Readability Score, Explained: What the Number Actually Means (and How to Raise It)
"Your readability score is 42" is the kind of output that looks scientific and explains nothing. What does 42 mean? Is it good? And more usefully — what do you actually change to raise it? This post takes the Flesch Reading Ease score apart piece by piece, maps the numbers to plain-English difficulty, and shows the two levers that move the score, with real before-and-after sentences.
The formula, written out
The Flesch Reading Ease score is calculated like this:
206.835 − (1.015 × words per sentence) − (84.6 × syllables per word)
Two inputs, two knobs. The average sentence length of your text, and the average syllable count of your words. That is all — the score is not measuring vocabulary sophistication directly, topic difficulty, or logic. It is measuring syntactic and phonetic load: how much the reader has to hold in working memory (long sentences) and how many sounds they have to decode per word (long words). Everything else in this post follows from that.
The scale, mapped to human difficulty
- 90–100 — Very easy. Grade 5. Children's books, emergency signage. Almost too simple for most commercial web content.
- 60–70 — Easy. Grade 7–8. The sweet spot for web content: Wikipedia, successful news sites, and most high-traffic SaaS pages land here. Readable by everyone without sounding childish.
- 40–50 — Plain but demanding. Grade 10–12. Government forms, mainstream magazines. Fine for an audience that chose to be there; a wall for a casual audience.
- 20–30 — Difficult. College level. Technical manuals, academic writing. Legitimate for expert audiences — and a genuine problem when the audience is the general public.
- 0–20 — Very difficult. Academic papers, legal documents. The floor of the scale; legal prose commonly lives here by design.
The target is not "maximize the score" — it is "match the audience." A tax-law explainer for accountants at 45 is perfectly calibrated; a tax-law explainer for freelancers at 45 is a bounce-rate factory. The score is a diagnostic, not a goal.
Lever 1 — sentence length (the big one)
The formula weights sentences at 1.015 per word over par and syllables at 84.6 per extra syllable, which makes long sentences the cheaper fix in practice: splitting one 32-word sentence into two 16-word sentences moves the score more than swapping a dozen two-syllable words. The rule of thumb: keep most sentences under 20 words, and let none drag past 30 unless the rhythm demands it.
Before (38 words, score cost visible): "While the compression process itself does not require any internet connection or server-side processing, it is important to remember that the quality of the output image will depend significantly on the quality settings you choose during the encoding phase." (39 words)
After (two sentences, same content): "Compression happens entirely on your device — no internet connection, no server. The output quality depends on the settings you choose while encoding." (20 words total) Same information. Half the load. The reader keeps moving instead of re-reading.
Lever 2 — syllables per word
Long words are not automatically bad — "algorithm," "compression," "browser" are the vocabulary of the topic, and replacing them with shorter words often makes the sentence less clear. The fix is surgical: swap the words that are doing decorative work, not the ones doing referential work. A quick test: if a synonym is shorter and more common in the reader's daily speech, take it. "Utilize" → "use." "In the event that" → "if." "Commence" → "start." Leave "compression" alone — "squishing" would score better and mean less.
The Readability Checker shows both levers side by side — words per sentence and syllables per word — so you can see which one is dragging a specific draft, rather than guessing.
Why readability matters beyond "user experience"
Three concrete mechanisms, in order of evidence strength:
- Comprehension is a prerequisite for everything else. If the reader does not understand the first paragraph, they do not see the offer, the link, or the value. Every downstream metric — conversion, time on page, sharing — starts from comprehension.
- Reading ease predicts abandonment on the open web. There is no "ask the writer for a simpler version" on the web. The reader's only tool is the back button, and hard text is the most reliable trigger for it.
- Readable content gets the backlinks. People cite, share and reference what they understood. A 60-score explainer gets embedded in other people's posts; a 30-score whitepaper gets downloaded, read once, and never cited.
A practical editing pass (5 minutes per 1,000 words)
- Score the whole draft — get the number and the two metrics.
- Find the 20-word-plus sentences (most editors will flag them; the checker's sentence count tells you how many there are) and split the worst three.
- Read it aloud. If you run out of breath mid-sentence, the reader's working memory has already given up. That breath test is surprisingly well correlated with the sentence-length metric.
- Re-score. You should see a 5–15 point movement from step 2 alone. Iterate once more if you are still far from the audience band.
The reference
The full history and variants (Flesch-Kincaid Grade Level, the original 1948 formula) are documented in the Flesch-Kincaid readability tests reference — worth reading if you want to know why the score tops out at 100, or what the Grade Level variant adds.
FAQ
- Does Google use readability in rankings? Not as a direct input — but the behaviours it predicts (abandonment, dwell time, citation) are exactly the signals that do move. Treat it as an instrument for the things that rank, not a ranking factor itself.
- What score should my site aim for? 60–70 for general audiences. 40–60 for professional audiences who came looking for depth. Below 40 only for expert/technical content where the vocabulary is non-negotiable.
- Is a higher score always better? No. An 85-score explainer about quantum error correction is not "better" — it is mismatched. Match the band to the reader, then optimize within it.
- Can I check a whole site at once? Paste each page into the checker. For a large site, start with the pages that drive conversions or top your traffic — that is where the score has money attached.