You paste your article into a readability tool and get a Flesch score of 58. Is that good? Another tool gives you a Fog Index of 14. Should you panic? A third shows a Coleman-Liau of 11. What does any of this mean, and does it actually matter?
Most content about readability scores tells you to "aim for 60–70" and leaves it there. That's about as useful as telling someone to "score well" on an exam. This guide goes deeper: what each formula actually measures, why they often disagree, which one to use for your specific content type, and — most importantly — how to improve your writing without making it worse by chasing numbers.
Why Readability Metrics Exist
Readability formulas were invented in the 1940s–1970s to solve a specific problem: government agencies and publishers needed a fast, consistent way to check whether documents were accessible to their intended readers — without having to recruit test panels for every piece of content.
Rudolf Flesch developed his Reading Ease formula in 1948 after studying how everyday Americans read. The US military used it to simplify technical manuals. Schools used early versions to match books to grade levels. Insurance companies used it to rewrite policy documents.
The formulas were always imperfect proxies. They measure surface features of text — sentence length and word length — as stand-ins for the much harder-to-measure quality of actual comprehension. A sentence can be short and still be confusing. A long word can be perfectly familiar to its audience.
Readability formulas measure the statistical properties of text, not actual comprehension. They don't understand meaning, context, audience knowledge, or whether your argument is logical. They are useful diagnostic tools, not quality scores.
Flesch Reading Ease: The Most Intuitive Score
Flesch Reading Ease produces a score from 0 to 100. Higher is easier to read. It's the most widely used readability formula for a simple reason: the output is immediately intuitive.
Two variables drive the score: average sentence length (words per sentence) and average word length (syllables per word). Shorter sentences and simpler words both push the score up.
| Score | Difficulty | Typical Audience | Examples |
|---|---|---|---|
| 90–100 | Very easy | 5th grade | Children's books, comics |
| 80–90 | Easy | 6th grade | Conversational writing |
| 70–80 | Fairly easy | 7th grade | Consumer content |
| 60–70 | Standard | 8th–9th grade | Most web content, Time Magazine |
| 50–60 | Fairly difficult | 10th–12th grade | Business writing |
| 30–50 | Difficult | College | Academic papers, legal docs |
| 0–30 | Very difficult | College grad | Scientific journals |
For most web content, 60–70 is the sweet spot. Below 50 means you're likely writing for a narrower audience than you intend. Above 80 is perfectly fine for simple how-to content, but might feel condescending in a professional context.
Flesch-Kincaid Grade Level: The School System Version
Flesch-Kincaid Grade Level uses the same two variables — sentence length and syllables per word — but produces a US school grade level instead of a 0–100 score.
A grade of 8 means a US 8th grader (roughly 13–14 years old) could read it comfortably. A grade of 12 means a high school graduate. A grade of 16 corresponds to a college degree.
Notice the large weight on syllables per word (11.8 vs 0.39 for sentence length). This formula is dominated by vocabulary difficulty. Using simpler words has a much bigger impact on your FK grade than shortening sentences.
Grade 7–9 is ideal for general-audience web content. The UK government's style guide targets Grade 9. The Plain English Campaign recommends Grade 6–8 for public communications.
Gunning Fog Index: The Business Writing Standard
Robert Gunning developed the Fog Index in 1952 for business writing. His insight was that "foggy" writing obscures ideas with unnecessarily long words and tangled sentences.
Complex words = words with 3+ syllables (excluding proper nouns and common compound suffixes)
The output is years of formal education required to understand the text on first reading. A Fog score of 12 = high school graduate can read it. Most major newspapers target Fog 10–12.
The Fog Index is particularly sensitive to polysyllabic words — words with three or more syllables. "Understand" (3 syllables) counts as complex. "Particularly" (5 syllables) counts as complex. This makes it a good tool for spotting jargon-heavy writing.
| Fog Score | Interpretation | Typical Publications |
|---|---|---|
| 6–8 | Very easy | Children's books, tabloids |
| 8–10 | Easy | Popular novels, TV scripts |
| 10–12 | Acceptable | Most newspapers, blogs |
| 12–14 | Hard | Business reports, textbooks |
| 14–17 | Very hard | Academic papers |
| 17+ | Extremely hard | Scientific journals, legal briefs |
SMOG Index: Built for Healthcare
SMOG (Simple Measure of Gobbledygook) was developed by G. Harry McLaughlin in 1969, initially for TV broadcasting, but it became the preferred readability metric in healthcare and public health because it was specifically validated against comprehension tests on patient education materials.
