The Great Website Diet
Enterprise homepages have lost up to 80% of their text over the years, which we call the written thinness risk. Drop any URL below to see how yours stacks up and whether AI can still read you.
25 free checks a day
Methodology and sources
How it works
We fetch an archived copy of the home page from the year you pick, trying several dates across that year and keeping the copy with the most readable text. Copies more than a year away from the chosen year are discarded rather than compared.
We then fetch the live home page today and measure both the same way: readable words, the share of the page code that is text, images and media per paragraph, and headings and lists.
The score (0–100) is based on today's live page only: text density 30%, visual-to-text balance 25%, text architecture (readable words) 25% and content structure 20%. 0–35 is low risk, 36–70 moderate, 71–100 high. The archived copy is shown as a trend for context and never changes the score.
Limits
Only the home page is measured, not the whole site. Pages built in Flash hid their text from the archive, so very thin old copies carry a Flash warning: read that comparison with care. The score is unaffected.
Some sites block automated visits; those show as No data rather than estimated figures. If no archived copy exists for the chosen year, today's score still shows, no comparison is made and the check isn't counted.
The score is an indicator of how much text AI engines have to read, not a measure of how they actually describe a brand.
Data sources
- Historical pages and page counts: Internet Archive Wayback Machine.
- Today's figures: the site's live home page and its public sitemap, fetched at the time of the check.
To cite: Seenford, Thinness Risk check, seenford.com/freetools/thinnessrisk, with the date of your check.