Text Readability Scores Explained: Flesch Kincaid, Gunning Fog, and Coleman Liau for Writers

You have probably experienced this: you land on a web page, start reading the first paragraph, and immediately bounce because the sentences feel like they were written for a PhD defense committee. The information might be excellent, but the writing is so dense that your brain gives up before you finish the second line. That reaction is not a character flaw. It is a measurable, predictable outcome, and readability formulas have been quantifying it since the 1940s.

Readability scores give you a number that represents how easy or difficult a piece of text is to read. They look at things like sentence length, word length, syllable count, and character count to produce a single metric you can act on. Whether you are a copywriter trying to write landing pages that convert, a UX designer writing microcopy for a mobile app, or a student formatting a research paper, understanding these formulas will make your writing measurably better.

This guide breaks down every major readability formula, explains the actual math behind each one, walks through a real calculation by hand, and shows you how to pick the right readability target for your specific audience.

What Readability Actually Measures (and What It Does Not)

Readability is a prediction. It estimates how likely a reader with a given education level is to understand your text on the first pass. Every readability formula works by analyzing surface features of your writing: how long your sentences are, how many syllables your words contain, or how many characters make up each word. These proxies correlate surprisingly well with actual reading comprehension, which is why the US military, healthcare organizations, and legal agencies have relied on them for decades.

What readability does not measure is quality, accuracy, or persuasiveness. A perfectly readable text can still be boring, factually wrong, or poorly argued. A Flesch Reading Ease score of 80 means the text is easy to parse; it says nothing about whether the ideas inside are any good. Think of readability as a filter for accessibility. It answers the question: “Can my target audience physically process these sentences without re-reading them three times?”

The core variables that nearly every formula relies on are average sentence length (measured in words per sentence) and word complexity (measured by syllable count or character count per word). Longer sentences force readers to hold more information in working memory, and longer words tend to be more abstract or specialized. When you combine long sentences with complex vocabulary, comprehension drops off a cliff.

This is why tools like a Word Counter are more useful than people realize. Counting your words and sentences is literally the first step in calculating any readability score. If you know your text has 200 words and 10 sentences, you already know your average sentence length is 20 words per sentence, and you can start evaluating whether that number is too high for your audience.

A Brief History of Readability Formulas

The story of readability research starts earlier than most people think. In the 1920s and 1930s, educators in the United States began developing word lists and vocabulary studies to figure out which textbooks were appropriate for which grade levels. Researchers like Edward Thorndike published massive frequency counts of English words, cataloging which words appeared most often in printed material. The logic was straightforward: words that appear frequently are words that most people know.

Rudolf Flesch and the 1948 Breakthrough

Rudolf Flesch, an Austrian immigrant who became one of America’s most influential writing consultants, published his landmark readability formula in 1948. His paper, “A New Readability Yardstick,” appeared in the Journal of Applied Psychology and introduced what we now call the Flesch Reading Ease score. Flesch’s insight was elegant: instead of requiring a massive vocabulary database, you could predict readability using just two variables, average sentence length and average syllables per word. His formula was simple enough to calculate with a pencil and paper, which made it wildly practical for editors, teachers, and government agencies.

Flesch was also a fierce advocate for plain language. His 1946 book The Art of Plain Talk argued that most professional writing was needlessly complex, and that clarity was not a sign of simplistic thinking but of disciplined communication. His formula gave that argument teeth. Now you could point to a specific number and say, “This government form scores a 22 on the Flesch scale; no ordinary citizen can understand it.”

Robert Gunning and the Fog Index (1952)

Robert Gunning was a business consultant who worked with newspapers and corporations to improve the clarity of their publications. In 1952, he developed the Gunning Fog Index, which introduced the concept of “complex words,” defined as words with three or more syllables. Gunning’s formula estimates the years of formal education a reader needs to understand the text. A Fog Index of 12 means you need roughly 12 years of education (a high school senior) to comfortably read the material.

Gunning famously used his index to analyze major American newspapers and demonstrated that publications with lower Fog scores had significantly larger readerships. The Wall Street Journal, despite covering complex financial topics, consistently wrote at a Fog level around 11 to 12. Meanwhile, academic journals regularly hit Fog scores above 17, which partially explains why almost nobody reads them voluntarily.

