What readability scores measure
Readability formulas estimate how difficult a text is to read using two structural signals: how long the sentences are and how long the words are. Longer sentences demand more working memory, and longer words tend to be less common and less familiar.
The important limitation is that these are the only things they measure. A formula cannot assess whether your argument is clear, whether your paragraphs are organised sensibly, or whether the vocabulary suits your audience. A text of short sentences and short words can still be incomprehensible.
Flesch Reading Ease
This is the most widely used score, running from 0 to 100 where higher is easier. Scores of 60 to 70 correspond roughly to eighth or ninth grade reading level and are the usual target for general web writing. Below 30 the text is genuinely difficult, appropriate for academic or technical writing but not for a broad audience.
Rudolf Flesch developed the formula in 1948, and it remains embedded in word processors and in some regulations — several US states require insurance policies to meet a minimum Flesch score.
The grade-level scores
Flesch-Kincaid, Gunning Fog, SMOG and ARI all express difficulty as a school grade level rather than a scale. They use slightly different inputs, which is why they disagree by a grade or two on the same text. That disagreement is normal and not a sign of error.
Gunning Fog and SMOG both weight words of three or more syllables heavily, so they penalise technical vocabulary more sharply. ARI uses character count instead of syllables, which makes it more reliable on text containing numbers and abbreviations where syllable estimation breaks down.
How to actually improve a score
Sentence length is the fastest lever. Any sentence over about 25 words is a candidate for splitting, and long sentences usually contain two ideas that would be clearer separately. Aiming for an average of 15 to 20 words with deliberate variation reads better than uniformity.
Word choice is the second. Replacing "utilise" with "use", "commence" with "start" and "in order to" with "to" costs nothing in meaning and improves both the score and the reading experience.
What does not help is gaming the formula. Chopping sentences arbitrarily produces choppy, disjointed prose that scores well and reads badly. The score is a diagnostic, not a target.
Readability and search
Readability is not a direct ranking factor. It matters indirectly: text people can read easily keeps them on the page, and engagement signals do affect performance. Writing for your actual audience beats writing for a number.
A caveat on syllable counting
Every readability tool estimates syllables algorithmically, since counting them properly requires a pronunciation dictionary. Estimates are close but not perfect, particularly with names, acronyms and unusual words, so treat small differences between tools as noise.
Analysed in your browser
Your text is never transmitted, so unpublished drafts stay private.
Frequently Asked Questions
What Flesch score should I aim for?
Around 60 to 70 for general web writing, which corresponds to eighth or ninth grade reading level. Technical audiences tolerate lower scores.
Why do the grade-level scores disagree?
They use slightly different inputs. Gunning Fog and SMOG weight long words more heavily; ARI uses character count rather than syllables. A gap of a grade or two is normal.
Does readability affect SEO?
Not directly. It matters indirectly, because text people can read easily keeps them on the page, and engagement affects performance.
How do I improve my score quickly?
Split sentences over 25 words and replace long words with shorter equivalents. Avoid chopping sentences arbitrarily, which scores well but reads badly.
How accurate is the syllable count?
It is estimated algorithmically, as in every readability tool. Names, acronyms and unusual words are the main source of small inaccuracies.