seo
Keyword Density in 2026: What the Number Tells You (and What It Never Did)
There is no ideal keyword density, and Google never published one. What the ratio measures, how stop words skew it, and the four jobs it is still good for.
Somewhere in every SEO checklist there is a line that says “keep your keyword density between 1% and 3%.” Google has never published a target density, and its own docs say the opposite: there is no magic number, and repeating a word to hit one is the thing that gets flagged. The ratio is still worth looking at, though. Treat it as a smoke alarm for your draft, not something to aim for. This post covers what the number actually is, why two tools give the same page different numbers, and the four jobs it still does well.
The raw counts below were computed in Node 22 with the same tokenizer and phrase-counting code that runs in our Keyword Density Analyzer. Paste the sample paragraphs into it and you should see the same raw counts.
What Google actually says
There are two primary sources worth reading, and both are short.
The SEO Starter Guide has a section called “Things we believe you shouldn’t focus on.” Under keyword stuffing it says: “Excessively repeating the same words over and over (even in variations) is tiring for users, and keyword stuffing is against Google’s spam policies.” The same section says “there’s no magical word count target, minimum or maximum” for content length.
The spam policies define keyword stuffing as “filling a web page with keywords or numbers in an attempt to manipulate rankings.” Their examples include blocks of city names a page is trying to rank for, and “repeating the same words or phrases so often that it sounds unnatural.”
Neither page gives you a percentage, in either direction. Google’s Search Central channel even published a video in 2011 titled What is the ideal keyword density of a page?, with the question framed as “0.7%, 7%, or 77%?” Fifteen years later, the official guidance still names no number.
So where does “1–3%” come from? You’ll mostly meet it inside SEO tools. Yoast, for example, gives a green light when keyphrase density is between 0.5% and 3% (3.5% in Premium, which also counts word forms). That’s a plugin’s rule of thumb, not Google’s threshold.
What the ratio actually is
Keyword density is term frequency divided by total words:
density = (times the term appears / total words) × 100That looks simple. In practice, three choices change the answer for the same text.
1. What counts as a word. Our analyzer lowercases the text, deletes every character that isn’t a letter, digit, underscore or whitespace, then splits on whitespace. So “Don’t” becomes dont, “e-mail” becomes email, and a lone hyphen typed as a dash disappears. A plain word counter that only splits on spaces counts that hyphen as a word. On the sample below, our Word Counter says 63 words and the density tool says 62. Neither is wrong. They just define a word differently.
2. Stop words. “The”, “it” and “a” are always the most frequent words in English prose. If a tool removes them first, the denominator shrinks and every other term’s density goes up.
3. Stemming. Does “season”, “seasoning” and “seasoned” count as one term or three? Tools that stem (or, like Yoast Premium, match word forms) merge them. Tools that don’t, keep them apart.
There is also a fourth, quieter difference for phrases. Our tool divides a phrase’s count by the number of phrases of that length in the text (a 62-word text has 61 two-word phrases). Other tools count every word inside each match and divide by total words. In the stuffed example below, “cast iron skillet” is 16.7% our way and 47% that way (6 matches × 3 words ÷ 38 words). Same text, same phrase, very different percentage.
A worked example
Here is a natural paragraph about seasoning cast iron:
Season a cast iron skillet before its first use. Wash it, dry it on the stove, then rub a thin layer of oil into the pan. Bake the skillet upside down for an hour. Seasoning builds up with every cook, so the pan gets more nonstick over time. Don’t soak it, and don’t run it through the dishwasher - that strips the seasoning.
| Counting method | Words | Top terms |
|---|---|---|
| Raw (our tool, stop words kept) | 62 | the 6 (9.7%), it 4 (6.5%), skillet / pan / seasoning 2 each (3.2%) |
| Stop words removed | 35 | skillet / pan / seasoning 2 each (5.7%) |
| Stop words removed + crude stemming | 35 | season 3 (8.6%), skillet 2 (5.7%), pan 2 (5.7%) |
The topic word “seasoning” is at 3.2%, 5.7% or 8.6%, depending only on how you count. The text didn’t change. That’s the whole problem with a fixed target. As a ranking rule, “keep it under 3%” isn’t well defined until you say which of these three you mean.
The stop-word and stemming rows were computed with a simple filter for illustration. Our analyzer has a stop-words toggle, but its word list may not match the one used here, so your filtered numbers can differ a little. It doesn’t stem at all. With the toggle off, “the” will usually top the 1-gram table, so either switch it on or skip past the function words to the first content word.
Now the stuffed version:
Cast iron skillet seasoning is the best way to season a cast iron skillet. Our cast iron skillet seasoning guide shows cast iron skillet seasoning steps. Season your cast iron skillet with cast iron skillet seasoning oil today.
| Text | Words | Unique words | Top 3-gram |
|---|---|---|---|
| Natural | 62 | 49 (79%) | every phrase appears once (1.7%) |
| Stuffed | 38 | 19 (50%) | “cast iron skillet” 6 times (16.7%) |
Under every counting method above, the stuffed paragraph looks extreme. That’s the signal worth having. Not “is it 2.4% or 2.9%” but “one phrase is doing a sixth of all the talking, and half the words are repeats.”
What density is still good for
Spotting stuffing. If one 2-gram or 3-gram sits far above everything else, read those sentences out loud. Usually that’s a template, a repeated call to action, or a product name in every heading. The unique-word count next to the total is a quick second check. A low ratio means a lot of repetition.
Spotting a page that never names its topic. This is the opposite failure. A post about “invoice templates” that only ever says “this document” or “the form” leaves readers (and search engines) guessing. If your main phrase doesn’t appear in the top 25 at all, add it where it belongs, like the intro, a heading, or the first answer.
Comparing drafts. Run the old version and the new one. If an edit doubled the count of a phrase you didn’t mean to push, you’ll see it right away. The absolute number doesn’t matter. The change does.
Finding repeated phrases. The 2-gram and 3-gram tabs are good at catching verbal tics: “in order to”, “it is important”, the same transition phrase five times. That’s an editing win even if you never think about search.
What replaced it
The idea that took density’s place is coverage. Google’s starter guide says its “language matching systems are sophisticated and can understand how your page relates to many queries, even if you don’t explicitly use the exact terms in them.” The same guide notes that varying your wording gives you “more chances to show up in Search simply because you are using more keywords.”
In practice, that turns three old questions into new ones:
| Old question | Better question |
|---|---|
| Did I hit 2%? | Does the page answer what someone searching this phrase actually wants? |
| Is the keyword in there enough? | Are the related things (tools, steps, named products, places) that a full answer would mention actually there? |
| How many more times should I add it? | Did I name the topic clearly in the title, the intro and a heading, then just write? |
Search intent matters more than any of these. A page on “cast iron skillet” can be a recipe, a buying guide or a care guide. The density table can’t tell which one you wrote. Only the reader can.
A practical routine
- Write the draft without looking at any numbers.
- Paste it into the Keyword Density Analyzer. Check the 2-gram and 3-gram tabs for anything that towers over the rest, and make sure your main phrase shows up at all.
- Run it through the Readability Analyzer. Rewriting stuffed sentences usually makes them shorter, too. (Readability scores have their own limits.)
- Once it’s live, put the URL through the SEO Analyzer to check the title, headings, structured data and the rest. The full pre-publish list is in the on-page SEO checklist, and how our tools compare with others is in the free SEO tools roundup.
A note on privacy: the analysis runs in your browser and your text is never uploaded. The page sends a usage ping that records the tool’s name (along with the country, device type and referrer our server sees on any request), but none of your words. That makes it safe for a confidential draft.