Variations in Style Tags counts how many times a keyword and its variations appear inside a page’s embedded CSS <style> blocks, returning a whole-word occurrence count. It reads source markup, not anything a visitor sees — treat it as a diagnostic, never an optimization target.
| Factor ID | RG-KWD-084 |
|---|---|
| Family | Keyword Usage & Density |
| Measurement | Variation occurrence count |
| Measured zone | Raw style-element contents |
What it measures
This factor counts keyword variations inside a page’s embedded CSS style elements. It looks only at the text held within <style> blocks — the styling instructions in the page source — and tallies how often the tracked keyword and its variations turn up there.
How RankGear measures it
RankGear extracts the style contents from the HTML with scripts removed, leaves the CSS text uncleaned, and totals case-insensitive whole-word variation matches. Because the CSS is not sanitized before counting, a match can land in a selector, a custom property, or a comment as readily as anywhere else in the block.
What RankGear looks for
<style>
/* class names, comments, and custom properties are all scanned */
.running-shoes-hero { color: #5457D6; }
:root { --running-shoes-accent: #FF6F61; }
</style>Scripts are stripped first; the remaining style text is left uncleaned, then each target term and its variations are counted as case-insensitive whole-word matches wherever they appear in that CSS — selectors, comments, or property names included.
How to optimize it
Do not optimize CSS for keywords. This is an observation, not a target — raising it by seeding keywords into class names or comments does nothing for readers and nothing you should chase. Use maintainable selectors, and place the topical information that actually matters in visible, semantic content where both people and crawlers read it.
Important considerations
- Class names, custom properties, and comments can all contribute to the count.
- Minifying or bundling your CSS changes the result, since it rewrites and strips the very text being scanned.
- The value does not represent visible content — nothing counted here is what a visitor reads on the page.
- Treat it as a source diagnostic only.
- This is a correlation-based prioritization signal within a measured result set, not a cause of ranking; a higher or lower count does not make a page rank or guarantee any position.
Related factors
Part of the Factors reference · how RankGear measures · glossary.