Keyword Frequency vs Keyword Density: What's the Difference?

Two numbers get used interchangeably in SEO advice and they measure genuinely different things. Confusing them is how people end up chasing a "2% density" target that means nothing on a 300-word page and something entirely different on a 3,000-word one.

Keyword frequency is a raw count. Keyword density is that count expressed as a percentage of total words. Frequency answers "how many times?"; density answers "how concentrated?". For the wider context on how word counts underpin every one of these measures, see The Complete Guide to Counting Words and Characters.

The Two Formulas

Frequency needs no formula — it is the number of occurrences of a term in a document. If "cold brew" appears 14 times, its frequency is 14.

Density normalises that count against document length:

density (%) = (keyword occurrences ÷ total words) × 100

Fourteen occurrences in a 1,400-word article gives a density of 1.0%. The same fourteen occurrences in a 400-word article gives 3.5% — identical frequency, wildly different concentration, and only the second one reads as repetitive.

Multi-word phrases add one wrinkle that most people get wrong. Because the denominator counts individual words, the numerator has to as well:

phrase density (%) = (occurrences × words in phrase ÷ total words) × 100

"Cold brew coffee" used 14 times in a 1,400-word article is (14 × 3) / 1400 × 100 = 3.0%, not 1.0%. Tools disagree on this — some apply the multiplier, some do not — which is the single most common reason two analysers report different densities for the same text. When you compare numbers across tools, check which convention each one uses before concluding anything.

Worked Example

Take a 1,200-word article about espresso machines:

Term	Frequency	Density (single-word denominator)
espresso	30	2.5%
machine	24	2.0%
espresso machine	18	3.0% (18 × 2 ÷ 1200)
portafilter	6	0.5%
grinder	4	0.33%

The frequency column tells you what the article is about. The density column tells you whether the article is overdoing it. Read together, they say the piece is solidly on-topic but leaning hard on one phrase — 3.0% for "espresso machine" is high enough that a read-through will almost certainly find sentences where the phrase was inserted rather than written.

When Frequency Is the Better Number

Frequency is the right tool whenever you care about what is in the text rather than how much.

Topic coverage audits. Ranking a document's terms by frequency, with stop words filtered out, produces a rough summary of its subject matter in about two seconds. If you wrote a guide to container gardening and "soil" appears twice while "pot" appears forty times, the frequency list has told you which section is thin.

Finding unintentional repetition. Crutch words are invisible while you write and obvious in a frequency list. Most drafters have one — "actually", "leverage", "essentially", "importantly". Seeing it at rank three is more persuasive than any style advice.

Comparing sections, not documents. Because frequency is not normalised, it is only meaningful within a fixed length. That is exactly what you want when checking whether a keyword appears in the introduction, or whether one section carries all the weight.

A Keyword Frequency Analyzer gives you the ranked list directly, including two- and three-word phrases, which surfaces patterns single words hide.

When Density Is the Better Number

Density is the right tool whenever you need to compare across documents of different lengths.

Benchmarking against competitors. Comparing your 900-word page to a rival's 2,600-word page by raw frequency is meaningless. Density makes them comparable.

Setting a repetition ceiling. Density is a reasonable proxy for how repetitive a piece feels. Above roughly 3% for one term, most readers notice.

Checking a template at scale. Product or location pages generated from a template often have wildly inconsistent lengths. Density catches the short ones where a boilerplate phrase now dominates.

The Keyword Density & Frequency Analyzer reports both figures side by side, which is usually what you actually want — neither number is very useful alone.

What Density Is Not

Density is not a ranking factor, and it has not been one for a long time. Google has never published a target density, and its spam policies name keyword stuffing — "repeating the same words so often that it sounds unnatural" — as a violation. Search systems weight terms using approaches like BM25, which applies saturating term frequency: the tenth occurrence of a word contributes far less than the second, and the fiftieth contributes almost nothing. Repetition has diminishing returns built into the maths.

So the useful mental model is a floor and a ceiling, not a target:

• Floor. If your main term appears zero or one time in a 1,500-word article, the page may genuinely not be about what you think it is. Fix that.
• Ceiling. Above about 3%, read it aloud. If any sentence exists to hold the keyword, cut it.
• Everything between is noise. Chasing 1.8% instead of 1.2% is not a use of anyone's afternoon.

The other reason to distrust a single density figure: most tools match exact strings, so "bike", "bikes" and "biking" count as three unrelated terms. A page can look under-optimised at 0.4% while covering the topic thoroughly across variants. This is one place a frequency list beats a percentage outright — you can see the variants and add them up yourself.

A Practical Workflow

Finish the draft first. Density calculated on a partial draft is meaningless because the denominator is still moving, and writing toward a percentage produces exactly the stilted prose the metric is supposed to detect.

Then run three checks. Get the total word count from a Word Counter so you know the denominator and can confirm the piece is near the right length for its format. Pull the ranked frequency list and read the top fifteen terms — if that list does not describe your article, the structure is wrong, not the density. Finally check density on your main term and its two closest variants, looking only for the floor and ceiling cases above.

For a broader pass, a Text Statistics Summary puts word count, sentence length and term distribution in one view, which is where repetition problems and readability problems usually turn out to be the same problem. And since all of these run entirely in your browser, an unpublished draft or a client's content brief never leaves your machine.

The Short Version

Frequency counts. Density normalises. Use frequency to understand what a document covers and where it repeats itself; use density to compare documents of different lengths and to catch stuffing. Treat 0.5–2% as the range natural writing tends to land in, not as a number to aim for — and if you find yourself editing a sentence to move the percentage, you are optimising the wrong thing.