Title vs meta description overlap checker

This checker normalizes both fields, measures their unique-word overlap, lists the words they share, and identifies when the description is simply the title repeated one or more times. Use the result to spot wasted description space before publishing, review large batches consistently, or explain a rewrite decision with a transparent metric instead of relying only on a quick visual impression.

What the overlap score tells you

The overlap ratio is a Jaccard score calculated from the unique normalized words in the title and description. The checker counts the words that appear in both fields and divides that count by the number of distinct words appearing across either field. A result of zero means the fields share no normalized words, while a result of one means they use exactly the same unique vocabulary. Values between those endpoints are evidence, not a universal pass or fail grade. Some repetition is useful because the title and description normally discuss the same page, product, or question. Excessive repetition can indicate that the description adds little context, but a low score can also reveal that the fields target different topics. Review the shared-word list alongside the ratio to see whether the overlap comes from meaningful subject terms or generic words. The counts include repeated occurrences in each field for context, although repeated occurrences do not increase the unique-word ratio. This keeps a keyword repeated five times from artificially inflating the comparison.

How text is normalized and repetition is detected

Before comparing the fields, the checker applies Unicode compatibility normalization, converts letters to lowercase using a fixed English locale, and extracts sequences of letters and numbers as words. Apostrophes inside words are retained, so a contraction remains one token, while punctuation around a word does not affect the match. The output is deterministic: it does not call a model, fetch a search result, use a clock, or depend on a browser locale. The repeated-title flag uses a stricter rule than the overlap ratio. It becomes true only when the complete normalized description word sequence is exactly the title sequence repeated one or more times. Therefore, a description containing all title words in a different order may receive a ratio of one but will not be marked as a repeated title. Likewise, adding even one genuinely new word makes the repetition flag false. Empty strings, whitespace-only values, and text containing no recognizable word are rejected because a meaningful ratio cannot be computed from either case.

Turning the result into a better description

Start with the repetition flag. If it is true, rewrite the description so it contributes information that the title cannot fit: a benefit, scope, audience, differentiator, constraint, or useful next step. Then inspect the shared words. Retain the specific terms that connect both fields to the page’s real subject, but avoid filling the description with repeated variants solely to change the score. A good description usually expands the promise made by the title in natural language. When the flag is false, treat the ratio as a diagnostic rather than a target. Compare similar pages within the same site, look for unexpected extremes, and read the actual fields before deciding. The checker does not predict rankings, measure click-through rate, judge character length, or prove that a snippet will be displayed by a search engine. It isolates one narrow editorial question so it can be combined with length checks, keyword review, structured data validation, and human proofreading. Automated workflows can call the API for $0.002 per item and route suspicious records to an editor.