At Microsoft Research’s annual technology demo day this week, researchers showed off a tool called Viral Search that attempts to measure virality in its more literal sense. That means not overall traffic over time, but the mechanics by which it passes from person to person over many generations, MIT Technology Review reports.

The software looked at 1.4 billion Tweets over the course of a year and produced a branching tree that shows how links spread, based on who follows whom and the sequence of retweets. From this, researcher Jake Hofman and a team created a new kind of Web metric that measures the average distance on the tree between any two people who tweeted the story.