Coleman and Liau’s Character-Based Approach (1975)

In 1975, Meri Coleman and T. L. Liau published a formula that took a different approach entirely. Instead of counting syllables (which is ambiguous and error-prone, especially for computers), they counted characters. Their Coleman-Liau Index uses the average number of letters per 100 words and the average number of sentences per 100 words. This made it much easier to implement programmatically, since counting characters is trivial for a computer, while counting syllables requires heuristic rules and a dictionary of exceptions. You can quickly get character-level data for any text using a Character Counter, which is exactly the kind of input the Coleman-Liau formula needs.

Flesch Reading Ease: The Formula and What the Scores Mean

The Flesch Reading Ease (FRE) formula is the most widely recognized readability metric in the English-speaking world. Microsoft Word has included it in its spell checker since the 1990s. Here is the actual formula:

FRE = 206.835 - (1.015 × ASL) - (84.6 × ASW)

Where:

• ASL = Average Sentence Length (total words ÷ total sentences)
• ASW = Average Syllables per Word (total syllables ÷ total words)

The result is a number, theoretically ranging from 0 to 100, though it can actually go below zero for extremely dense academic prose. Higher scores mean easier reading. Here is how to interpret the output:

The constants in the formula (206.835, 1.015, and 84.6) were derived through regression analysis on a large corpus of texts that had been independently rated for difficulty by reading comprehension tests. Flesch calibrated the weights so that the output would land on a 0 to 100 scale that intuitively maps to difficulty.

One important detail: the formula penalizes syllable density (ASW) much more heavily than sentence length (ASL). The coefficient for ASW is 84.6, compared to just 1.015 for ASL. This means that swapping out a few multi-syllable words for shorter synonyms will move your score more than shortening your sentences by a word or two. If your score is too low and you need to bring it up, start by replacing complex vocabulary.

Flesch-Kincaid Grade Level: Mapping Text to School Grades

The Flesch-Kincaid Grade Level formula was developed in 1975 by J. Peter Kincaid and his team under contract with the US Navy. The military needed a way to ensure that technical manuals were written at a level that enlisted personnel could actually understand. Instead of an abstract 0 to 100 scale, the Kincaid variant outputs a US school grade level.

FKGL = (0.39 × ASL) + (11.8 × ASW) - 15.59

The inputs are the same: average sentence length and average syllables per word. But the output is a grade number. A score of 8.2 means the text should be understandable to someone in the 8th grade (around 13 years old). A score of 14.0 implies you need a college sophomore’s reading level.

The US Department of Defense adopted the Flesch-Kincaid Grade Level as a mandatory standard. Department of Defense Instruction 5025.13 requires that most documents intended for a general military audience be written at or below an 8th grade level. This is not because military personnel cannot read at higher levels; it is because when you are reading a repair manual for an aircraft engine at 2 AM under stress, you need the text to be absolutely unambiguous.

For web writers, the Flesch-Kincaid Grade Level is often more actionable than the Flesch Reading Ease score because it maps directly to an audience. If you are writing a blog post for a general audience, aim for grade 7 to 9. If you are writing documentation for developers, grade 10 to 12 is acceptable. If your score comes back above 14, you have a problem regardless of your audience.

The Gunning Fog Index: Cutting Through Dense Writing

The Gunning Fog Index takes a slightly different angle. Instead of counting all syllables, it focuses specifically on “complex words,” which Gunning defined as words with three or more syllables. The formula is:

Fog Index = 0.4 × (ASL + PHW)

Where:

• ASL = Average Sentence Length (words per sentence)
• PHW = Percentage of Hard Words (words with 3+ syllables, expressed as a percentage of total words)

A few rules modify the “hard words” count. Proper nouns are excluded. Compound words made from simple words (like “nevertheless” or “basketball”) are sometimes excluded depending on the implementation. Verbs that become three syllables only because of conjugation (like “created” or “however”) are debatable, and different implementations handle them differently.

The output represents the years of formal education needed. A Fog Index of 8 means an 8th grader can read it. A Fog Index of 17 means you need a graduate degree. The ideal targets by publication type:

• Mass-market fiction and tabloid news: 6 to 8
• Quality newspapers (New York Times, Guardian): 10 to 12
• Business reports and white papers: 12 to 14
• Academic journals and legal contracts: 15 to 20

The Fog Index is particularly useful for identifying two specific problems: sentences that are too long, and vocabulary that is unnecessarily technical. If your Fog score is high but your sentence lengths are normal, you know the issue is word choice. If your sentence lengths are extreme but your vocabulary is simple, you know you need to break up your paragraphs. Analyzing your word choices with a Word Frequency Counter can reveal whether a small number of complex, repeated terms are dragging your Fog score up. If you find that one six-syllable technical term appears 40 times in a 2,000 word article, that single word is significantly inflating your complexity score.

