Keyword density is not a ranking factor: what to check instead

The advice is old, and it has outlived its time

You can still find tools that score your page for keyword density and tell you to raise it. The premise comes from early search engines that ranked pages by counting term frequency, which made repeating a phrase genuinely effective.

Those systems were replaced because repetition was trivially gameable. Search engines now assess whether a page answers the question a searcher asked, using language models that understand a passage rather than counting its words. Google's own published guidance has said for years that there is no ideal keyword density and that writing for it is not something to aim at.

If a tool tells you to repeat a phrase more often to rank better, the tool is describing how search worked before most of its users were online.

Why the advice survives anyway

Three reasons, and none of them is evidence. It is easy to measure, so it feels actionable. It gives a content team a number to hit, which is comforting when the real goal is unquantifiable. And the phrase persists in articles that are themselves written to rank for 'keyword density', which is a neat illustration of the problem.

Worse, it is actively harmful in the specific way it is most often applied: padding. A writer told to hit 2% density on a topic that only needs 400 words will produce 1,200 words of strained repetition, and both readers and search systems judge that page more harshly than a tight version of the same answer.

What a word count is genuinely good for

None of this means word counts are useless. They are useful for several honest things, just not ranking.

• Reading time, which sets reader expectation and affects whether someone starts at all.
• Meta descriptions and titles, where a hard character ceiling truncates anything that runs long.
• Checking you actually covered the brief. A 1,500-word brief answered in 300 words usually means a section was skipped.
• Measuring your own output over time, which is a workflow signal rather than a ranking one.
• Comparing depth against the pages already answering the same question, so you can see whether you have something to add.

So how long should a page be

As long as it takes to answer the question fully, and not longer. That is unsatisfying as advice and it is the honest answer.

The useful version of the question is not 'how many words' but 'what would make this page the best answer to this specific query, and have I included it?' Sometimes that is 600 words. Sometimes it is 3,000, because the question has four parts and each needs a worked example. The length is an output of that decision, never an input to it.

There is no length at which thin content becomes valuable. A 3,000-word page with no specific information ranks worse than a 500-word page that answers the question, and it wastes more of the reader's time doing it.

What to check instead of density

1. Intent match. Search the query yourself. What is the searcher actually trying to do — find a definition, compare options, complete a task? Does the page let them do it?
2. Coverage. Have you answered the obvious follow-up questions, or does the reader have to search again the moment they finish?
3. Specificity. Named numbers, worked examples, stated assumptions. Vague content cannot be checked, and content that cannot be checked does not build trust.
4. Structure, for a human. Headings that describe their sections, paragraphs short enough to scan, one idea per paragraph.
5. Accessibility of the language. If a reader has to re-read a sentence to parse it, the sentence has failed regardless of what it contains.

Notice that every one of those is judged by reading the page, not by running a tool over it. That is the point, and it is also why the density checker is still popular: it offers a number where the real work offers a judgement.

Where repetition does still matter

One place repetition genuinely damages you is at page scale rather than sentence scale. If a dozen pages on a site share the same template copy, headings and structure, the site as a whole looks repetitive — and that judgement is made about the site, not about any one page.

The fix is not density management. It is making each page specifically about its own subject, which is a content decision and not a word-frequency one. On this site, for example, every tool's guide heading names the tool and every partner recommendation carries copy written for that specific tool, precisely because identical blocks repeated across a site are the actual risk.

Count what is worth counting

The Word Counter gives you words, characters, sentences, paragraphs and estimated reading time, with the counts that have hard limits — titles and meta descriptions — separated from the ones that are informational. It will show you a keyword frequency figure if you ask, because seeing your own repetition is useful; it will not tell you what number to aim for, because there is not one.