Variation Count in Top 30KB tallies how many times your keyword’s variations appear inside RankGear’s compatibility view of the first 30,000 UTF-16 units of the page, plus the alt text pulled from that same early range. It returns a raw occurrence count — a topicality observation from the highest-weight part of the source, not a target to inflate.
| Factor ID | RG-KWD-029 |
|---|---|
| Family | Keyword Usage & Density |
| Measurement | Variation occurrence count |
| Measured zone | Truncated Google-density corpus plus image alt text |
What it measures
This factor counts keyword variations within RankGear’s compatibility representation of the first 30,000 UTF-16 units of the page. Variations are related forms of your target keyword, and the count reflects how clearly the earliest, highest-weight portion of the source establishes the page’s topic. Because it reads a truncated corpus, it captures what a density-style reader would see near the top of the document rather than across the whole page.
How RankGear measures it
RankGear truncates the raw HTML to the first 30,000 UTF-16 units, extracts alt text from that truncated portion, strips selected noncontent elements, appends the extracted alt values to the text, then counts case-insensitive whole-word variation matches across the resulting corpus. The order matters: truncation happens before stripping, and only alt text found inside the truncated source is added.
| Stage | What happens |
|---|---|
| Truncate | Raw HTML is cut to the first 30,000 UTF-16 units |
| Extract alt | Alt text is read from that truncated portion |
| Strip | Selected noncontent elements are removed |
| Append | The extracted alt values are appended to the text |
| Count | Case-insensitive whole-word variation matches are tallied |
How to optimize it
Make your early content clearly topical, use related terms naturally, and provide accurate alt text for relevant images that sit within the measured source range. Treat the number as an observation, not a quota: because truncation happens before element stripping, front-loading genuine, on-topic content and keeping template bloat off the top of the source matters more than repeating variations. A higher count relates to how position within a measured result set tends to look; it does not by itself cause a page to rank.
Important considerations
- Truncation occurs before element stripping, so heavy template markup at the top of the source reduces how much body content fits inside the 30KB budget.
- Some remaining markup may still contribute matches even after noncontent elements are stripped.
- Only alt text found inside the truncated source is appended — images below the cut do not contribute.
- Source order and template size affect what enters the corpus, so two pages with identical visible copy can produce different counts.
- This is a correlation-based prioritization signal, not a cause of ranking; a higher count tracks with position inside a measured result set and does not guarantee or make a page rank.
Related factors
Part of the Factors reference · how RankGear measures · glossary.