Keyword Density, Explained

After reading this you can measure how often a word or phrase appears in a page, express that as a density percentage, and judge whether the result is a useful diagnostic or a distraction.

What keyword density measures

Keyword density is the share of a document's words that a specific term takes up. If a 500-word article uses the phrase "cold brew" 10 times, that phrase accounts for 20 of the 500 words (each occurrence is two words), so its density is 20 divided by 500, or 4%.

The number tells you something narrow but real: what the text is literally about, in raw counts, once you strip out grammar words like "the" and "of". It does not tell you whether the page will rank. Search engines moved past raw term frequency as a primary signal many years ago. Treat density as a mirror that shows you the words you actually wrote, which is often different from the words you think you wrote.

Here is the hook. Writers routinely believe they covered a topic thoroughly, then run the text and find the main term appears twice while a filler word appears fifteen times. The checker settles that argument with counts.

When to use it, and when to skip it

Reach for a density check in three situations. First, when auditing a page that targets a specific query and you want to confirm the query terms are present at all. Second, when a draft feels repetitive and you want to see which word is causing it. Third, when comparing your text against a competitor's on the same topic to spot vocabulary you left out.

Skip it as a writing target. There is no threshold that triggers a ranking boost. Aiming for a fixed percentage such as "2% for the main keyword" produces stiff, padded prose and wastes your time.

The formula and the intuition

Density for a single word is its count divided by the total number of counted words, times 100.

Here f is the frequency (how many times the term appears), N is the total number of words counted, and D is the density as a percentage. For a phrase of m words, each occurrence still counts as one phrase instance, but the denominator N stays as the total word count, so a two-word phrase appearing 10 times in 500 words gives D = 10/500 \times 100 = 2\% as a phrase, even though it consumes 20 individual word slots.

The important subtlety is N. This tool filters stopwords before counting, so N is the number of meaning-carrying words, not every word on the page. That choice raises every density figure compared to counting raw words, because the denominator shrinks. Always know which N you are dividing by before comparing two tools.

A worked example you can reproduce

Reading the phrase lists

Single words tell you the topic. Two- and three-word phrases tell you the framing. A page about running shoes that shows running shoes and trail running high in the two-word list is on topic. If the top phrase is click here or read more, your visible text is thinner than you thought, because navigation and calls to action are outweighing the substance.

Phrase counts are always lower than the counts of their component words. In the sample, quick fox appears once even though fox appears 4 times, because only one of those appearances sits directly after quick. Do not expect phrase densities to add up to word densities; they measure adjacency, not membership.

Common mistakes

The most frequent error is chasing a target percentage. Writers read a blog post claiming "keep your keyword at 1 to 2%" and edit until the number lands there. The claim has no basis in published ranking documentation, and the edits usually make the prose worse.

The second mistake is counting HTML noise. If you paste a full page source without stripping tags, class names and script contents can leak into the counts. This tool strips tags, scripts and styles before counting, so its numbers reflect visible text only. If another tool gives you wildly different results, check whether it counted the markup.

The third mistake is ignoring word forms. run, runs, running and ran are counted separately because they are different strings. A topic can be well covered while no single form scores high. Read the list as a family of related terms, not one line item.

How density behaves as text grows

Density is a ratio, so adding more unrelated text drives any single term's percentage down even when its count is unchanged. Suppose fox appears 4 times. In 15 counted words the density is 26.67%. Add 35 more counted words on other subjects without adding fox, and N becomes 50, so density falls to 4/50 \times 100 = 8\%. The word appears exactly as often; the page just says more.

This is why long, thorough articles show low densities for their main term and why that is fine. A 2000-word guide with a main phrase at 0.5% is not under-optimized; it simply has room for everything the topic needs.

Related tools

Density looks at the words in your text. To check the other half of page quality, how heavy the page is to load, use the Page Size & Request Analyzer, which reports page weight, request counts and the heaviest resources. A page can have ideal vocabulary and still fail visitors if it takes eight seconds to load.

Related tools

Page Size & Request Analyzer

Page weight, request counts and the heaviest resources.

Meta Tag & Open Graph Generator

Generate meta tags and preview the Google, Facebook and X cards.

Redirect Chain Checker

Follow every hop: 301s, 302s, loops and chain length.

SERP Snippet Preview

Pixel-accurate preview of your Google result on desktop and mobile.

UTM Campaign URL Builder

Build tagged campaign URLs with correct encoding.

robots.txt Tester

Test which URLs a robots.txt blocks, rule by rule.