Fixing a Low Flesch Reading Ease Score
Flesch Reading Ease turns two counts, words per sentence and syllables per word, into a score where higher means easier. A low score tells you a page asks a lot of its readers, but not where or why.
This guide shows how to work out which of the two inputs is dragging your score down, how to strip out text that distorts the count, which passages to rewrite first for the biggest gain, and how to pick a sensible target for your audience instead of chasing an arbitrary number.
What the score is made of
Rudolf Flesch published the Reading Ease formula in 1948 for printed English. It has only two moving parts, and knowing how much each one weighs is the key to fixing a low score efficiently.
Reading Ease = 206.835 โ (1.015 ร words per sentence) โ (84.6 ร syllables per word)
The first input is average sentence length. Each extra word in the average sentence costs about one point. The second input is average word length measured in syllables. Its multiplier is far larger, so small changes in word choice move the score a lot: raising the average from 1.4 to 1.5 syllables per word costs about 8.5 points, roughly the same as adding eight words to every sentence.
Two worked examples
| Text | Words per sentence | Syllables per word | Reading Ease |
|---|---|---|---|
| A typical corporate services page | 20 | 1.5 | 206.835 โ 20.3 โ 126.9 = about 60 |
| The same page after rewriting | 14 | 1.4 | 206.835 โ 14.2 โ 118.4 = about 74 |
The rewrite gained about 14 points. Roughly six came from shorter sentences and eight from plainer words. That split is common, and it is why cutting sentences alone often disappoints: people chop a paragraph into fragments, see the score rise by three or four points and give up.
The bands Flesch described
| Score | Flesch's description |
|---|---|
| 90 to 100 | Very easy |
| 80 to 89 | Easy |
| 70 to 79 | Fairly easy |
| 60 to 69 | Standard |
| 50 to 59 | Fairly difficult |
| 30 to 49 | Difficult |
| 0 to 29 | Very difficult |
Scores can fall below zero for extremely dense text and rise above 100 for very simple text. Neither is an error; the formula has no hard limits.
Why pages end up with low scores
A low score almost always has one of a handful of causes. Spotting which applies to your page saves a lot of rewriting in the wrong place.
- Latinate business vocabulary. Words such as facilitate, implementation, functionality, organisational and approximately each carry four to six syllables. A page that uses a few dozen of them will score low even with short sentences.
- Nouns made from verbs. "Carry out an evaluation of" instead of "evaluate", "provide assistance with" instead of "help". These add words and syllables at the same time.
- A few monster sentences. One 60-word sentence in a 400-word page raises the average noticeably. Several of them can sink a page that is otherwise clear.
- Stacked qualifications. Legal, financial and medical copy often hedges every statement with clauses, which lengthens sentences and adds long words together.
- Text that is not prose. Product codes, long URLs, units, email addresses and table cells are counted as words and syllables by most tools. They can distort the score in either direction.
- Specialist terms your readers genuinely need. "Photovoltaic", "anaesthetic" or "conveyancing" may be the right word. They lower the score, and that is sometimes acceptable.
The last cause is the one to leave alone. The aim is to remove difficulty that the subject does not require, not to pretend the subject is simpler than it is.
Diagnosing a low score
Run the page through the Readability Checker. It reports Flesch Reading Ease alongside Flesch-Kincaid Grade Level, Gunning Fog and average sentence length, and it flags individual sentences that need attention. Those extra numbers tell you where to look.
Reading the numbers together
| What you see | Likely cause | Where to start |
|---|---|---|
| Low Reading Ease, average sentence length under about 18 words | Long words | Word choice and nominalisations |
| Low Reading Ease, average sentence length well over 20 words | Long sentences | The flagged sentences, longest first |
| Gunning Fog much higher than the Flesch-Kincaid grade | Many words of three or more syllables | Word choice |
| Score changes a lot when you remove a table or list | Non-prose text skewing the count | Score the prose on its own |
| Score is fine but readers still struggle | Structure, order or missing context | Headings, opening paragraph, examples |
Estimating syllables per word yourself
If your tool does not show syllables per word, you can work it out backwards. Rearranging the formula:
syllables per word = (206.835 โ 1.015 ร words per sentence โ Reading Ease) รท 84.6. A page
with a score of 45 and an average of 16 words per sentence comes out at about 2.05 syllables per word.
Everyday English prose usually sits well below that, so this page has a vocabulary problem, not a sentence
problem.
