Spearman correlation is a coefficient from -1 to +1 that measures whether a factor’s values and ranking positions move together in rank order — without assuming the relationship is a straight line.
| Term | Spearman Correlation |
|---|---|
| Category | Statistics and Evidence |
| Also known as | Spearman’s Correlation, Spearman rho |
| Where it appears | Results table and workbook Overview |
What it means in RankGear
RankGear reports a Spearman correlation for each factor in the Results table and again in the workbook Overview. To calculate it, RankGear ranks the pages in a SERP twice — once by their value for that factor, once by their actual ranking position — then measures how closely the two orderings agree. Because it works on rank order rather than raw numbers, it detects any consistently increasing or decreasing (monotonic) relationship, not only a straight-line one, and a single extreme page will not distort it the way it can distort a value-based measure.
How to interpret it
Read the sign first, then the magnitude, then the sample it came from. A positive value means higher factor values tend to accompany better positions; a negative value means the two move in opposite directions; a value near zero means there is no consistent monotonic relationship in that set of results. The closer the number sits to +1 or -1, the more consistent the pattern. Always weigh it against the sample size behind it — a coefficient drawn from a handful of URLs can swing widely — and check whether a low reading reflects a genuinely flat relationship or simply a factor that was missing or unmeasured for most pages.
| Coefficient | What it tells you |
|---|---|
| +0.5 to +1.0 | Higher factor values consistently line up with better positions. |
| -0.5 to +0.5 | Weak or no monotonic relationship in this SERP. |
| -1.0 to -0.5 | Higher factor values consistently line up with worse positions. |
Example
Suppose a run analyzes twenty ranking URLs and the word-count factor returns a Spearman correlation of +0.62. That says the longer pages in this SERP tend to hold the stronger positions in rank order — and because the measure uses ranks, one unusually long page near the bottom will not inflate the figure. A second factor might read -0.08, a signal that its values shuffle almost independently of position and offer little to prioritize around.
Important considerations
- A correlation describes association, not cause. A strong Spearman value shows a factor tracks position in this SERP; it does not prove that changing the factor will move a page.
- Small SERP samples are unstable — with few URLs, one page reshuffling the ranks can swing the coefficient sharply, so treat readings from thin samples as tentative.
- Spearman captures monotonic, rank-order agreement, not straight-line fit; compare it with Pearson when you want to know whether a relationship is also linear.
- The coefficient is a comparative indicator drawn from the pages in front of it, not a score Google assigns — read it as evidence about this result set, not as a ranking weight.
Related terms
Part of the RankGear glossary · how RankGear measures · the 870 factors.