Keyword Density: What the Percentage Means and What It Does Not

Density is usually the number of times a keyword appears divided by the total word count, expressed as a percentage. On a 1000-word page, ten occurrences are one percent. That arithmetic is fine as a descriptive statistic. It becomes a problem when it is treated as a control knob: force the percentage up and the prose becomes repetitive; force it down and you may remove the only clear statement of the topic.

Search systems evaluate topical relevance with far more than exact phrase counts. Synonyms, related entities, headings, links and user behaviour all matter. Historical advice that named a magic density band was never a stable rule, and pages written to hit a number usually read as if they were. Prefer covering the subtopics a reader expects over repeating the same string.

On a mixed Chinese and English page, total words is an ambiguous denominator. If Chinese is counted in characters and English in words, a single combined density is not comparable to either monolingual baseline. Measure density inside each language segment separately, or state clearly which unit the total uses. Otherwise a Chinese paragraph can look under-optimised simply because characters were treated as words.

Whether the counter folds plurals, tenses and stop words changes the percentage. Brand names and product codes often should be counted exactly; common verbs often should not dominate the list. Decide whether you care about an exact phrase or a family of related terms, and keep that decision consistent when comparing pages.

Density is a useful sanity check after drafting. Extremely high density on a short page is a signal to rewrite. Near-zero density on a page that claims to be about a topic is a signal that the topic is only in the title. Use it as a diagnostic, not as a goal to optimise toward.

A clear first mention in the introduction and a descriptive heading usually do more for comprehension and for search than three extra repetitions in the body. If the page cannot state its topic in one plain sentence near the top, density elsewhere will not rescue it.

Write the page for the reader first. Then run a density check and look only for extremes: accidental repetition, missing primary topic, or a bilingual denominator problem. Fix those. Do not rewrite sentences solely to move a percentage from 1.1 to 1.4.

A healthy page about UUID generation mentions versions, collisions, storage and misuse even if the exact phrase UUID generator appears only a few times. Density cannot see that coverage. After a density pass, list the subtopics a sceptical reader would expect and check that each has a sentence. That checklist prevents both keyword stuffing and hollow pages that repeat a title without teaching anything.

Searchers type UUID generator while good prose says generate a version 4 identifier or mint sortable IDs. Exact-match density will look low even when the page fully answers the query. Measure exact match only as a secondary signal. Primary quality is whether a newcomer can act after reading, not whether a phrase appears every hundred words.

Copying the density of a top-ranking page assumes that density caused the ranking. Those pages usually also have better links, fresher examples and clearer structure. Steal the outline and the questions they answer, not the raw occurrence count. When two pages cover the same subtopics, the clearer writer almost always wins over the one with more exact matches.