Reading Time Estimator

What a reading time actually promises

A reading-time label is a small contract with the reader: spend this long and you will have finished. Get it right and people start articles they would otherwise skip. Get it wrong in the optimistic direction and you have trained them to distrust the label.

The arithmetic is trivial — words divided by words-per-minute — so all the honesty lives in the speed you assume and in whether your content is the kind that a word count can describe.

Where 238 comes from

This tool defaults to 238 words per minute for silent reading. That figure comes from a meta-analysis of silent reading research rather than from convention, which is why it is not a rounder number.

Round defaults are everywhere: 200, 250, sometimes 300. They are easy to remember and they quietly change your estimate by twenty percent or more. If you have reason to believe your audience reads faster or slower — a technical readership working through unfamiliar terms, say — change the number rather than accepting a default someone chose for tidiness.

Three speeds, three questions

The tool reports three figures because writers ask three different questions.

• Reading, 238 wpm by default. The one to put on an article.
• Speaking, 150 wpm. Fixed, because it is a property of speech rather than of the reader. Roughly two-thirds of reading speed, which is why a five-minute article becomes an eight-minute talk. Script writers who plan by reading time consistently over-run.
• Skimming, 450 wpm. What a lot of web readers actually do. Useful as a floor: if the skim time is already long, the piece is long regardless of how well it is written.

When the word count stops working

Word count predicts duration well for prose and badly for everything else.

Code blocks are read line by line, often re-read, and sometimes copied out and run — a hundred words of code is not a hundred words of narrative. Tables are scanned rather than read. Diagrams take time that no counter can see. Reference material is not read front to back at all.

The practical adjustment is to round up on technical content rather than reporting the raw figure. Under-promising and over-running is the failure mode readers remember.

A note on comparing tools

If your CMS shows a different reading time than this page does, the cause is almost always a different assumed speed — many platforms hardcode 200 or 250 — or a different approach to stripping markup before counting. Neither is more correct. They are different assumptions about the same text, which is a good reason to know what assumption you are publishing.

What speeds does it use?

Reading defaults to 238 words per minute, speaking aloud is fixed at 150, and skimming at 450. The reading speed is adjustable; the other two are there for comparison.

Why 238 and not a round 200?

Because it comes from a meta-analysis of silent reading studies rather than a guess. Round numbers like 200 or 250 are common on other sites precisely because they are round. If your audience is slower or faster, change the figure — that is what the control is for.

Why are speaking and skimming shown too?

Because they answer different questions. If you are writing a script, the number that matters is the spoken one — roughly two-thirds the reading speed, which is why a five-minute article becomes an eight-minute talk. Skimming is what most people actually do with a web page.

Does it account for images, code blocks or tables?

No. It counts words. An article dense with code samples or diagrams takes longer to work through than its word count suggests, because those are read differently — often slower than prose, sometimes skipped entirely.

How does it count words?

By splitting on whitespace, so hyphenated compounds count as one and a numeral counts as a word. That is the same convention most word counters use, which keeps the estimate comparable to other tools.

Should I put a reading time on my articles?

It tends to help readers decide whether to start, and an honest one builds more trust than an optimistic one. If your content is technical, consider rounding up rather than down — under-promising the time and over-running is worse than the reverse.

Why does the estimate differ from the one my CMS shows?

Different assumed speed, almost always. Many platforms hardcode 200 or 250 words per minute and some strip markup differently before counting. Neither is more correct; they are different assumptions about the same text.

Is my text sent anywhere?

No. Counting happens in your browser and nothing is stored or transmitted.