Coleman-Liau Index: The Character Counter’s Formula

The Coleman-Liau Index solved a practical problem that plagued the syllable-based formulas. Counting syllables by computer is surprisingly hard. The word “area” has three syllables, but a naive algorithm might guess two. The word “business” has two syllables, but it has eight letters, making it look complex to a character-based approach. Every syllable-counting algorithm relies on heuristic rules (“count the vowels, subtract silent e’s, handle diphthongs…”) that fail on edge cases.

Coleman and Liau sidestepped this entirely:

CLI = (0.0588 × L) - (0.296 × S) - 15.8

Where:

• L = Average number of letters per 100 words
• S = Average number of sentences per 100 words

No syllable counting at all. Just letters and sentences. This makes the Coleman-Liau Index trivially easy to compute programmatically, which is why many automated text analysis tools prefer it. A Character Counter gives you the exact letter count you need, and a Word Counter gives you the word and sentence counts. From those three numbers, you can calculate L and S in seconds.

The output, like Flesch-Kincaid, is a US grade level. A Coleman-Liau score of 10 means a 10th grader should understand the text. The trade-off is that character count is a less precise proxy for word difficulty than syllable count. A short word like “flux” (four letters, one syllable) is harder than “together” (eight letters, three syllables), but the Coleman-Liau formula would rate “together” as more complex. In practice, these edge cases average out over a large enough text sample, and the formula performs comparably to syllable-based methods on texts of 200 words or more.

SMOG Index and Automated Readability Index: Two More Worth Knowing

SMOG Index

The SMOG (Simple Measure of Gobbledygook) Index was developed by G. Harry McLaughlin in 1969. McLaughlin specifically designed it for healthcare materials, where miscommunication can literally endanger lives. The formula is:

SMOG = 3 + √(number of polysyllabic words in 30 sentences)

“Polysyllabic” means three or more syllables. You take a sample of 30 sentences (10 from the beginning, 10 from the middle, 10 from the end of your text), count every word with three or more syllables, take the square root, and add 3. The result is, again, a grade level.

The SMOG Index tends to produce scores about two grade levels higher than Flesch-Kincaid for the same text. This is intentional. McLaughlin argued that the Flesch-Kincaid formula overestimates comprehension because it measures what a reader can struggle through, not what they can comfortably understand. SMOG targets 100% comprehension, while Flesch-Kincaid targets roughly 50 to 75% comprehension. For healthcare and safety-critical documents, you want the stricter standard.

Automated Readability Index (ARI)

The ARI, developed in 1967, is another character-based formula that avoids syllable counting:

ARI = (4.71 × characters/words) + (0.5 × words/sentences) - 21.43

It produces a grade level score. The ARI was originally designed for real-time readability monitoring on military teletypewriters, which is why it uses characters instead of syllables. You cannot count syllables on a teletype machine, but you can count every character as it prints. Today, the ARI remains useful as a quick sanity check alongside other formulas. If three different formulas all say your text is at grade level 12, you can be fairly confident that is accurate. If one says 8 and another says 14, you need to investigate why.

How to Calculate Readability by Hand: A Worked Example

Let us work through a real example. Here is a short passage (you would normally want at least 200 words for an accurate score, but this will illustrate the process):

“The committee reviewed the proposed amendments to the environmental protection policy. Several members expressed concerns about the implementation timeline. The revised document requires all manufacturing facilities to reduce emissions by thirty percent within five years.”

Step 1: Count the basics.

Use a Word Counter or count manually:

• Total words: 38
• Total sentences: 3
• Total syllables: 67 (the-com-mit-tee-re-viewed-the-pro-posed-a-mend-ments… you count each one)

Step 2: Calculate the averages.

• ASL (Average Sentence Length) = 38 ÷ 3 = 12.67 words per sentence
• ASW (Average Syllables per Word) = 67 ÷ 38 = 1.76 syllables per word

Step 3: Plug into the Flesch Reading Ease formula.

FRE = 206.835 - (1.015 × 12.67) - (84.6 × 1.76)
FRE = 206.835 - 12.86 - 148.90
FRE = 45.08

A score of 45.08 falls in the “Difficult” range (30 to 49), which means this text is appropriate for a college-level reader. That makes sense: the passage uses words like “amendments,” “implementation,” and “manufacturing,” which are all multi-syllable professional terms.

Step 4: Calculate Flesch-Kincaid Grade Level for comparison.