The fix, step by step
Step 1
Score only the body text
Copy the main content of the page into the Readability Checker without the navigation, footer, cookie banner, product codes, prices or specification tables. Keep headings out too, because many tools count a heading with no full stop as part of the next sentence. Note the Reading Ease score, the average sentence length and the other measures in a spreadsheet so you have a clean baseline that reflects only the prose your visitors actually read.
Step 2
Work out which input is the problem
Look at the average sentence length first. If it is above about 20 words, sentence length is costing you points. If it is below about 18 words and the score is still low, the vocabulary is the problem. You can confirm by estimating syllables per word from the formula. Knowing which lever to pull stops you spending an hour chopping sentences when the real drag comes from long, abstract words.
Step 3
Rewrite the costliest passages first
Paste each section or paragraph into the checker separately and rank them by score. Rewrite
the lowest-scoring passages first, because a few dense paragraphs usually account for most of the damage.
Within each one, start with the sentences the checker flags. Split any sentence that carries more than
one idea, lead with the subject and verb, and cut padding such as in order to or at this point in time.
Step 4
Swap long words that have plain equivalents
Search the text for words ending in -tion, -ment, -ality
and -isation, and for stock business words such as utilise, facilitate
and commence. Replace each with a shorter word only where the meaning survives: use,
help, start. Turn nouns back into verbs, so "make a decision" becomes "decide". Keep
the specialist terms your readers use and search for themselves, and define any that newcomers might
not know.
Step 5
Set a target for your audience
Decide what score is realistic for the people reading this page rather than copying a figure from a plugin. A page for the general public can usually reach the 60s or higher. A page for accountants, engineers or clinicians may sit lower and still be clear to them. Compare against your own best-performing pages on similar topics, and write the target into your style guide so every author aims for the same range.
Step 6
Re-score, read aloud and record the result
Run the rewritten body text through the checker again and record the new score beside the baseline. Then read the page aloud from start to finish. Stop wherever you stumble, lose the thread or sound robotic, because a higher score achieved with choppy fragments or vague words is not an improvement. Publish, then check the page again after later edits so the score does not drift back down.
Before and after: where the points come from
The two rewrites below target different inputs. Seeing them side by side shows why you should diagnose before you edit.
A vocabulary problem
Before: "Our organisation facilitates the implementation of comprehensive maintenance solutions. Utilisation of our services ensures operational continuity."
Only two sentences, of nine and seven words, yet the score is extremely low, probably below zero, because nearly every word has four or more syllables.
After: "We service and repair your equipment so it keeps running. Book us once and we handle the rest."
The sentences got slightly longer, and the score rose sharply, because the syllable count per word fell by about half.
A sentence problem
Before: "If you would like to change your delivery date you can do this through your account up until 48 hours before the scheduled delivery, after which point you will need to contact our customer service team by phone, as changes cannot be made online once the order has been passed to the courier."
Fifty-three words in one sentence. The words themselves are mostly short.
After: "You can change your delivery date in your account up to 48 hours before delivery. After that, please phone us. We cannot change it online once the courier has your order."
Three sentences, 31 words. The meaning is intact and the reader gets the deadline first.
Where the score comes from on common platforms
Different tools apply the same formula but count differently, so expect small disagreements.
- WordPress. Yoast SEO and some other SEO plugins show a Flesch Reading Ease result in the editor for English and several other languages. They score the post content, which may include text from blocks you would not consider prose, such as tables or buttons. Treat the plugin's traffic light as a prompt, and use a checker on the body text for a figure you can compare over time.
- Shopify and other hosted shops. There is no built-in readability score. Copy the product or page description into a checker. Remove size tables and specification lists first, as they are rarely full sentences.
- Microsoft Word. Word can show Flesch Reading Ease and Flesch-Kincaid Grade Level after a spelling and grammar check, once readability statistics are switched on in the proofing options. It is useful for drafts written outside the CMS.
- Static sites and headless CMSs. Score the source Markdown or rich text rather than the rendered page, so code samples, navigation and repeated components do not enter the count.
Edge cases and when to leave a low score alone
Non-English pages
The 1948 formula and its constants were built for English. Scoring a French, German or Welsh page with it gives numbers that do not mean what the bands say. Adapted versions exist for some languages, and some tools switch formula automatically. If yours does not, judge non-English pages by sentence length and a native reader's opinion rather than by the English score.
