Keyword density: what it measures and what a healthy range looks like
How keyword density is calculated, why the 'ideal percentage' advice is mostly folklore, and how to use density as a diagnostic rather than a target.
On this page5 sections
Keyword density is the share of your text made up of a given word or phrase. Say a term appears 12 times in a 900-word article: that is 12 divided by 900, or 1.3 per cent. The arithmetic is trivial. What people get wrong is what to do with the number.
The formula
density = (occurrences of the term ÷ total words) × 100
For multi-word phrases there is a subtlety: a three-word phrase occupies three of your words each time it appears, so some tools multiply the numerator by the phrase length. Whichever convention a tool uses, comparing two pages measured the same way is fine; comparing across tools is not.
What a healthy range looks like
There is no threshold that search engines reward. What exists is a range that naturally-written text tends to fall into, and edges outside it that signal a problem:
| Density | What it usually means |
|---|---|
| Under 0.3% | The page may not clearly be about the topic at all — often a sign of vague, abstract writing. |
| 0.5%–2% | Where most well-written, on-topic pages land without anyone trying. |
| 2%–4% | Noticeably repetitive. Read it aloud; you will hear it. |
| Over 4% | Reads as stuffed. Real readers notice before any algorithm does. |
Why the 'ideal percentage' advice persists
In the late 1990s, search engines really did weight raw term frequency, and stuffing worked. Everything since — TF-IDF weighting, then semantic embeddings — has moved ranking away from counting words toward modelling meaning. A page can rank for a phrase it never uses verbatim, and a page repeating a phrase 40 times can rank for nothing at all.
What to look at instead
- Coverage: do you address the sub-questions a reader would have? Missing sub-topics hurt more than missing keywords.
- Distribution: a term appearing 12 times in one paragraph and nowhere else suggests the article is only briefly on-topic.
- Natural variants: singulars, plurals, synonyms and related concepts. Writing that only ever uses one exact phrase reads mechanically.
- The obvious places: title, opening paragraph, one or two subheadings. Not because of a rule, but because that is where a reader checks they are in the right place.
- Stop words: 'the' will always top a raw frequency list. Any useful density report filters them out.
A practical workflow
- 1Write the piece without thinking about density at all.
- 2Run a density check and read the top 15 terms. Ignore the percentages at first.
- 3Ask: does this list describe the article? If a stranger read only these terms, would they guess the topic?
- 4If the answer is no, the problem is focus, not frequency — the fix is structural.
- 5If one term is above 3–4 per cent, replace half its occurrences with pronouns or variants and re-read for rhythm.
Done this way, density stops being a score to chase and becomes what it is actually good at: a quick, honest summary of what your text talks about most.
Was this guide useful?
Your vote is anonymous and helps us decide what to write next.
Written by
The WordCounter4U Team
Editorial
We build and document the tools on this site. Our guides state their methodology so you can check the numbers yourself.