FKGL = (0.39 × 12.67) + (11.8 × 1.76) - 15.59
FKGL = 4.94 + 20.77 - 15.59
FKGL = 10.12

A grade level of 10.12, meaning a 10th grader (about 15 years old) should be able to understand this. For a government policy document, that is actually not terrible. For a blog post targeting general consumers, it is too high.

Step 5: Count complex words for the Gunning Fog Index.

Words with 3+ syllables: “committee” (3), “reviewed” (2, does not count), “proposed” (2, does not count), “amendments” (3), “environmental” (5), “protection” (3), “several” (3), “expressed” (2, does not count), “concerns” (2, does not count), “implementation” (5), “timeline” (2, does not count), “revised” (2, does not count), “document” (3), “requires” (3), “manufacturing” (5), “facilities” (4), “emissions” (3), “thirty” (2, does not count), “percent” (2, does not count).

Complex words: 10 out of 38 = 26.3%

Fog Index = 0.4 × (12.67 + 26.3)
Fog Index = 0.4 × 38.97
Fog Index = 15.59

A Fog Index of 15.59, which means this passage needs a reader with roughly 16 years of education (a college graduate). The Fog Index is higher than the Flesch-Kincaid because it harshly penalizes polysyllabic words, and our sample passage is loaded with them.

What Readability Scores Mean for SEO

Google has never publicly stated that readability scores are a ranking signal. There is no line in their algorithm documentation that says, “We run Flesch-Kincaid on every page and boost articles at grade 8.” But that does not mean readability is irrelevant to search performance.

Readability affects SEO indirectly through user behavior metrics. When someone clicks a search result and immediately bounces back because the content is too dense, Google records that as a short dwell time. When users consistently bounce from a page, it signals to Google’s ranking systems that the page is not satisfying user intent, and the page gradually drops in the search results.

Studies from SEMrush, Backlinko, and other SEO research firms have found a correlation between readability and ranking performance. Content that scores in the Flesch Reading Ease range of 60 to 70 tends to rank better for informational queries than content scoring below 30. This is not causation; it is correlation. But the mechanism is clear: readable content keeps people on the page, reduces pogo-sticking (jumping back to search results), and increases the chance of earning backlinks (because people share content they actually understand).

Google’s own writing guidelines for Search Quality Raters emphasize “Expertise, Authoritativeness, and Trustworthiness” (E-E-A-T). Part of trustworthiness is accessibility. A medical article that scores 25 on Flesch Reading Ease might be expertly written, but if ordinary patients cannot understand the health advice, it fails the “helpful for the audience” test.

For practical SEO purposes, here is the process: write your content, then run it through a text readability checker. If your Flesch Reading Ease is below 50 and you are targeting a general audience, revise. Shorten your longest sentences. Replace multi-syllable jargon with simpler alternatives where the meaning does not change. Use the Clean Text tool to strip out any formatting artifacts or extra whitespace that might have crept in during editing, and then re-check your score.

Ideal Readability Targets by Audience and Content Type

Not all content should aim for the same readability level. A children’s book and a legal brief serve entirely different audiences, and treating them with the same readability target would be absurd. Here are concrete recommendations based on the content type:

Blog Posts and Marketing Content

Target a Flesch Reading Ease of 60 to 70 (grade 7 to 9). This covers the largest segment of the English-reading population. Marketing copy, in particular, should err on the side of simplicity. You are competing for attention with every other tab in the reader’s browser. Every unnecessary complex word is friction. Every 40-word sentence is a potential exit point.

News Articles and Journalism

Target a Flesch Reading Ease of 55 to 65 (grade 9 to 11). Newspapers have been optimizing for readability since the 1950s, and the best ones hit this range consistently. The Associated Press Stylebook implicitly encourages this level of clarity, even though it does not directly reference readability scores.

Academic Papers and Technical Documentation

Target a Flesch Reading Ease of 30 to 50 (grade 12 to 16). Here, precision matters more than accessibility. If the accurate term is “endoplasmic reticulum,” you cannot replace it with a simpler word. But you can still control sentence length. Even in academic writing, sentences over 35 words start causing comprehension issues regardless of the reader’s expertise.

Social Media and Mobile Content

Target a Flesch Reading Ease of 70 to 80 (grade 5 to 7). People read social media on tiny screens while waiting in line at the grocery store. Their attention span is measured in seconds, not minutes. Short words, short sentences, short paragraphs.

Healthcare and Legal Documents for Public Use

Target a Flesch Reading Ease of 65 or higher (grade 6 to 8). The American Medical Association recommends that patient-facing health materials be written at a 6th grade reading level. Given that nearly half of American adults read below a 6th grade level according to NCES data, this is not an arbitrary standard; it is a public health imperative.

