Foundational pillar / check reading-level

Reading ease: our editorial readability guideline

Our editorial guideline for AI-answer extraction, not a Google rule: we compute the Flesch reading ease of the page copy and compare it with our 50 to 70 guideline. No engine publishes a readability score; Bing asks for content that is easy to understand without external context.

By Shimon Carroll, Founder, SEO for AI Agents · Last updated

What this check measures

The Flesch reading ease score of the visible body text, computed from average sentence length and average syllables per word. Higher is easier to read. Alongside the score we report the sentence-length distribution (a typical length, the long end of the range and how many sentences run very long) so you can see whether a few very long sentences drive it. Pages with too little prose for a stable score (forms, listings, dashboards) are reported as not assessable and never scored.

Why it matters

Plain, clear writing helps visitors understand what you offer and decide to act, and it helps answer engines quote you accurately. Bing's Webmaster Guidelines ask for content that is easy to understand without external context; Google asks for content that is helpful and easy to understand. Neither publishes a readability formula or threshold, and Flesch is a rough proxy: expert pages for professional audiences can be clear and still score low. Treat the number as a strong prompt to reread your most important copy, not as a ranking factor.

How we score it

Below 30 (very hard to read): a MEDIUM finding at medium confidence, labelled as our editorial guideline (50 to 70 is the guideline). 30 and above: passes, with the standard Flesch band named. Not measured, and never scored, when the page has too little running prose, when the text does not split into plausible sentences, or when the formula leaves its 0 to 100 scale. The finding is never high or critical, and it never takes the first action-plan slot while a measured defect exists.

Confidence-flag rules

MEDIUM confidence on a measured score. The score is computed only on validated prose blocks (navigation, headings, tables, code and short list items are excluded). A value off the 0 to 100 scale is reported as unmeasured rather than clamped into a band, so a confident label always rests on real running prose.

Common mistakes

  • Rewriting accurate expert copy into vague generalities just to raise the number.
  • Treating a readability score as a ranking factor; Google does not publish one.
  • Ignoring sentences that genuinely run on far too long in the copy that matters most.

How to fix it

Read the main copy aloud. Where a sentence is hard to follow, split it or use a plainer word, and define jargon your customers may not know. Keep precise technical terms your audience expects. Do not change copy that already reads clearly just to move the score.

Primary sources

Changelog

  • · The finding now also reports how long your sentences run (a typical length, the long end of the range and how many run very long), from the same prose blocks. Evidence only: the score, its band and the severity are unchanged.
  • · First published page for this check, framed as our editorial guideline with Google and Bing sources. Clearly hard prose is a medium finding; a score computed off running prose or off the 0 to 100 scale is reported as unmeasured. The finding no longer describes text as "doctoral-level".
← Back to the methodology indexRun an audit