Term Frequency compresses the total number of keyword-variation matches in a page’s raw body HTML into a single logarithmic score, returned as 1 + log10(count). It counts raw markup and attributes too, so read it alongside clean-text metrics rather than treating it as a target to inflate.
| Factor ID | RG-KWD-024 |
|---|---|
| Family | Keyword Usage & Density |
| Measurement | Log-transformed variation count |
| Measured zone | Raw body HTML without scripts |
What it measures
This factor measures how many keyword-variation matches appear in the raw body of the page, then compresses that total into a logarithmic score. Rather than reporting a plain count, it expresses term usage on a scale where each additional order of magnitude adds roughly one point — so the gap between a handful of mentions and a great many is deliberately flattened.
How RankGear measures it
RankGear counts case-insensitive whole-word variation matches in the script-free body HTML and returns 1 + log10(count). Matching is case-insensitive and limited to whole words, and <script> contents are stripped before counting — but the remaining raw HTML, including tags and attribute values, is searched as-is.
score = 1 + log10(match_count)| Match count | Score (1 + log10) |
|---|---|
| 0 | −∞ (edge case) |
| 1 | 1.0 |
| 10 | 2.0 |
| 100 | 3.0 |
| 1,000 | 4.0 |
How to optimize it
Use relevant terms naturally across useful content and meaningful markup, emphasizing breadth and clarity rather than repeated exact wording. Because the score is logarithmic, adding more copies of the same phrase yields sharply diminishing returns — a page already in a healthy range gains almost nothing from extra repetition. Treat this number as an observation about how heavily a term appears, not a quota to hit.
Important considerations
- Markup and attributes can contribute to the raw-body count, so the figure is not a clean-text word count.
- Logarithmic scaling reduces the effect of large count differences — a page with 100 matches scores only one point above a page with 10.
- A zero match count produces negative infinity mathematically, so a term that never appears is a distinct edge case rather than simply a low score.
- Compare this against clean-text metrics before acting, since raw-HTML counting can overstate on-page usage.
- This is a correlation-based prioritization signal that tracks term usage within a measured result set; on its own it does not make a page rank or guarantee a position.
Related factors
Part of the Factors reference · how RankGear measures · glossary.