Revolutionizing Path Analysis

A user logs onto your platform and starts navigating. He or she may sign up, then browse through a few products, make a purchase and sign out. This is a common path which thousands of users click through everyday day. Path analysis so far has helped us analyze sequences of events like these, visualized with the use of sankey diagrams.

But is that really enough to get in-depth answers to your business questions? The sankey diagram, which is basically a flow diagram, was innovative when it was invented by Matthew Henry Phineas Riall Sankey to visualize the energy efficiency of steam engines in 1898… It is however limited for the analysis of digital events, requiring a lot of time, money and effort to extract meaningful results, which is why most analytics providers don’t bother offering it.

CoolaData views path analysis as an inseparable part of every analytics dashboard, but we also changed the way analysts use path, breaking it down to illustrate the duration of each step in the path. In geektalk we think of it as “sankey meets sessionization.”

Each color here represents an event. The wider the color, the more time went by between events.

Now we can use path analysis to analyze loyalty, churn and ad ROI. CoolaData has built a widget which is comprised of two different layers: an initial layer we call Path Type, and a more in-depth layer called Path Version.

A sankey path can analyze top five paths to a purchase. But CoolaData’s path can analyze the top five paths of users performing three consecutive purchases. An example would be a user who purchased a shirt, then a day later purchased socks and finally also purchased shoes. All users who performed three consecutive purchases are visualized in one path to find what they have in common. Not only can we analyze these unique top paths, but we can also break them down to Path Versions, which are defined by any additional property we want to analyze.


So for example, we could break down a Path Type by duration, which means we would get groups of users who have the same duration time to events. A duration is the time it takes a user to go from one event to the next. These are our Path Versions. The image above illustrates a group of users who are spending a relatively high amount of time at the level up and checkout phase, represented by the pink and blue lines. These are the segments that are taking the longest to complete.

If it took a user 45 seconds to sign up after logging in to our website, 45 seconds would be the duration. This could help us recognize flaws, as flawed UX may be the reason for it taking too long to complete certain actions. On the other hand, we could also use this information to figure out where to promote campaigns: if users from the UK who performed three consecutive purchases have the shortest duration time of events and moved through our platform fast, they’re not around long enough to see ads. But if people from France are taking their time performing the same events, we might benefit from investing in French ad campaigns, since they spend more time on the platform and are exposed to more advertising. We can also see exactly where they spend the most of their time and target campaigns accordingly.

CoolaData’s path analysis can also pull the top five Path Versions out of all defined Path Versions. Each and every event in the path can be analyzed by average, median and distribution.

Contact us today to see a demo of CoolaData’s new path analysis, you’ll get real behavioral insights for almost any activity your users perform.

Share this post

Leave a Reply

Your email address will not be published. Required fields are marked *