How AI Engines Evaluate Text Complexity When Citing Sources

Large language models like those powering ChatGPT, Perplexity, and Google’s AI Overviews are trained on massive corpora of text spanning the full difficulty spectrum. When these systems generate answers and select sources to cite, they are implicitly making readability judgments, even if they do not run a Flesch formula under the hood.

AI systems tend to prefer citing sources that explain concepts clearly and completely. A well-structured article with clear headers, short paragraphs, defined terminology, and logical flow is more likely to be surfaced by an AI overview than a dense academic paper with 50-word sentences. This is partly because the AI’s own training optimizes for generating coherent, accessible responses, and it is easier to extract and paraphrase information from clearly written source material.

For content creators, this means optimizing for readability is not just about human readers anymore. It is about making your content machine-parseable and citation-worthy. Write encyclopedic, factual paragraphs that can stand alone as answers to questions. Use precise definitions. Structure your content with H2 and H3 headers that signal topic shifts. These practices help both human readers and AI retrieval systems.

Vocabulary diversity also plays a role. If your text repeats the same key terms excessively, AI systems may view it as thin or keyword-stuffed. On the other hand, using too much varied vocabulary for the same concept can confuse the extraction process. The sweet spot is natural language with moderate repetition of core terms and varied supporting vocabulary. Running your draft through a Word Frequency Counter can show you whether your keyword distribution looks natural or if certain terms are over-represented.

Practical Tips for Improving Your Readability Score

Knowing the formulas is only half the battle. Here is how to actually improve your scores without dumbing down your content:

1. Break long sentences. Any sentence over 25 words is a candidate for splitting. Look for conjunctions (“and,” “but,” “which”) as natural breaking points. Two 15-word sentences are almost always clearer than one 30-word sentence.

2. Replace complex words when a simpler synonym exists. “Utilize” becomes “use.” “Commence” becomes “start.” “Approximately” becomes “about.” These substitutions cost you nothing in precision and gain you several points on every readability formula.

3. Use active voice. “The report was reviewed by the manager” (8 words, passive) becomes “The manager reviewed the report” (6 words, active). Active voice naturally produces shorter, more direct sentences.

4. Strip unnecessary qualifiers. Words like “very,” “really,” “quite,” “somewhat,” and “relatively” add syllables without adding meaning. Cut them ruthlessly.

5. Read your text out loud. If you stumble over a sentence while reading aloud, your readers will stumble too. Your ear is a surprisingly good readability checker.

6. Prepare your text before analysis. Before running any readability formula, make sure your text is clean. Extra line breaks, stray special characters, and formatting artifacts can throw off word and sentence counts. The Clean Text tool strips all of that out in one pass, giving you an accurate baseline to work from.

7. Test on multiple formulas. No single formula tells the whole story. Run your text through Flesch, Fog, and Coleman-Liau. If all three flag the text as too complex, you have a genuine readability problem. If only one does, investigate why. It might be a syllable-counting edge case or an unusually long paragraph that skews one metric.

Frequently Asked Questions

What score should I aim for? That depends entirely on your audience. For most online content aimed at a general audience, a Flesch Reading Ease between 60 and 70 hits the sweet spot. For academic work, 30 to 50 is normal and expected. For children’s content or public health materials, aim for 70 and above.

Are readability scores accurate for non-English text? The formulas discussed in this article were calibrated for English. They rely on English-specific relationships between syllable count, word length, and comprehension difficulty. Some researchers have adapted formulas for Spanish, French, German, and other languages, but you should not apply the standard Flesch or Fog formulas to non-English text and expect meaningful results.

Can I game readability scores by just shortening everything? You can artificially inflate your Flesch score by writing nothing but five-word sentences with one-syllable words, but the result will read like a children’s primer and will not be appropriate for adult audiences. Readability scores are a guide, not a god. Use them to identify problem areas, then apply human judgment to decide what to fix.

Do readability formulas account for formatting, images, or layout? No. Every formula discussed here analyzes text only. They do not consider bullet points, subheadings, white space, font size, images, or any other visual element that affects real-world readability. A wall of text at Flesch 65 is harder to read than the same text broken into short paragraphs with subheadings, even though both score identically. Readability formulas measure linguistic complexity, not visual design.

Stop guessing whether your content is too complex for your audience. Count your words and sentences with the Word Counter, check your character distribution with the Character Counter, and start making data-driven editing decisions today.