Readability: how hard the copy is to read

What this screen is

Readability tells you how hard the writing is in the newsletters you track, and whether it is hard enough to be costing engagement. Each collected newsletter is scored on Flesch reading ease and Flesch-Kincaid grade. Those scores then roll up into portfolio averages, two monthly trend lines, a per-brand table and a row per edition.

Both scores are arithmetic, worked out from the body text rather than judged by AI, so the same email always scores the same. Behind almost every panel sit two fixed target bands, Flesch 60 to 70 and grade 7 to 8. Those are best-practice numbers rather than measurements of your own data, and every tile, trend line, chart and brand tag on the screen judges against them.

Two panels plot readability against an estimated open rate. Newsletrix subscribes to competitor newsletters like any other reader, so it never sees a sender's real opens and that axis is an estimate. There is no date range control either. Every panel covers everything you have collected for the selected brands.

Open it when

• You are rewriting your newsletter and want your market's reading level.
• A competitor outperforms you and you suspect their copy is simply easier to read.
• You want a shortlist of brands writing too densely, above grade 8.5, or suspiciously thin, below grade 6.5.
• You want the exact word, sentence and character counts behind one edition.

Screen anatomy

The screenshot is from an earlier release; the map after it walks the current layout.

1. KPI tile row Average word count, Average reading time, Flesch reading ease, Flesch-Kincaid grade, Median word count. The marked strip judges that score against a fixed band rather than against your own history.
2. Flesch reading ease - monthly trend One point per calendar month, green band drawn between 60 and 70, points inside it drawn green.
3. Flesch-Kincaid grade - monthly trend The same chart for grade, band between 7 and 8. A line above it month after month is the clearest call to simplify.
4. Word count - monthly trend Average body word count per month.
5. Correlation cards Readability against estimated open rate. The badge, an r value plus a strength word, is the message. Both ignore the brand filter.
6. Per-brand readability Status chips, then a sortable row per brand with a tag and a recommendation. The N column says how far to trust it.
7. Detailed metrics table One row per newsletter. Flesch RE is green from 60 up, yellow from 30 to just under 60, indigo below 30.
8. Page header Title, subtitle and a help icon opening a plain-language explainer.
9. Filter card The Brand multi-select, placeholder All brands. Clear all appears once a brand is selected.
10. Understanding readability scores A fixed blue box with the Flesch scale, grade levels and lexical diversity.

Every panel explained

Average word count, Median word count and Average reading time

How long a typical edition is in the current filter.

How it is calculated. Word count is how many words the body text holds, and the average tile is their mean. The median tile takes the upper middle value of the sorted counts, which is not a true median when the number of editions is even. Reading time assumes a flat 200 words per minute, worked out per edition and then averaged.

How to read it. A mean well above the median means a few long editions are dragging it up, so quote the median instead. Anything above five minutes is long-form. Bear in mind that the count includes footers and image alt text.

Flesch reading ease and Flesch-Kincaid grade tiles

The two headline scores, each the mean to one decimal. Under each one sits a benchmark strip reading, for example, You 59 · Best practice 60-70 · below range.

How it is calculated. Flesch reading ease is 206.835 - 1.015*(words/sentences) - 84.6*(syllables/words), clamped to 0-100. Flesch-Kincaid grade is 0.39*(words/sentences) + 11.8*(syllables/words) - 15.59, floored at 0. A sentence ends at a full stop, question mark or exclamation mark, and syllable counts are estimated rather than looked up in a dictionary. The strips compare against fixed bands, 60 to 70 with an ideal of 65, and 7 to 8 with an ideal of 7. Because lower is better for grade, a value under 7 is still scored as good, even though the strip labels it below range.

How to read it. Higher Flesch is easier, so below 50 reads academic and above 80 juvenile. Lower grade is simpler. Both formulas run on the same two levers, long sentences and long words, so treat the pair as a single piece of evidence.

The two monthly trend lines

One average per calendar month across your history, with the target band shaded behind the line.

How it is calculated. Each point is the average for one calendar month of the send date, oldest month on the left. The vertical scale fits whatever is on show and always keeps both band edges in view.