Polysyllables = words with 3+ syllables. Technically requires ≥30 sentences for accuracy.
SMOG tends to give higher (harder) scores than Flesch-Kincaid for the same text. If your content is for a general health audience, keep SMOG below 8. Patient education materials ideally target SMOG 6–7.
The key difference from Fog: SMOG counts all polysyllabic words without exceptions. Proper nouns, hyphenated words, and numbers all count if they have 3+ syllables.
Coleman-Liau Index: No Syllables Required
The Coleman-Liau formula, developed in 1975, solves a practical problem: counting syllables is difficult to automate accurately. English syllabification rules have hundreds of exceptions. Coleman and Liau found that average word length in characters was a better predictor of reading difficulty than syllable count.
L = average letters per 100 words · S = average sentences per 100 words
Output is a US grade level, similar to Flesch-Kincaid. The main advantage is that character counting is perfectly reliable in any language and any encoding — making it the most consistent formula across different implementations.
Every readability tool implements syllable counting differently. English has no fully reliable algorithmic syllable counter — there are too many exceptions. A tool that counts "opened" as 2 syllables and one that counts it as 3 will produce noticeably different Flesch scores on the same text. Treat scores as ranges, not precise values.
Which Formula Should You Use?
Does Readability Affect SEO?
Google has confirmed that readability is not a direct ranking factor. But the relationship between readability and rankings is more nuanced than a simple yes/no.
What readability does affect:
- Bounce rate — readers abandon difficult content quickly. High bounce rate is a negative engagement signal.
- Dwell time — clear, well-structured content keeps readers on page longer.
- Social sharing — people share content they understand. Complex writing gets shared less.
- Backlinks — other sites link to content that explains things clearly. Impenetrable writing attracts fewer links.
- Return visits — readers come back to sources they find reliable and easy to understand.
All of these are behavioral signals that Google's algorithms measure and factor into ranking decisions. Readable content doesn't rank higher because Google counted your Flesch score — it ranks higher because it earns better engagement metrics.
If you're writing for developers, doctors, or lawyers, a Flesch score of 40 is perfectly appropriate. Chasing a high readability score on technical content produces vague, useless writing. Match your language to your actual audience, not to a generic target number.
How to Actually Improve Readability
These techniques improve every readability score simultaneously because they address the underlying issues all formulas measure:
1. Shorten your sentences
Aim for an average sentence length of 15–20 words. Not every sentence needs to be short — variety is important — but long sentences (30+ words) should be rare. Every conjunction ("and," "but," "however") is a potential sentence break.
2. Replace polysyllabic words where possible
Not every long word can or should be simplified. "Cardiovascular" in a medical article stays. But "utilization" becomes "use." "Subsequently" becomes "then." "Implement" can become "do" or "run" in most contexts. Challenge each long word: is there a shorter one that means exactly the same thing to your audience?
3. Use active voice
Passive voice adds words, hides the subject, and tends to produce longer sentences. "The report was written by the team" → "The team wrote the report." Active voice is cleaner, shorter, and reads faster.
4. Break up dense paragraphs
Readers scan before they read. Large blocks of text signal difficulty. Keep paragraphs to 3–5 sentences for web content. Use white space generously. Each paragraph should cover one idea.
5. Add subheadings
Subheadings don't affect readability scores, but they dramatically affect actual reading comprehension. A reader who can see the structure of your argument before they read it will understand it better than one who encounters it linearly.
6. Cut filler phrases
Phrases like "it is important to note that," "in order to," "due to the fact that," and "at this point in time" add words without adding meaning. Every word that doesn't earn its place makes the reader work harder.
The Limits of Readability Scores
Readability formulas cannot measure:
- Whether your argument is logical or well-structured
- Whether your examples are relevant and helpful
- Whether you've covered the topic completely
- Whether the content matches what the reader is actually trying to learn
- Whether the writing is engaging or interesting
- Whether technical terms are appropriately explained
A document with a Flesch score of 80 can be perfectly clear and completely useless. A document with a Flesch score of 40 can be exactly right for its audience.
Use readability scores to catch obvious problems — brutally long sentences, unnecessary jargon, passive-voice overload — and then use your own judgment, real reader feedback, and engagement data to refine from there.