Content analysis
Use Keyword Density to Edit, Not Chase a Score
Repeated language can make a page clear, or make it exhausting. Colquhoun.shop’s Keyword Density analysis helps you inspect that repetition. The useful outcome is better writing, not reaching a supposedly perfect percentage.
Start with the text the tool actually read
The native page analysis counts words in text extracted from public HTML. Before interpreting the table, review that text. Navigation, repeated calls to action and other page elements may contribute to the counts. The result is not necessarily an isolated count of your main article.
Scripts do not run during this extraction. If an important section appears only after JavaScript runs, it may be absent. The reader also has a 50,000-character output limit and bounded parsing; a truncated result describes the returned text, not the entire page. These limits matter more than a small change in percentage.
Read counts and percentages together
The analysis produces single-word, two-word and three-word phrase lists. It normalizes case, retains common words such as “the” and “and,” and includes overlapping phrase occurrences. Each list shows up to 30 of the most frequent entries. An entry missing from the displayed list is not proof that it never appeared.
Each percentage uses the number of available positions for that phrase length. In a six-word sample there are six single-word positions, five two-word positions and four three-word positions. That means percentages from different phrase-length lists do not have identical denominators.
Consider this deliberately repetitive six-word sample: “local local local service helps businesses.” The single word “local” appears three times, giving 50%. The two-word phrase “local local” appears twice in five available positions, giving 40%. The phrase overlaps: the middle “local” belongs to both occurrences.
Neither number is a quality grade. It is a precise answer to a narrow counting question. Different tools may remove common words or use different denominators, so compare methods before comparing percentages.
Use repetition as an editing prompt
Choose one page and one reader goal. For a service page, that might be understanding who the service helps, what is included and how to take the next step. Examine frequent phrases against that goal.
- Necessary repetition: the service name or a technical term may need to appear several times for clarity.
- Template repetition: repeated navigation or buttons may explain a high count without revealing a problem in the main copy.
- Unhelpful repetition: several sentences may repeat a claim instead of supplying examples, limits or evidence.
Read suspicious passages aloud. Replace a redundant sentence with a useful detail, not simply a synonym chosen to lower the count. If “fast service” appears repeatedly, the better improvement may be a concrete description of the process rather than another adjective.
Make a controlled before-and-after comparison
Record the URL, date, returned word count, truncation status and the phrases you are investigating. Keep an authorized copy of the relevant text outside the application if you need an editing record. After the page is updated, check the same URL and verify that the returned text includes your changes.
Compare the actual wording as well as the numbers. Removing a long footer can change a phrase’s percentage even if the paragraph you edited remains untouched. A change in extraction scope can therefore masquerade as a writing improvement.
For a simple review, track three things: which passage changed, why the change helps the reader, and whether the revised passage still conveys the intended meaning. The density result is supporting evidence, not the decision-maker.
Do not turn a count into a promise
There is no ranking target built into this calculation. It does not measure keyword search volume, competitor visits, search position or the probability of an AI citation. A lower percentage is not automatically better, and a high percentage is not automatically a penalty.
Use the Colquhoun.shop dashboard to investigate language patterns, then make the editorial judgment yourself. For the broader workflow of reviewing a competitor’s messaging, see turning page text into a content review. Keep the distinction clear: observation can inform your writing, but copying a competitor’s wording or chasing their density is not a strategy.