Where the numbers come from
The calculation is simply word count divided by words per minute. What matters is choosing a defensible rate.
A widely cited 2019 meta-analysis by Brysbaert reviewed a century of reading research and found the average silent reading rate for adults reading non-fiction in English is around 238 words per minute. Reading aloud is considerably slower, at roughly 183 wpm.
That analysis also found the frequently repeated figure of 300 wpm to be too high, and claims of 1,000+ wpm through speed reading to be unsupported once comprehension is measured.
What changes the real rate
Individual and situational variation is wide — the average conceals a range roughly from 175 to 300 wpm for competent adult readers.
Text difficulty matters most. Technical documentation, legal text, and unfamiliar terminology can halve reading speed. Familiar narrative prose can be considerably faster.
Purpose changes behaviour. Skimming for a specific fact is much faster than reading for comprehension, which is faster than studying for retention.
Medium has a measurable effect. Screen reading is typically slightly slower than print, and mobile reading slower again due to narrower line lengths and more scrolling.
Language differs substantially. Chinese and Japanese are measured in characters rather than words, and reading rates across languages vary considerably when measured in words.
Why publishers show read times
Read time estimates appeared widely after Medium introduced them, and research on their effect is mixed but broadly positive.
The main benefit is that they let readers make an informed decision before committing. Someone with three minutes will start a 3-minute piece and abandon a 20-minute one partway — the estimate reduces that mismatch, which tends to improve completion rates.
The counter-effect is real too: a long estimate can deter readers who would otherwise have started. Some publishers now show estimates only on shorter pieces, which is arguably a manipulation of the reader's decision rather than an aid to it.
Calculating it well
A few refinements make estimates more accurate:
- Add time for images. A common approach adds 12 seconds for the first image, 11 for the second, and so on down to 3 seconds, reflecting decreasing attention.
- Account for code blocks. Reading code is much slower than prose — often half the rate or less.
- Round sensibly. “7 min read” is more useful than “6.8 min read”. Rounding up is the safer error.
- Set a floor. Anything under a minute should show as “1 min read” rather than a fraction.
Speaking pace, which is a different question
If you are writing a script, the relevant rate is much slower. Comfortable presentation pace is around 130 words per minute; conversational speech runs 120–150; audiobook narration is typically 150–160.
The practical implication catches people out: a 5-minute talk is roughly 650 words, which is far less than most people expect — about one and a half pages of double-spaced text. Writing a 1,500-word speech for a 5-minute slot guarantees either rushing or overrunning.
Add time for pauses, audience reaction, and any demonstration. Most experienced speakers write to roughly 80% of their allotted time.
Frequently asked questions
How many words per minute do people read?
Around 238 words per minute for adults reading non-fiction in English silently, based on a 2019 meta-analysis by Brysbaert. Reading aloud averages roughly 183 wpm.
Is speed reading real?
Rates above roughly 400 wpm generally involve skimming rather than full reading, and comprehension drops measurably. Claims of 1,000+ wpm with full understanding are not supported by the research.
How many words is a 5-minute speech?
Roughly 650 words at a comfortable presentation pace of 130 words per minute. That is far less than most people expect, and writing more guarantees rushing or overrunning.
Should articles display a reading time estimate?
It helps readers decide before committing and tends to improve completion rates. The counter-effect is that a long estimate can deter readers who would otherwise have started.