Skip to content
×

How do readability scores differ across topics?

Readability scores can differ across topics, but comparisons must use the same language, scoring tool, and sampling conditions. In a small study published by Ahrefs in 2021, the average Flesch Reading Ease (FRE) scores for Google’s top ten results were 69.7 for food, 60.2 for marketing, and 49.6 for engineering. These findings come from a specific sample of US English-language searches. They are not standards that every article should meet, nor can they be applied directly to Traditional Chinese content.

Higher scores for food do not mean engineering articles are less well written

FRE evaluates text using average sentence length and word length. Longer sentences and words, on average, produce lower scores; higher scores indicate text rated as easier to read. The study reported the following topic averages:

  • Food: 69.7, labeled fairly easy to read in the study.
  • Marketing: 60.2, labeled a standard reading level in the study.
  • Engineering: 49.6, labeled difficult to read in the study.

The author interpreted this ranking as scores falling as topics become more complex. However, the comparison itself cannot prove that topic complexity caused the score differences, nor does it establish that engineering writers pay less attention to readability. Different topics and audiences may naturally require different vocabulary.

Before comparing scores, check how the sample was selected

The study first selected several large websites in engineering, marketing, and food, then sampled keywords for which those sites ranked in the top ten in US English-language searches. Each category contained 5,000 unique keywords. In total, the study analyzed 15,000 keywords and their search results pages. The sample covered different search-volume ranges, and obvious branded keywords were removed.

Some pages were excluded because their FRE scores could not be calculated reliably: 19% in engineering, 9% in marketing, and 10% in food. The average FRE scores reported above therefore come from a filtered sample, not a census of all content in the three fields. The study data also provide no comparative results for Traditional Chinese or other scoring tools.

General and specialist audiences need different thresholds

The Ahrefs article suggests that aiming for a higher FRE score may help when writing for a general audience with varying levels of reading comprehension. A pancake recipe, for example, is less likely to need lengthy, complex wording and may naturally score higher. That does not mean a recipe must achieve a score of 69.7.

Technical content, by contrast, may need to retain complex vocabulary. The author uses an article explaining FRE and SEO as an example: replacing necessary terminology solely to raise the score could make the explanation harder to understand. Whether wording is appropriate should depend first on the audience and the topic, not just on the score.

When editing, treat the score as a clue rather than a target

The author recommends not chasing a tool’s “perfect score.” Instead, keep the text flowing smoothly, avoid unnecessarily complex wording, and identify problems by reading aloud and asking others to review the text. If an article’s score is well outside the usual range, it may be worth reviewing. But the decision to revise should depend on whether long sentences and technical vocabulary hinder understanding, not simply on a desire to match another topic’s average.

Readability scores across topics: web perspectives and Q&A

Does the higher proportion of excluded engineering pages raise or lower the average score?

The available data cannot establish either the direction or the size of the effect. In engineering, 19% of pages were excluded because their FRE scores could not be calculated reliably, compared with 9% in marketing and 10% in food. However, the material provides neither readability scores nor their distribution for the excluded pages. It is therefore impossible to say how including them would change the averages, or whether the gaps among the three categories would widen or narrow.

If two topics have different FRE scores, can we tell whether sentences or words account for the difference?

Not from the average scores alone. FRE uses both average sentence length and word length, and both affect the result. To understand which factors are associated with the gap, we would need to examine sentence-length and word-length data for each topic. The available study data are not sufficient to separate their effects.

Does sampling 5,000 keywords per category mean that 5,000 distinct articles were compared in each category?

That cannot be assumed. The study states that each category contained 5,000 unique keywords and that their search results pages were analyzed. However, it does not specify how many unique content pages were included in each category. So 5,000 keywords should not be taken to mean 5,000 distinct articles.

References

  1. Flesch Reading Ease: Does It Matter for SEO? (Data Study)ahrefs.com · 2021-07-06T16:00:20+00:00 — Discusses: Does the higher proportion of excluded engineering pages raise or lower the average score?; If two topics have different FRE scores, can we tell whether sentences or words account for the difference?
Keep the conversation going

Join the discussion 0 comments

What are you working through? Share a question or experience and let’s find the next step.

No comments yet. Start the conversation.

Share your thoughts

← Keep exploring
LET’S TALK

Start with your brand.

Tell me about your website and what you want to achieve.

WhatsApp tony@procrustes.info
Scan to chat on WhatsApp