Check a live URL or paste a draft. Density for one- to four-word phrases, variants grouped, a verdict in plain words, and where each term sits on the page.
Most density tools give you a word, a count and a percentage. This one tells you what the percentage means, groups the forms of a word that belong together, and shows where on the page each term actually appears.
Every phrase from one to four words gets a count and a percentage of the words counted, and your target keyword gets a band in plain language. Under half a per cent is thin. Half to three per cent is healthy. Three to five is high. Above five is stuffed. The bands are relative to the page’s own word count, not to a universal target, and a page under 300 counted words is not banded at all, because the percentages are too small a sample to call.
“Search engine” and “search engines”, “rank” and “ranking”, “page” and “pages” each count as one term with the split shown beside it. Splitting them, which most tools do, makes your real use of a topic look weaker than it is and pushes it down the table.
Six columns beside every term show whether it appears in the title, the meta description, the H1, a subheading, image alt text or anchor text. A phrase used forty times in the body and never in a heading is a different problem from one used four times in the right places.
Above the table the tool prints the words it counted, split into body text and anchor text, against the words on the page, and says how many it excluded. Navigation, footers, sidebars, scripts and citation markers do not count. The number it counted for the scope you are looking at is the number every percentage is divided by.
Switch between all content and body only, which drops the anchor text and recalculates every percentage against the smaller count. Then read one-word terms, two-word, three-word or four-word phrases. The one-word view is a word frequency counter in its own right.
Each tab opens on the top 25 phrases by the current sort, with a Show all control when there are more, up to 500 per tab. Copy the table and you get the full set whether or not you expanded it, or download it as CSV, or share a link that opens the same check for someone else. The link runs the check again when it is opened, so it shows the page as it is then, not as it was.
Google’s spam policies name keyword stuffing as a practice that can get a page demoted or removed. They do not name a percentage. That is why this tool bands your density against the page itself rather than against a number nobody has published.
One phrase far above everything else, a high or stuffed band on the verdict, and ticks in every placement column because the same words were pushed into the title, the H1, the alt text and the anchors. Read the two- and three-word tabs as well: a page can look fine on single words and still repeat one exact phrase every second paragraph.
Cut the repeats where a pronoun or a natural variant reads better, and check that the grouped variants are carrying the topic rather than the exact phrase. Re-run the check on the live page afterwards, because the count that matters is the one on the page Google fetches, not the one in your editor.
A thin band is the more common result, and the quieter one. A page written about a topic can go 1,500 words without using the term people search for, and it is then ranking on a phrase it barely contains. The grouped row for your target keyword shows whether the topic is present under other forms or genuinely absent.
Density is one of the checks in the on-page SEO checker, which grades the whole page against the same keyword.
Not as a ranking factor. Google has said for years that it does not score pages on a keyword percentage, and nothing about how modern search reads a page suggests otherwise. If you came here for a target number to hit, there is not one, and any tool that gives you one is guessing.
It still matters as a diagnostic, and that is the only claim this page makes for it. Density is the fastest way to see what a page is actually about as opposed to what you meant it to be about, whether the term you want to rank for is present at all, and whether one phrase is doing so much of the work that the writing has started to show it. Those are questions about the page, not about the algorithm, and they are worth two minutes before you publish.
A 2 per cent density on a 300-word page and on a 3,000-word page are different things, and neither is good or bad on its own. The verdict bands against the page’s own count so that thin means thin for this page, and a page too short to band gets no verdict rather than a misleading one.
The checker takes any public URL, so a keyword density analysis of the page ranking above you is as easy to run as one of your own.
Run the top result for your query with your own target keyword in the field. The pinned row shows how hard they use the term and where; the rest of the table shows which other phrases carry the page.
Switch to body only. A competitor’s navigation and footer can hold hundreds of anchor-text words that have nothing to do with the article, and the all-content view counts them.
Same keyword, same scope, same tab. The comparison that matters is rarely the percentage; it is the placement columns, and whether they have the term in the places you do not.
We count the words a reader sees in the main body of the page, not every word in the raw HTML. Navigation, footers, sidebars and ads are excluded. Script and style content never counts. On pages with heavy citations or footnotes, a Wikipedia article for example, we strip the reference markers before counting, so a page dense with “[12][13][14]” does not get padded word and density figures.
Word count and density are only meaningful when both are measured the same way, so a tool that counts the full raw page, or does not strip citations, will report a different total from ours. Neither number is wrong. They answer slightly different questions, and this one answers the question a reader would.
The count that matters for the table you are looking at is the one printed above it for the scope you have selected. Switch from all content to body only and that number falls, because the anchor text has left along with the words it contributed, and every percentage in the table is recalculated against the smaller count. Every figure in the table can be reproduced by hand from the count above it.
Paste a public URL, or switch to Paste text for something that is not live yet. Both run the same analysis.
Optional, and worth doing. It pins that term to the top of every table with its rank among phrases of its length, and it is what the verdict is written about.
The band tells you whether the term is thin, healthy, high or stuffed on this page. The six columns beside it tell you whether it is in the places that count, which is usually the more useful answer.
A page that reads well on single words can still lean on one exact three-word phrase. Switch tabs before you decide anything.
Fixing a page is the easy half. Knowing whether it worked means watching the position, week after week, against every other page you have changed. RankCaddy tracks Google organic and local rankings daily, blends your Analytics and Search Console data into one view, and flags the pages quietly competing with each other.
Start your 14-day free trialFrom $24 a month for 250 daily keywords, with unlimited projects and users on every plan. No card needed for the trial.
Keyword density is the percentage of a page’s words made up of a given word or phrase. It is a rough signal of how much a page emphasises a topic, useful for spotting under- or over-use, but not itself a Google ranking factor.
Search engines treat close variants, plurals, verb forms and minor spelling differences, as the same underlying term. If we counted “rank” and “ranking” separately, your real use of that concept would be split across two rows and look weaker than it is. Grouping them gives you the number that reflects how much a topic is emphasised.
No. Google has never published a target percentage, and chasing one is a good way to make your writing sound unnatural. This tool bands your density as thin, healthy, high or stuffed relative to the page’s own word count so you can spot real outliers rather than chase a number that does not exist.
Paste the page’s URL or the text itself into the checker above, add the keyword you want it to rank for, and read the verdict at the top of the results. The table beneath it lists every phrase from one to four words with its count, its percentage of the words counted, and whether it appears in the title, meta description, H1, subheadings, alt text or anchor text.
Density is a measurement: how much of the page a term makes up. Stuffing is a judgement: using a term so often, or in so many places, that the page reads as written for a ranking rather than for a person. Google names stuffing in its spam policies and does not name a density figure, which is why this tool bands density against the page itself and flags the high and stuffed bands rather than quoting a percentage as a rule.
Not as a ranking factor, and keyword density in SEO has not been one for a long time. It still matters as a quick read of what a page is actually about, whether your target term is present at all, and whether one phrase is carrying more of the page than it should. Use it to check a page, not to tune one.
Not this one. This checker reads a page; it does not know where the page ranks. Position data over time is what RankCaddy’s rank tracking does, and it is a paid product with a 14-day free trial rather than a free tool.
No. Pasted text is analysed and the result is shown to you; we do not keep it. A shared link opens the tool and runs the check again rather than saving a copy of the result.