Very short pages
On a page with 80 words, one long product name or one unusual sentence can move the score by ten points. Scores on short text are unstable. Look at the individual sentences instead.
Regulated wording
Financial promotions, medical information and terms and conditions sometimes contain wording you cannot change. Score the parts you control, and consider adding a plain summary above fixed wording rather than rewriting it. This is not legal advice; check any change to regulated text with whoever signs it off.
Expert audiences
A page written for specialists can score in the 30s or 40s and still be the right page for its readers. Check that the difficulty comes from necessary terms and not from padding. If an expert would also find it slow, rewrite it.
The score and search
Google has not said it uses Flesch Reading Ease or any readability formula in ranking. Raising the score is worth doing because readers find the page easier, not because a number is being measured by a search engine.
Keeping scores from sliding back
- Write a short word list. Pair the long words your team keeps using with the plain words you prefer. Put it where writers will see it, such as the top of the content brief template.
- Check drafts, not only published pages. Scoring a draft before it reaches the CMS is quicker than rewriting live pages later.
- Watch reused boilerplate. Company descriptions, guarantees and service summaries are copied onto many pages. Fixing the source text once raises the score everywhere it appears.
- Review supplier and agency copy. Text written by third parties often arrives in a denser house style. Score it before it goes live.
- Re-check after big edits. Additions tend to be written quickly and in a heavier style than the original. A quick re-score catches that.
For related fixes, see fixing sentences over 20 words and fixing a high reading grade level, which works on the same inputs from the grade-level side.
1. What is a good Flesch Reading Ease score?
There is no single correct figure. Flesch described 60 to 69 as standard and 70 to 79 as fairly easy, and many writers aim for the 60s or above for pages meant for the general public. Pages for professionals can sit lower and still work well for their readers. The better question is whether the people you write for can read the page quickly. Use your own best pages on similar topics as the benchmark, and set a range rather than one exact number.
2. Why is my Flesch score low when my sentences are short?
Because the formula weighs word length heavily. Each 0.1 increase in average syllables per word costs about eight and a half points, which is roughly the effect of adding eight words to every sentence. A page full of words such as implementation, functionality or organisational can therefore score low even with short sentences. Look for long abstract words and nouns made from verbs, and replace them with shorter everyday words where the meaning stays the same.
3. Why do different tools give me different Flesch scores?
All of them use the same formula, but they count the inputs differently. Syllables are counted by rules rather than a dictionary, so tools disagree on words such as area or business. They also disagree on where a sentence ends when they meet headings, bullet points, abbreviations and decimal points, and some include navigation and footer text. A gap of a few points between tools is normal. Track progress in one tool rather than comparing numbers across several.
4. Will improving my Flesch score help my Google rankings?
Not directly as far as anyone can show. Google has not said it uses a readability formula, and there is no public evidence that Flesch Reading Ease is part of ranking. What Google does say is that it aims to reward helpful content written for people. A page that is easier to read is more likely to satisfy visitors and to be quoted accurately by search features and AI assistants. Improve the score for readers, and treat any search benefit as a side effect.
5. Can I get a high score by just using very short sentences?
You can raise the number that way, but you will often make the page worse. Chopping every sentence into fragments of five or six words makes text feel jerky and can break the links between ideas. It also only works on one of the two inputs, so a page full of long words will still score poorly. Aim for a mix of short and medium sentences, one idea per sentence, and plainer vocabulary where it fits.
6. Should technical or legal pages have a high Flesch score?
They should be as easy to read as the subject allows, which may still be a lower score than a consumer page. Keep the terms your readers need and search for, and define them on first use if some readers may not know them. Remove the difficulty that the subject does not require, such as padding, stacked clauses and nouns made from verbs. For regulated wording you cannot change, a plain summary above it helps readers without altering the approved text.
7. Does the Flesch formula work for languages other than English?
The original formula was built for English, and its constants reflect English word and sentence patterns. Applying it unchanged to other languages gives scores that do not match the descriptive bands. Some languages have adapted versions, and some tools switch formula automatically depending on the page language. If your tool does not, use sentence length and feedback from a native speaker who matches your audience as your guide for non-English pages.
The Readability Checker calculates Flesch Reading Ease, Flesch-Kincaid Grade Level and Gunning Fog, average sentence length, passive voice share, paragraph length and jargon density, and flags the sentences that need attention.
Open Readability Fixes โ