Short definition
Spearman rank correlation — a statistical measure of how strongly two rankings move together, from −1 (exactly opposite) through 0 (no relationship) to +1 (exactly alike).
Also known as: Spearman's rho, rank correlation, Spearman coefficient.
Two lists, one question: do they move together? That is what Spearman rank correlation answers. It is the tool we use to check whether things like speed scores line up with search positions — and it produces a single number between −1 and +1.
What Spearman rank correlation is
Spearman rank correlation — usually written as ρ (rho) or rs — measures how strongly two rankings go together. You rank your data twice: for example, nineteen website builders by their PageSpeed score and the same nineteen by their position in Google. Then you check, for each agency, how far apart the two ranks sit. If the lists move together, the result approaches +1. If one climbs while the other falls, it approaches −1. If there is no pattern, the value sits around 0.
The difference with a regular (Pearson) correlation is subtle but important: Spearman uses only the order, not the underlying values. The gap between position 1 and 2 in Google can be a world of clicks apart; to the measure it counts as one rank step. That makes it robust against outliers and skewed distributions — exactly what you get with scores and rank positions. The math behind the measure is explained clearly on Wikipedia.
How to read the number
The reputation of the number follows a rough convention. Values around ±0.8 or stronger point to a clear relationship; between ±0.4 and ±0.7 the relationship is moderate; anything closer to zero is weak to absent. But those thresholds are conventions, not laws: with a small sample of around twenty pages, an apparently solid value can still be chance. The sample size always belongs with the outcome, just like the question of what exactly was compared to what.
More important still: a correlation describes co-movement, never a cause. Two rankings can move together because a third factor drives both — a strong brand, for example, produces both good positions and a polished website. Anyone who turns a correlation straight into a causation claim goes further than the number supports.
Why we use it
In our lab and field data research we wanted to answer exactly this kind of question: does a high PageSpeed score come with a higher position for competitive searches? Spearman rank correlation is the right tool because positions are ranks and scores are not neatly distributed.
The outcome in September 2026: +0.32 between lab score and position number, meaning agencies with better scores actually sat slightly lower in Google on average. Field LCP — the speed real visitors experience — scored −0.07, no relationship. Together, those two numbers say something no single PageSpeed report can: speed alone does not sell positions. What does track with position is the authority of the page — and that is a different, bigger story.