Keyword density is a diagnostic, not a target
There is no magic percentage that ranks a page. Keyword density as a ranking factor died years ago — modern search engines model topics and meaning through language models, not term frequency. Anyone selling you a "2.5% optimal density" is selling 2005.
What density analysis is genuinely useful for is diagnosis. It reveals things you can't see while writing: that your "ultimate guide to CSS Grid" mentions grid exactly twice, that one filler word dominates your copy, or that you've unconsciously used four different terms for the same concept — which dilutes the topic signal and confuses readers as much as crawlers.
What a natural density looks like
Well-written long-form content usually lands between 0.5% and 2% for its main phrase without any deliberate effort. If yours is far below that, the page may genuinely not be about what you think. If it's far above, it will read badly — and awkward repetition is far more likely to cost you readers than to gain you rankings.
Rather than hitting a number, aim for natural coverage: the phrase in the title, once in a heading, in the opening paragraph, and wherever it genuinely fits in the body. Then use related terms and synonyms freely — search engines understand them, and readers prefer them.
Reading time and matching search intent
The reading-time figure here uses roughly 200 words per minute, the accepted average for adult silent reading of general web content. It's useful for two reasons: it's what readers see on Medium-style layouts and it sets expectations, and it's a proxy for whether your content depth matches the query.
That match matters more than length itself. Someone searching "what is a rem unit" wants a two-minute answer; someone searching "css grid complete guide" expects fifteen. Publishing a 4,000-word essay for a quick-answer query performs badly, and so does a thin page for a research query. Check what currently ranks for your target phrase and aim for comparable depth.
Reading the most-used words table
The table excludes stop words and very short terms, so what remains is the vocabulary your page actually emphasises. Read it as a summary of your topic — if a stranger saw only that list, would they know what the page is about? If the top entries are generic ("things", "make", "good"), the writing is probably vaguer than it should be.
It's also a useful check for accidental repetition. Seeing the same verb eight times usually means the prose needs varying, which is a readability win regardless of what any search engine thinks.