Flesch-Kincaid & SMOG Readability Scorer

What readability scores measure

Readability formulas are a family of statistical tools that estimate how difficult a piece of text is to read by counting features like sentence length and word complexity. They do not measure the quality of your argument or the accuracy of your claims, but they do give you a reliable proxy for the reading effort your audience is likely to need, which is why they are widely used in education, government, and content marketing.

What the formulas share

Readability formulas estimate how difficult a piece of text is to read by analysing sentence length and word complexity.¹ The Flesch Reading Ease formula came first, and the Flesch–Kincaid Grade Level formula followed later for U.S. Navy training materials.² Both tests reduce a passage to three countable features (words, sentences, and syllables) and then combine those counts with different weighting factors. Because they count the same core features, the two Flesch-based scores correlate closely even though their scales move in opposite directions.

What the formulas cannot capture

None of the formulas measure whether an argument is clear, well-structured, or accurate. They only count syllables, words, and sentences, and they do not account for every difference between readers or the effect of content, layout, and retrieval aids.³ A paragraph packed with short, familiar words scores well even when its logic is impenetrable, while a tightly reasoned argument full of technical terms can score poorly. Think of the scores as a proxy for reading effort, not a measure of quality.

Flesch-Kincaid Grade Level and Reading Ease

Both Flesch-Kincaid formulas use the same three inputs: word count, sentence count, and syllable count. Grade Level maps the result to a US school year, so a score of 8.0 suggests a text is readable by a typical eighth-grader. Reading Ease uses a 100-point scale where higher numbers mean easier reading. Microsoft recommends 60 to 70 for most standard files, while Grade Level is recommended around 7.0 to 8.0 for most documents.² The two scales are not directly convertible because they weight sentence length and syllable count differently.

For most general-audience web content, including blog posts, product pages, and documentation introductions, a Grade Level around 7 to 8 and a Reading Ease of 60 to 70 is a practical target. Specialist writing can sit higher when the audience expects domain vocabulary, but the score is a guide rather than a rule. Aiming for plain language without dumbing your content down is the balance most editors settle on.

A short example shows the arithmetic in action. The pangram The quick brown fox jumps over the lazy dog. has 9 words, 1 sentence, and 11 syllables, which works out to a Flesch Reading Ease of 94.3 and a Flesch-Kincaid Grade Level of 2.3. Both scores land far above the 60 to 70 target range for typical web content, which makes sense: short, familiar one-syllable words are exactly what the formulas reward.

The two scales also have two different jobs. Reading Ease was designed as a continuous 0-to-100 gauge of general adult reading effort, where a higher number always means easier reading, while Grade Level was built to map training material onto a US school year, which is why its output reads as a grade placement rather than a quality score. That origin decides which number to quote: ease suits benchmarking against broad consumer content, because the scale is continuous and comparable across any two texts, while grade suits audience placement, because it answers the direct question of who the text fits.

The two numbers rarely fight. Because both scales reduce the same passage to the same three counts, they move together for nearly every edit, and a genuine split usually means one input is dominating: either long sentences are dragging the grade while the vocabulary stays simple, or dense polysyllables are sinking the ease while the sentence length holds steady. When you report a score to someone else, quote the band rather than the raw decimal, because the tool labels the Reading Ease result with its band, and the band communicates faster than a number like 63.4 ever will.

A score from this tool can legitimately differ from the same text scored in Word or another checker. English does not mark syllables in writing, so every scorer estimates them with heuristics, and sentence boundaries are also an estimate wherever abbreviations and decimals appear. Two careful tools can therefore land a fraction of a grade apart on the same passage, and a drift of a few tenths is measurement noise rather than a contradiction. Compare your scores within one tool over time, and treat cross-tool differences as rounding disagreement rather than a verdict on either implementation.

SMOG Grade explained

SMOG (Simple Measure of Gobbledygook) was developed by G. Harry McLaughlin and published in 1969 as an alternative to formulas such as Gunning Fog.⁴ It focuses on polysyllabic words, those with three or more syllables, and it is widely used for checking health messages.⁵ The full formula counts polysyllabic words across three 10-sentence samples, takes the square root of the nearest perfect square, and adds 3 to produce a grade estimate.

