Variations in HTML Tags counts how many times your keyword’s variations appear anywhere in a page’s rendered HTML source — visible text, attributes, URLs, and markup alike — and returns that total. Despite the name, it is not limited to tag names, and it is not a clean-content measure: source-only strings count too.
| Factor ID | RG-KWD-061 |
|---|---|
| Family | Keyword Usage & Density |
| Measurement | Variation occurrence count |
| Measured zone | Entire rendered HTML source |
What it measures
Despite its name, this factor does not look only inside tag names. It counts keyword variations throughout the complete rendered HTML source of the page. Any place a variation string appears in that source — visible copy, attribute values, link targets, or the surrounding markup — can add to the total.
How RankGear measures it
RankGear applies case-insensitive, whole-word variation matching directly to the raw rendered HTML and returns the total. Overlapping matches are resolved with longest-first consumption: when several variations could match the same span of text, the longest variation is counted first and the characters it covers are consumed, so a shorter variation nested inside it is not double-counted.
| Where a variation can appear | Counts toward the total? |
|---|---|
| Visible page text | Yes |
| Attribute values (alt, title, class, data-*) | Yes |
| URLs and link targets | Yes |
| Surrounding markup and template code | Yes |
How to optimize it
Raise the value the honest way: improve the visible topical content and the meaningful metadata a reader or crawler actually uses. Do not manipulate source-only text or tag names to inflate the count — that adds strings no reader sees and does not improve the page. Read this factor as an observation about how saturated the whole source is with your topic’s vocabulary, not as a target to maximize.
Important considerations
- Text, attributes, URLs, and markup can all contribute to the count, so the number reflects the entire source, not just the copy a visitor reads.
- Template code can influence the value — boilerplate, framework output, and repeated components may add variations that have nothing to do with the article body.
- This is not a clean-content metric. A high number does not mean the visible text is rich; it may reflect source noise.
- Read it alongside visible-text factors so you can tell genuine topical depth from source-level repetition.
- Correlation is not causation. This count relates to where pages sit within a measured result set; it is a prioritization signal, not a lever that makes a page rank.
Related factors
Part of the Factors reference · how RankGear measures · glossary.