ³⁷.

Speed and time spent typing are different things. The authors note that the 10 to 19 group was fastest even though it spent less time on its phones than people aged 20 to 39 ³⁸. The age pattern is an association in a self-selected sample of people who chose to take a typing test. It tells you where averages sit, not what a particular person can reach. For the desktop equivalent, see average typing speed by age.

Language experience separated typists as well. Native English users were the fastest group, at 37.8 WPM, and people who never type in English were the slowest, at 25.6 WPM ³⁹.

A touch typing course made only a small difference to phone speed. People who had taken one for desktop keyboards averaged 35.7 WPM on phones, slightly below the 36.4 of those who had not ⁴⁰. Our touch typing guide explains the desktop method and what it does and does not change.

Swipe typing did not raise average speed, and its gain shows up after practice

Swipe typing, which the study calls gesture typing, did not raise average speed across the sample. People who reported gesture typing averaged 32.2 WPM overall, against 34.8 for people with no text-entry aid ²⁴. That is a small gap, and the study does not measure how long people had practised, so it cannot say whether the technique pays off with time.

The authors report that gesture typing and word prediction were both linked with fewer keystrokes per character ⁴¹, a measure of how many key presses each character takes ( keystrokes per character). Fewer keystrokes did not give the gesture group a higher average speed in this sample, so fewer keystrokes did not show up as higher measured speed in this sample.

The strongest evidence for swipe typing is about practice. A field study that Palin and colleagues cite found that people were almost 10 WPM faster after practice with a gesture keyboard than with tapping ³. We did not open that study, so read the figure as the Palin team's report of it. The lesson for a reader is to judge swipe typing by your own speed after you have practised, not on the first day. The swipe typing question sets out when it is worth switching.

Tablets have no large speed benchmark, and the controlled tablet study found word prediction slowed typing

We found no study of tablet typing at the scale of the phone data. The one controlled study we could read tested several virtual QWERTY keyboards with twenty young adults ¹². Its authors state that text entry on touch devices is slower and more error-prone than on a traditional computer keyboard ⁴², and that word prediction made the keyboard slower still ³¹. The authors conclude that a virtual QWERTY keyboard without visual changes was the fastest method among the ones they tested ⁴³, so the plain keyboard was the best of the options they tried.

We did not use speed figures from product reviews, since we could not check how they were measured. This guide therefore does not state a tablet average, and it does not say how much a physical keyboard closes the gap for tablets, because we did not find a controlled measurement of that. If you dictate on an iPad instead, read dictation on iPad.

Speaking into a phone is about three times faster than typing on it

In a controlled study on an iPhone, English speech entered text at 153 WPM, against 52 WPM on the built-in keyboard, a rate 2.93 times faster ^{4 5}. Mandarin showed the same pattern, with speech at 123 WPM against 43 WPM on the Pinyin keyboard ^{44 45}. Over 1,000 words, the keyboard took about 12.7 minutes longer than speaking, before any editing ⁴⁶.

Speech is not error-free. Its final text had slightly more uncorrected errors, at 1.30 percent, against 0.79 percent for the keyboard ⁴⁷. So the faster route also leaves more cleanup, which matters for long documents.

The study used short phrases in a laboratory, and its authors say results outside the laboratory still need study ⁴⁸. The revised paper also says that upper-bound speech recognition performance has greatly improved compared with earlier systems ⁴⁹, so the speed gap may be a floor rather than a ceiling. For the wider comparison, read speaking vs typing speed and dictation vs typing. On an iPhone, the iPhone dictation guide shows how to start. If you want speech on a computer, Yap, which we make, is a dictation app for Mac and Windows, and you can compare download options.

Average phone speed is about 36 WPM, and autocorrection is the setting the data supports

Measured phone speed sits near 36 WPM, and the evidence points to three things a reader can act on. First, keep autocorrection on. Its users were the fastest group, and the data gives no reason to turn it off for speed ²⁰. Second, use both thumbs where you can, since two thumbs was the fastest posture measured ¹⁴, although that is an association rather than a tested effect. Third, give swipe typing time, because the benefit in the cited field study came after practice ³.

For long text, the choice is between speed and cleanup. Speech entered text about three times faster than a phone keyboard ⁴, but it left more uncorrected errors ⁴⁷, so plan time for editing. The how to type faster on a phone question gives the step-by-step version of these points, and the two thumbs versus one finger question sets out the posture evidence on its own.