Percentile
IntermediateA percentile is a value telling you what share of the data falls below a given point; the 75th percentile, for instance, is the value below which 75% of the data lies.
A percentile generalizes the median (the median is just the 50th percentile) and is the main tool for reading skewed distributions. Its power is that, instead of one central number, it lets you go to the edges: the 95th percentile of load time is 'the worst experience the slowest 5% of users have', and it is that edge which drives users away, not the flattering average. In measuring web speed, the 75th percentile has become the standard because it takes the slow tail seriously yet is not caught by individual extreme outliers. The common mistake is reporting average speed; the mean hides the 10% of users who are slow beneath the majority's fast numbers, whereas the percentile exposes that pain.
How it is calculated
nth percentile = the value at or below which n% of the sorted data falls
The output is in the data's own unit. The 50th percentile is the median; high percentiles (75, 90, 95) are used to see the worst experience and a skewed distribution.