Sample size is the number of usable observations behind a calculation — in RankGear, the count of ranking pages that were successfully measured and included in a statistic.
| Term | Sample Size |
|---|---|
| Category | Statistics and Evidence |
| Also known as | n |
| Where it appears | Statistics, legends, and report metadata |
What it means in RankGear
Sample size sits next to the averages, medians, and spreads shown in RankGear’s Statistics view, in chart legends, and in the metadata attached to saved reports. It records how many pages actually contributed to a given number — not how many URLs were requested, but how many returned usable data for that measurement. Because a competitive analysis weighs your page against the pages currently ranking for a keyword, the pool is small to begin with, and any page that fails to load, blocks the fetch, or lacks the element being counted simply drops out of the total.
How to interpret it
Read every statistic together with its sample size. A mean built on three pages carries far less weight than the same mean built on nine, so let n temper how firmly you act on a figure. It also pays to know what a low or zero value means: a measurement can be present and counted, missing because the page could not be measured, or genuinely zero because the element was absent on pages that were measured. A zero that means “absent” is real data; a zero that means “not measured” is not, and the two should never be averaged together.
| Sample size | What it tells you |
|---|---|
| Larger n | More stable statistics; comparisons and targets are more trustworthy. |
| Small n (a handful of pages) | Treat averages and spreads as directional, not precise. |
| Excluded pages | Pages that could not be measured are dropped from n rather than counted as zero. |
Example
Say you analyze a keyword whose top ten results include two pages that block RankGear’s fetch and one that returns an error. The word-count statistic then reports a sample size of seven, not ten. Rerun the analysis a week later and nine pages measure cleanly — the average word count can shift noticeably even though the underlying pages barely changed. Most of that movement is the larger, steadier sample, which is exactly why the second run’s target deserves more confidence than the first.
Important considerations
- Statistical association is not causation: a sample size tells you how much evidence stands behind a number, not that matching that number will move rankings.
- SERP samples are small by nature, so a single page can swing an average — weight your conclusions by n.
- Provider metrics are comparative indicators drawn from the pages in the sample, not Google’s own scores.
- A larger sample narrows uncertainty but never removes it; distribution and outliers still matter.
Related terms
Part of the RankGear glossary · how RankGear measures · the 870 factors.