What is The Ideal Keyword Density for SEO?

There is no ideal keyword density for SEO. Google has never published a target percentage, and the range you keep reading is a WordPress plugin’s scoring threshold rather than a rule from any search engine. Counting your keyword mentions is not what decides whether a page ranks.

That answer is unsatisfying when you have a draft open and an orange bullet telling you the keyphrase appears too few times. So it is worth tracing where the number came from, because once you know that, the bullet stops reading as a verdict.

What is keyword density, and how is it calculated?

Keyword density is how many times a keyword appears in a page divided by that page’s total word count, expressed as a percentage. Twelve mentions of a phrase in a 1,200-word article works out at one in a hundred, which is where a figure like “one per cent” comes from.

The formula is simple. What it measures is not, because two people counting the same page will disagree about nearly every input: whether headings count, whether the navigation and footer count, whether “keyword” and “keywords” are the same word, and whether a two-word phrase counts as one hit or two.

Nobody ever standardised any of that. Density was a reasonable proxy in the late nineties, when search engines really did rank documents partly on how often a term appeared on them, and it survived as a habit long after the ranking systems it described were retired.

Where does the keyword density percentage everyone quotes come from?

The 0.5% to 3% range that dominates every answer on this question comes from Yoast SEO, the WordPress plugin, where it is the band that turns the keyphrase density bullet green. Yoast documents that range openly as its own scoring rule, not as guidance from Google.

It is one of the most installed plugins on the web. Every user sees a traffic-light verdict on a number, a fair share of them go looking for what the number is supposed to be, and the answer that comes back is Yoast’s own threshold, repeated by people who no longer remember where they picked it up. That is the whole chain, and it explains why the figure is quoted with such confidence by sources that never name one.

The counting rules are worth knowing if you plan to chase that bullet anyway. All the words of your keyphrase have to land in the same sentence, though not in the order you typed them, and Yoast expects a longer keyphrase to appear less often than a short one. The plugin is not measuring the thing most people assume it is measuring.

Has Google ever published a keyword density target?

No. Not in the documentation, not in the spam policies, and not from anybody who has spoken for Google in twenty years of being asked.

Matt Cutts, who ran Google’s webspam team at the time, took the question directly in a 2011 webmaster video. His answer was that the first mention or two of a word genuinely helps Google work out what a page is about, and that you hit diminishing returns quickly after that. He also told people to be sceptical of anyone selling them keyword density software. Fifteen years later, the position has not moved.

Asked on Reddit at the end of 2021 whether keyword density was still a ranking factor, John Mueller answered with a single word: no. Search Engine Roundtable recorded that exchange in January 2022.

Why do two tools report different keyword density for the same page?

Because they count different things and disagree about the denominator. Run one page through two density checkers and a gap of half a percentage point between them is ordinary, not a bug in either of them.

There are two counting styles you are likely to meet:

• Phrase counting - Yoast and most content editors look for your whole keyphrase and score each occurrence as a single hit, however many words that phrase contains.
• Single-word counting - Frequency tools split the text into individual words and report each word’s share of the total, so a two-word phrase never appears as a row at all.

The Word Frequency Analyzer by New SEO Tools works the second way, and its percentage column is always calculated against every word in the text you pasted, stop words included, even when you switch on the filter that hides them from the list. It also keeps “tool” and “tools” as separate rows with the shared lemma in its own column, so you can see whether you used one form ten times or two forms five times each.

Neither number is wrong. They answer different questions, which is a decent reason to stop treating either one as a score you have to hit.

What actually counts as keyword stuffing?

Keyword stuffing is filling a page with keywords or numbers to manipulate rankings, and Google’s definition of it carries no threshold whatsoever. The spam policies give three examples: lists of phone numbers with nothing of value around them, blocks of text listing the cities a page wants to rank for, and repeating words or phrases so often that the writing sounds unnatural.

Read that third one again. The test is how the sentence sounds, not what a calculator says about it. You can trip it comfortably inside a band a plugin calls green, and you can sit well above that band without tripping it at all, when the repetition is what the subject genuinely demands.

An article about redirects is going to say “redirect” a lot. That is the word for the thing. Nobody at Google is confused by it.

The real failure is the sentence that got bent out of shape to fit a phrase in. “Our free SEO tools online help you find free SEO tools fast with free SEO tools” is what stuffing sounds like, and spotting it takes no software at all. Reading the paragraph out loud is the entire check. The Content Readability Checker makes a reasonable second pass, because it scores sentence length, passive voice and grade level, and prose that has been mangled to hit a keyword target tends to show up as sentences noticeably clumsier than the ones around them.

Is there a word count Google prefers?

No, and this one Google answers in a sentence you can quote back at people. Its guidance on creating helpful content asks whether you are writing to a particular word count because you heard Google has a preferred one, then answers its own question in brackets: no, we don’t.

Density and word count are the same mistake in different clothes. Both take a number that happens to describe good pages and treat it as the thing that made them good. Long pages often rank because a thorough answer needs the room, not because length earns anything on its own.

What should you check instead of density?

Check placement and coverage, since both are things you control and both change what a page is actually about. There are 4 checks worth more than any percentage:

• Does the phrase appear at all - More pages fail this than you would expect. When the exact wording someone types never appears in your text, Google is left to infer it.
• Is it in the title and the first paragraph - Position carries weight that repetition does not.
• Is one word dominating for no good reason - A frequency list where your brand name outranks every noun in the article is a page written for the wrong reader.
• Are the related terms there - A page about redirect chains that never mentions status codes, hops or loops reads thin at any density.

How do you check word frequency on a page you have already published?

Copy the body text of the page, paste it into a frequency tool, and read the top twenty rows. The Word Frequency Analyzer ranks every word by how often it occurs, gives each one’s share of the text and its part of speech, and redraws the same list as a bar chart and a word cloud.

Paste the article only. Your header, footer and navigation belong to the template rather than the page, they repeat everywhere, and leaving them in inflates the denominator while burying the words you care about under “home”, “about” and “contact”.

Then read the list as a description of your page instead of a score. If the top nouns are the subject you meant to write about, you are finished. If a filler word you never noticed using is sitting at rank three, that is worth an edit for the reader’s sake, and it has nothing to do with rankings.

Final Thoughts

Stop measuring density. Write the page for the person who searched, put the phrase where it belongs, and let the count land wherever it lands. If a plugin hands you an orange bullet for a page you would happily give a reader, the bullet is the thing that is wrong.