Shahin

Path Length

Advanced
Updated Jul 11, 2026 Shahin editorial 2 min read
PronunciationEN/pæθ lɛŋkθ/

Path length is the total number of steps in a path, including repeats and returns; unlike depth, it counts every move even when the user goes backward.

Path length counts all steps and is therefore the most sensitive metric to looping behavior: a path where the user goes back and forth between two pages many times has a large length without really advancing. This makes path length misleading on its own and it should always be read beside path depth; a large gap between length and depth signals loops and confusion. The distribution of path length, like depth, is right-skewed, and a tail of very long sessions pulls the mean up, so the median or a length histogram works better than the mean. The practical use of path length is in deciding how to analyze: the length distribution tells you how many steps a flow diagram needs to cover most behavior, and if most sessions are short a deep path report has little value.

How it is calculated

Path length = total number of sequential steps in a session including repeats and returns; usually reported as the median or the distribution of length across sessions in the range

Length counts every move, which is why it differs from depth; loops and returns raise length but not depth. The counting basis is the session and the unit is a step count. Because the distribution is right-skewed, the median or a histogram is more telling than the mean.