Word Count, Reading Time, and Speaking Time Explained
Learn how word count, reading time, and speaking time are calculated. See the standard formulas behind every blog timer with our free word counter tool.
Three Numbers on Every Blog Post
The word count, estimated reading time, and estimated speaking time that appear on blog posts, documentation pages, and Medium articles are all derived from the same base metric: how many words are in the text. The formulas are simple, standardized, and you can verify them yourself.
These numbers serve different audiences. The writer cares about word count (length targets for SEO). The reader cares about reading time (commitment before they start). The speaker cares about speaking time (will this fit in my allotted slot).
Word Count
A word counter splits text on whitespace and punctuation boundaries and counts the resulting tokens. Most counters define a word as a sequence of non-whitespace characters bounded by spaces, newlines, or the start/end of the text.
The word count for "The quick brown fox jumps over the lazy dog." is 9. The period is attached to "dog" and does not create a separate token.
Edge cases that trip up basic word counters:
- Hyphenated words: "state-of-the-art" is one word to some counters, four words to others. Most agree on treating hyphenated compounds as one word.
- Contractions: "don't", "it's", "they're" are each one word.
- URLs and code:
https://example.com/pageorimport React from "react"may be counted as multiple "words" because punctuation splits them. For code blocks, this inflates word count artificially. - Numbers with commas: "1,000,000" is one word, not three.
Reading Time
The standard formula assumes an average reading speed of 238 words per minute for English prose. This number comes from a meta-analysis of reading speed studies and is the default used by Medium, Dev.to, and most blogging platforms.
reading_time (minutes) = word_count / 238
A 1,200-word article has an estimated reading time of about 5 minutes. A 500-word article is about 2 minutes.
Most platforms round up to the nearest minute and display whole numbers ("5 min read"), not decimals ("5.04 min read").
For technical content with code blocks, tables, and diagrams, the effective reading time is higher because readers pause to study examples. The formula is a rough heuristic, not a precision measurement. A 1,000-word tutorial with five code snippets takes longer to read than a 1,000-word opinion piece. The formula doesn't account for this.
If you're displaying reading time on your own site, you have two options:
- Use the standardized formula (238 wpm). Readers recognize this convention.
- Adjust for content type: 200 wpm for technical content, 260 wpm for casual blog posts. This is more accurate but less standard.
Speaking Time
The speaking time formula uses a slower rate: 130 words per minute for English speech at a comfortable pace. This is the rate used by Toastmasters and speech writing guides.
speaking_time (minutes) = word_count / 130
A 650-word speech takes about 5 minutes to deliver. A conference talk abstract of 1,300 words is about 10 minutes of speaking time.
Speaking time varies more than reading time. A nervous speaker rushes through at 160 wpm. A deliberate speaker with pauses for slides might deliver at 100 wpm. The 130 wpm formula is the middle ground.
If you're preparing a talk, the formula gives you a starting estimate. Then rehearse and time yourself. The actual delivery time is the only number that matters. Use the formula for the first draft, then adjust.
Character Count and Density Metrics
Beyond word count, text statistics tools often provide:
- Character count (with spaces): total characters including spaces. Useful for platforms with character limits (Twitter, SMS, meta descriptions).
- Character count (without spaces): used for some SEO tools and as a rough measure of content density. Two texts with the same word count but different character counts differ in word length, which affects reading level.
- Sentence count: split on
.,!,?boundaries. Used for readability formulas like Flesch-Kincaid. - Paragraph count: split on double newlines. Useful for assessing text structure. Short paragraphs (2-4 sentences) read faster. Long paragraphs (8+ sentences) signal dense, reference-style content.
Reading Level
Some counters estimate reading grade level using formulas like Flesch-Kincaid or SMOG. These combine average sentence length and average syllables per word to produce a U.S. grade level. A score of 8.0 means "readable by an 8th grader."
For technical writing aimed at developers, reading level scores are often high (10-14) because jargon and code increase syllable counts. This is normal and not a problem. The audience is technical. Writing "database migration" as "the process of moving information from one storage container to another" lowers the reading level but frustrates the reader.
The Flesch-Kincaid formula:
grade = 0.39 * (words / sentences) + 11.8 * (syllables / words) - 15.59
A score of 8.0 means the text is accessible to an 8th-grade reader. Legal documents often score 12+. Technical documentation scores 10-14. News articles target 8-10.
When to Check
- SEO: word count correlates with ranking depth. Articles under 300 words are considered thin content by most SEO tools. Check with the Word Counter.
- Blog post previews: add reading time to your article header. Readers decide whether to commit based on time investment.
- Speaking engagements: check speaking time for talks, presentations, and video scripts.
- Social media: character counts for tweets, LinkedIn posts, and meta descriptions. Twitter's 280-character limit, LinkedIn's 3,000-character post limit, and Google's 155-160 character meta description display limit all depend on character counting.
Try it yourself: open the Word Counter. Paste a paragraph of text. Check the word count, character count, reading time, and speaking time. Verify the reading time formula: divide word count by 238. Verify the speaking time: divide word count by 130.