How to read it. Direction matters more than any single point. A falling Flesch line, or a rising grade line, means the market is writing more densely. The pill under the chart judges the latest month only. Averages are unweighted and there is no minimum sample, so a one-newsletter month is drawn exactly like a fifty-newsletter one.

Word count - monthly trend

A bar per month for the average body word count, labelled Newsletter avg. word count. An empty filter reads No data available. Rising bars with a falling Flesch line mean editions are getting longer and harder at once.

The two correlation cards

Each card plots one dot per newsletter, with a fitted trend line, a dashed line at the industry median open rate (retail, the default, is 15.68 percent) and one at the ideal, Flesch 65 or grade 7.

How it is calculated. The chart covers the 500 most recent newsletters carrying both a score and an estimated open rate, and the n beside the title says how many that came to. The line is a least-squares fit, r a Pearson correlation. Below three points, or with no variation on one axis, the card reads Not enough data to plot a correlation. The badge turns r into a word. An r of 0.5 and above is strong, 0.3 moderate, 0.1 weak, and anything under that none.

How to read it. Read the badge, not the dots. A Flesch line rising to the right, or a grade line falling, both say easier copy goes with higher estimated opens.

Per-brand readability

One sortable row per brand, under three chips counting the statuses.

How it is calculated. Each row averages that brand's newsletters, and a brand appears as soon as one of its editions has both scores. Status compares the average grade against the 7 to 8 band with a half-grade tolerance. Above 8.5 is Simplify, below 6.5 is Too easy, and anything in between is On target. The recommendation comes from the same numbers, for example, Writes at grade 9.4. Consider simplifying (target 7-8). Colours switch at the same points, plus Flesch 55 and 80. Default sort is grade, highest first.

How to read it. Sort by Grade for the densest writers, then by Open rate to see whether they are punished for it. One newsletter is enough to place a brand here, so check the N column first.

Detailed metrics table

Every newsletter in the filter, one row each, from Subject through to FK grade.

How it is calculated. One row per newsletter, newest first. Avg. chars per word is characters divided by words, reading time is word count divided by 200, and characters include spaces. A missing Flesch or grade prints N/A, while a missing count prints 0. Nothing is paged or capped, so a large account produces a very long table.

How to read it. Scan the Flesch RE colour break at 60, then read the sentence count and characters per word beside it: long sentences and long words need different fixes.

Understanding readability scores

A fixed box repeating the generic Flesch scale (90 to 100 very easy, 60 to 70 standard, 0 to 30 very difficult), grade levels as US school grades, and lexical diversity as unique words over total words. Those generic bands are not the email target band used everywhere else on the screen.

Metric definitions

Filters and controls

Four things are deliberately absent. There is no date range picker and no search box, no export or download button, and no group or tag selector. Every panel covers your whole history for the selected brands. For a date-scoped view, work from Newsletters instead.

How to use it

1. Read the score tiles through their benchmark strips In range, below range or above range is the verdict against the fixed 60-70 and 7-8 bands.
2. Check direction on the trend lines Is the line drifting toward or away from the band, and is the latest month in or out? Discount thin months.
3. Put length next to difficulty If the word count bars rise while the Flesch line falls, cut length before rewording sentences.
4. Sort the per-brand table by Grade, then check N The top rows are your densest writers; drop single-digit samples.
5. Treat the correlation cards as a sanity check Because the open rate axis is an estimate, a strong result is only a hypothesis to test.
6. Drill into the detailed table Find the worst Flesch scores, note their sentence count and characters per word, then open them in Newsletters.

Limits and caveats

• Syllable counts are estimated rather than looked up in a dictionary, and both formulas inherit that error.
• Only a full stop, question mark or exclamation mark ends a sentence, so links, abbreviations and bullet lists inflate the sentence count and flatter both scores.
• Both formulas are built for English. Non-English newsletters still get a score, and it is not meaningful.
• Flesch is clamped to 0-100 and grade floored at 0; a newsletter scoring exactly 0 is left out of that average but still listed in the detailed table.
• Monthly trends are unweighted averages with no minimum sample size.
• Every open rate here is an AI estimate. The scatters cap at the 500 most recent qualifying newsletters and report a correlation with no significance test.
• A newsletter shows up here once it has readable body text, usually within a few minutes of arriving.

FAQ

Related screens