DEV Community

Michael Lip
Michael Lip

Posted on • Originally published at zovo.one

Word Count Is the Most Misunderstood Metric in Writing

What counts as a word? It seems obvious until you have to define it precisely. Is "well-known" one word or two? Is "don't" one word or two? Is "300" a word? Is a URL a word? Different word counters give different answers because there is no universal definition.

Microsoft Word counts "well-known" as one word (separated by a hyphen, not a space). Google Docs counts it as two words. Most web-based counters split on whitespace, making "well-known" one word but "well known" two words. The discrepancy matters when you are writing to a strict word limit.

Why word count matters

Academic submissions. Journals and conferences have strict word limits. A 6,000-word limit means 6,000 words, not 6,050. Exceeding the limit can result in desk rejection without review.

SEO content. Search engine optimization research suggests that comprehensive content (1,500-2,500 words) tends to rank better for informational queries. Too short and Google may not consider it comprehensive. Too long and readers bounce before finishing.

Publishing contracts. A book contract specifies a word count range. A 70,000-80,000 word novel. Deviating significantly affects production costs, pricing, and marketing.

Social media. Twitter has a 280-character limit. LinkedIn posts cap at 3,000 characters. Character count matters more than word count for these platforms, but word count determines how many characters you need.

Beyond word count: the metrics that matter

Character count (with and without spaces). Required for social media, SMS, and any character-limited context. "Hello world" is 11 characters with spaces, 10 without.

Sentence count. Average sentence length correlates with readability. Short sentences (10-15 words) are easy to read. Long sentences (25+ words) are harder. A mix is natural. All short is choppy. All long is exhausting.

Paragraph count. Web content should have short paragraphs (2-4 sentences). Academic writing allows longer paragraphs. The paragraph count relative to word count tells you about the density of the writing.

Reading time. The average adult reads 200-250 words per minute for non-technical content. A 1,500-word article is a 6-7 minute read. Displaying estimated reading time sets reader expectations and has been shown to increase engagement (Medium popularized this).

Speaking time. The average speaking rate is 125-150 words per minute. A 10-minute presentation should be approximately 1,250-1,500 words. Going over means either rushing or going over time.

Readability scores. The Flesch-Kincaid Grade Level estimates the US school grade needed to understand the text. A score of 8 means an eighth-grader can understand it. Most popular writing targets a grade level of 7-9. Technical documentation might be 12+.

Flesch-Kincaid Grade = 0.39 * (words/sentences) + 11.8 * (syllables/words) - 15.59
Enter fullscreen mode Exit fullscreen mode

The definition problem in code

Implementing a word counter requires defining the split rule:

// Naive: split on spaces
text.split(/\s+/).filter(Boolean).length;

// Better: split on word boundaries
text.match(/\b\w+\b/g)?.length || 0;

// More inclusive: handle contractions, hyphens, Unicode
text.match(/[\w'-]+/g)?.length || 0;
Enter fullscreen mode Exit fullscreen mode

Each regex gives different results. The word boundary \b approach splits "don't" into "don" and "t" (two words). The character class approach keeps "don't" as one word. Neither is definitively "correct."

For Japanese, Chinese, and other languages without spaces, word counting requires a tokenizer that understands word boundaries -- a fundamentally harder problem than English word counting.

The practical tool

I built a word counter at zovo.one/free-tools/word-counter that provides word count, character count (with/without spaces), sentence count, paragraph count, reading time, speaking time, and readability scores. It matches Microsoft Word's counting behavior for consistency, since that is what most people consider the "correct" count.

Paste your text, get every metric you need. It is the fastest way to verify that your content hits the target length.

I'm Michael Lip. I build free developer tools at zovo.one. 500+ tools, all private, all free.

Top comments (0)