SMOG is designed around three 10-sentence samples, so very short inputs are less stable than longer samples.⁵ That is why the tool shows "Need more text" for very short inputs. Because SMOG weights polysyllabic words more heavily than Flesch-Kincaid does, it can produce different grade estimates from the same passage. You should treat early scores on brief drafts as provisional rather than final.

How sentence highlighting works

The sentence-highlighting view annotates your text one sentence at a time so you can see exactly which passages are pushing your readability scores upward. Long sentences raise the average words-per-sentence input that every major formula relies on, so flagging them visually turns an abstract number into a concrete list of revision targets you can work through in order.

What gets flagged and why

The "View Highlights" view annotates your text sentence by sentence. Any sentence with 25 or more words is highlighted in amber. Long sentences are a leading cause of high readability scores because they raise the average words-per-sentence input used by the formulas.⁶ By surfacing each long sentence visually, the view lets you find the passages that drag your average down without re-reading the whole draft.

Turning highlights into revisions

To revise, click "Edit Text" to return to the textarea, shorten the flagged sentences, and watch the scores update automatically. A single long sentence in an otherwise short-sentence document can shift the Grade Level noticeably, so the threshold is intentionally conservative. Working from the highlighted list keeps your revisions focused and measurable.

After splitting or trimming each flagged sentence, re-check the scores to confirm the change moved the number in the right direction. Highlighting turns the abstract formula into a concrete set of edit decisions, so you can trust that each rewrite is working without reading the raw numbers every time. Iteration is the whole point of the feature.

The moves that move the number

Only three levers actually move the number. Splitting or trimming long sentences lowers the average words per sentence, the input the formulas weight most heavily. Swapping a polysyllabic word for a shorter equivalent, wherever the meaning survives the swap, lowers the syllable count that feeds every formula. Cutting filler words shortens sentences without sacrificing content, so it moves both counts at once. Every editing lever that exists maps to one of those three inputs, which is why the same small set of moves improves all three scores together and why one disciplined editing pass is usually enough.

The boundary is just as important. Because the formulas cannot see structure, argument order, or voice, some edits improve real clarity while leaving the score untouched: reorganizing a paragraph so the key point lands first, cutting a tangent, or fixing a misleading analogy all read as changes to the humans you write for and as nothing at all to the arithmetic. A reorganized draft can therefore read dramatically better at an identical score. Treat the number as a floor to clear rather than a target to max out, and let judgment own everything above the floor.

Matching scores to your audience and content type

The right readability target depends on who will read your content and what they need from it. Marketing copy for a general consumer audience often benefits from a Reading Ease above 65 and a Grade Level below 9 because readers scan rather than study, and complex sentences make re-reading more likely. Specialist material can sit higher when precise terminology is part of the job.

SEO content, product documentation, and instructional guides each have different effective ranges. Product documentation for software developers often needs a higher grade level to describe APIs and configuration options accurately. Instructional guides for novices benefit from shorter sentences and familiar vocabulary because clarity matters more than precision. CapyToolkit calculates your score instantly as you type, so you can revise sentences, check the result, and converge on your target without switching between tools or submitting text to a remote API.

Readability is not a ranking factor. No search engine publishes a readability formula as part of its ranking signals, so a score never ranks your content directly, and chasing a number for the crawler's sake is effort spent on the wrong target. What the score does affect is what people do after they arrive: text that reads easily gets finished, and finished readers stick around long enough for anything else to happen. Plain-language writing therefore serves readers first, and any downstream engagement effect arrives second, not the other way around.

Some fields practice readability as policy rather than advice. With SMOG as the customary yardstick, health materials are the classic case, and consent forms follow the same pattern: plain-language checklists tell writers to run readability tests on the draft and revise the hard sentences before the document goes out, because the reader cannot opt out of understanding what they are signing. That is the context where a score stops being advisory and becomes a gate the draft has to clear, and it is why some institutions publish a fixed target band their forms must meet.