Word and Character Count Explained: Content Limits, Reading Time, and Why Counters Disagree
Counting words sounds like the one task computers cannot get wrong, yet paste the same text into Microsoft Word, Google Docs and an online counter and you can get three different numbers. Counting characters is worse: an emoji can be one, two or five "characters" depending on who is counting. When a limit matters, such as an essay's word count, a meta description, a tweet or a database column, you need to know which definition the limit uses, and what reading speed sits behind a "5 min read" label.
How a word counter decides what a word is
Most counters split text on whitespace and count the pieces. The differences come from everything that is not a plain space: hyphens, dashes, slashes, numbers, URLs and punctuation standing alone. "Well-known" is one word in most counters. "word—word" with an em dash and no spaces may be one or two. A URL is usually one word, however long. A lone & or — surrounded by spaces counts as a word in simple counters and not in smarter ones.
// Simple: split on whitespace (what most quick counters do)
const words = text.trim().split(/\s+/).filter(Boolean).length;
// Language-aware: Intl.Segmenter knows words in languages without spaces
const seg = new Intl.Segmenter('ja', { granularity: 'word' });
[...seg.segment('東京タワーに行きました')].filter((s) => s.isWordLike).length;
// 5: 東京タワー | に | 行き | ま | したChinese, Japanese and Thai do not separate words with spaces, so whitespace counting treats a whole sentence as one word. For those languages, character count is the usual measure of length, and dictionary-based segmentation such as Intl.Segmenter is needed for real word counts.
When a limit is strict, as with essays, grant applications and publishers, use the tool the limit-setter uses, usually Microsoft Word or Google Docs (Tools > Word count, or Ctrl+Shift+C), and do not rely on another counter agreeing to the last word.
Characters: with spaces, without, and what a character is
Character counts come in two flavours: with spaces, the usual meaning for limits such as meta descriptions and tweets, and without spaces, sometimes used for translation pricing. Check which one a limit means.
The deeper question is what counts as one character. Programming languages count differently from people:
const s = '👩💻'; // "woman technologist" emoji
s.length; // 5 UTF-16 code units (JavaScript's .length)
[...s].length; // 3 Unicode code points
[...new Intl.Segmenter('en', { granularity: 'grapheme' }).segment(s)].length; // 1 what users see
new TextEncoder().encode(s).length; // 11 bytes in UTF-8A form that validates maxlength with JavaScript's .length rejects text with emoji earlier than users expect. A database column limit may be in characters (PostgreSQL varchar(n), MySQL VARCHAR(n) with utf8mb4) or in bytes and UTF-16 units (SQL Server varchar and nvarchar). For user-facing limits, count graphemes; for storage limits, count what the storage counts.
Common content limits
SEO title tag ~50-60 characters (Google truncates by width, about 600 px)
Meta description ~150-160 characters on desktop (truncated by width)
X (Twitter) post 280 characters; links count as 23; many emoji and CJK count as 2
SMS 160 characters (GSM-7); 70 if any character needs Unicode,
such as an emoji; longer messages are split into 153/67-char parts
Instagram caption 2,200 characters
LinkedIn post 3,000 characters
App Store subtitle 30 charactersSearch engines truncate titles and descriptions by pixel width, not by a character count, so a title of wide capital letters is cut earlier than one of narrow lower-case letters. Treat the numbers as guides and check a preview. The SMS rule surprises people most: a single emoji or curly apostrophe switches the whole message to Unicode encoding, and the limit drops from 160 to 70 characters, so the message may be billed as two or three parts.
How reading time is calculated
Reading time is word count divided by an assumed reading speed. A 2019 meta-analysis by Marc Brysbaert found adult silent reading of English non-fiction averages about 238 words per minute. Medium uses 265 words per minute and adds time for images. Speaking is slower: presentations and podcasts run at roughly 130 to 160 words per minute.
const wordsPerMinute = 238;
const minutes = Math.max(1, Math.round(wordCount / wordsPerMinute));
// 1,200 words -> about 5 min readAdjust for content. Technical articles with code are read more slowly than narrative text, and readers skim lists and tables. Round to whole minutes; "4.6 min read" suggests a precision the estimate does not have.
Word counts are also converted into pages for manuscripts, essays and assignments. The result depends on font, size and line spacing, roughly 500 words per single-spaced page and 250 double-spaced in 12-point Times New Roman; a words to pages calculator applies those assumptions for you.
Using word count as a target
Word count is a constraint, not a quality signal. Google has repeatedly said it does not use word count as a ranking factor; pages rank when they answer the query well. A 300-word page that answers a narrow question completely can outrank a 3,000-word article that buries the answer. Use length limits to keep writing focused, such as a summary under 150 words or UI copy within the space available, rather than to pad content to a number.
The same applies to keyword counts. Repeating a phrase to hit a density target reads badly and is treated as spam; if you want to see which phrases dominate a draft, a keyword density checker shows the one-, two- and three-word phrases you use most, which is more useful for spotting accidental repetition than for chasing a percentage.
Sources and further reading
Marc Brysbaert, "How many words do we read per minute? A review and meta-analysis of reading rate", Journal of Memory and Language (2019). MDN documents Intl.Segmenter and string length in UTF-16. X's developer documentation describes weighted character counting, 3GPP TS 23.038 defines the GSM-7 SMS alphabet, and Google Search Central covers title links and snippets.
Frequently asked questions
Why do word counters give different numbers?
They define a word differently: hyphenated words, dashes without spaces, numbers, URLs and stand-alone symbols are counted differently by different tools. For strict limits, use the same tool as whoever set the limit.
Does character count include spaces?
Usually yes: limits such as meta descriptions, tweets and form fields count spaces. Some translation and publishing contexts use counts without spaces, so check which is meant.
How is reading time calculated?
Divide the word count by an average reading speed, commonly about 238 words per minute for adult silent reading of non-fiction, and round to whole minutes.
Why is an emoji counted as more than one character?
Many systems count UTF-16 code units or bytes, not what users see. A single emoji can be 2 to 11 units depending on the method; X counts most emoji as 2.
How many characters can an SMS have?
160 with the basic GSM-7 alphabet, but only 70 if any character needs Unicode, such as an emoji or curly quote. Longer messages are split into parts of 153 or 67 characters.
Does word count affect SEO rankings?
Not directly. Google says word count is not a ranking factor; what matters is whether the page answers the query well.