The Variation Clustering Score is a legacy composite statistic that summarizes where your generated keyword variations land across the cleaned body content and how spread out those positions are. It is a descriptive distribution measure, not a target — higher or lower is not universally better, and missing positions can make the result non-finite.
| Factor ID | RG-SRP-040 |
|---|---|
| Family | Term Position Distribution |
| Measurement | Composite position/spread score |
| Measured zone | Cleaned, lowercased, trimmed body content |
What it measures
This factor reports a legacy clustering score derived from all generated keyword-variation positions. It rolls the location and dispersion of those variations into a single number, describing both roughly where the variations sit in the body and how tightly or loosely they are grouped.
How RankGear measures it
RankGear combines the offsets of every generated variation, then computes their mean, their population variance, and the UTF-16 length of the cleaned body content. Those inputs feed a single formula:
score = 100 × (mean / body_length) × min(1, population_variance / mean)| Component | What it represents |
|---|---|
| mean | Average of the combined variation offsets |
| body_length | UTF-16 length of the cleaned, lowercased, trimmed body |
| population variance | Spread of the combined variation offsets |
The min(1, …) term caps the variance factor at 1, so wide dispersion contributes up to but not beyond that ceiling. If some variation positions are missing, the calculation can yield a non-finite result.
How to optimize it
Treat this score as an observation, not a target. Organize the page around coherent subtopics and use appropriate variation language naturally within those sections. Optimize for topical clarity, not for a manufactured distribution pattern — engineering the number itself does not improve the underlying content.
Important considerations
- Positions are UTF-16 offsets, not words, sentences, pixels, or HTML byte positions.
- Cleaned body content excludes markup and may differ from the raw source.
- The generated variation set affects this and other variation-based factors.
- Missing variation positions can produce a non-finite score.
- These are descriptive distribution statistics; lower or higher is not universally better, and the score relates to — but does not cause — ranking within a measured result set.
Related factors
Part of the Factors reference · how RankGear measures · glossary.