The path analysis techniques lets you do things like:
- Find the top pages that new users open after opening the home page.
- Discover what actions users take after an app exception.
- Uncover looping behavior which may indicate users becoming stuck.
- Determining the effects of an event on subsequent users actions.
Path analysis is similar to other Analytics features, but offers some advantages:
- Like funnels, path analysis explores the steps users take through your site or app, but path analysis gives you ad hoc exploration of multiple paths vs. analysis of a single, pre-defined path.
- The users flow and behavior flow reports perform similar analysis based on pages, while path analysis can visualize paths using both page or screen views and events.
Create a path analysis
To create a new path analysis:
- Sign in to Google Analytics.
- Navigate to your property.
- In the bottom left, click Analysis.
- In the Analysis Hub, select an existing analysis or create a new one. Show me how.
- In Tab Settings, select the Path analysis technique. Alternatively, you can click + to add a new tab and select Path analysis from the list.
- Select the kind of data to use as the starting point of your analysis:
- On the right, under STARTING POINT, click to choose a dimension, or drag a dimension from the Tab Settings > NODE TYPE list and drop it in the STARTING POINT.
- Currently, you can choose Event name or Screen name node types. See How path analysis works below for more explanation of steps and nodes.
- Choose a value for the dimension.
Your new path analysis appears. On the left is the starting point you selected. To the right is STEP +1, which shows the top 5 screens your users viewed or events they triggered after that starting point.
See your users' next steps
To see the next steps your users took, click a data point in the graph. (Data points in a path analysis are called nodes.)
Click a node to expand it and add a new step. Click the node again to collapse it. Nodes that appear in gray are at the end of the user's path and can't be expanded.
By default, the graph shows the top 5 nodes in a step. Click + More to add up to 20 nodes per step. Additional nodes beyond the top 20 are grouped into an "Others" node.
How path analysis works
Path analysis uses a tree graph to illustrate the event stream, the collection of events users triggered and the screens they viewed. (Tree graphs are also called Sankey diagrams). A path analysis graph consists of the following elements:
- Starting point
- Node type
- Add node
The starting point is the screen or event that begins the path you want to analyze. It appears as the leftmost column in the visualization.
For example, begin an ecommerce path with a page that offers choices for different categories of shoes (e.g., men's, women's, athletic, etc.)
Steps are the columns in the graph. Each step after the starting point represents the successive actions of viewing a screen or triggering an event.
For example, STEP +1 is the list of screens viewed or events triggered by your shoe shoppers after opening the footwear product page starting point.
Nodes are the data points within steps, representing the number of users or events at that point in the path.
For example, the Men's Shoes node in STEP +1 represents the number of shoppers who opened that page, or the number of events that were triggered from that page.
Node type denotes the dimension values you'll see in each step of the graph. You set the node type for the starting point when you create a new path analysis. You can switch node types for a step using the menu above the step.
A path is a specific sequence of nodes occurring across one or more steps, within a specified time frame.
How paths are calculated
Path analysis takes the starting point you provide and then examines the event stream to find the next screens viewed or events triggered by your users immediately after that starting point. These screens and/or events are then aggregated together into paths. The numbers shown in each node represents the total number of users or events that contributed to that particular point in the path.
Paths and sessions
Paths are calculated from the user's event stream, using the first instance of the dimension value you select as the starting point.
Paths can span one or more sessions, depending on the date range you choose. A new session begins if a user is inactive for 30 minutes. If a path spans multiple sessions, the data for a node is an aggregation of all sessions.
Single session example
In the same session, a user opens the following screens:
Home > Product A > Home > Product B
The images below illustrate how path analysis visualizes this user journey, using the first instance of the Home screen_view event as the starting point, and then adding steps as you expand the nodes:
Initial view of the path showing the starting point and STEP +1.
Path with STEP +2 expanded.
Path with STEP +3 expanded.
Over the course of a week, in two different sessions, the user performs the following actions:
Session 1: Home > Product A > Product B > Home > Product C
Session 2: Home > Product A > Checkout
An initial view of this path for Event count looks like this:
You can explore this path in two ways:
To see the aggregated paths, expand the Product A node. This adds Step +2 and shows each session as its own individual path.
To see the complete cross-session path, expand all the nodes to show the 8 steps following the starting point.
Change the type of data shown in each step
Node type determines what kind of information to display in a step. Use the menu at the top of a specific step to display that node type in that step.
Change the analysis metric
By default, path analysis calculates the event count for each node in the graph. You can easily apply a different metric calculation:
To apply a metric:
- On the left, in Variables, select one of the supported metrics from the list and drag it to the Variables target in Settings.
Path analysis currently supports the following metrics:
The Event count metric counts the number of events triggered for each node of a path. Event count is the result of aggregating across all users and all sessions in the analysis time frame.
For example, when a user opens the home page, navigates to a product page, then returns to the home page before navigating to another product page, all within 30 minutes, the path shows two home screen_view events for the home page and one screen_view event for each product page in the first step.
The Total users metric represents the number of unique users who viewed a screen or triggered an event in the analysis time frame.
For example, when a user opens the home page, navigates to a product page, and then returns to the home page before navigating to another product page within the timeframe selected, the path will show one home page user on the starting point and one product page user for each product in the first step.
Include or exclude data from path analysis
You can customize the data shown in path analysis to focus on the most relevant information using the following options:
Select which nodes appear in the graph
To choose which nodes appear in each step of the graph:
- At the top of a step, click Edit .
- Use the checkboxes to include the screens or events you want to see in that step.
The values you select determine which individual nodes to display in that step. Unselected values are grouped into the + More node.
To remove a node from the graph, see exclude nodes below.
You can remove nodes if they aren't relevant. Excluding a node only hides it from the graph, but doesn't change how the path is calculated. To exclude data from the event stream before it is used to calculate paths, consider applying a segment.
To exclude a node from the graph:
- Right-click the node.
- Click Exclude node.
- Decide how to exclude the node:
- Selected only. This option excludes the node from the current path in the current step. Any subsequent steps along this path from this node are removed from the analysis.
- From all paths. This option excludes the node from any path in any step of the analysis.
You can see a list of the nodes you've excluded in Tab Settings, under NODE FILTERS. To re-add nodes you've excluded, hover over one of the filters and click X.
View unique nodes only
This option allows you to perform a path analysis that displays only changing values of your nodes.
For example, if you want to see how your users navigate through the pages on your website or screens in your app, and there are multiple events tracked for each page or screen, this option will display only one node for each even if the user performed many consecutive events on that page or screen. However, If the user then comes back to that page or screen after they visited another page or screen, it will show again in the path.
Segments let you define the subsets of users or events you want to include in or exclude from an analysis. For example, you can explore the paths taken by converting or non-converting users to find out where their user journeys diverge. You could also create a path analysis based on a subset of important events. This is particularly helpful if you have many events in a session but want to see the path across only a specific subset of them.
Learn more about building segments.
To apply a segment:
On the left, drag an existing segment from the Variables panel to the segment target in the Tab Settings panel.
Segments are applied to the event stream before the path analysis is calculated. This means that events or users you've excluded in the segment are not part of the analysis' event stream, and therefore are not part of the path calculation.
You can apply filters to the analysis based on any of the available dimensions and metrics. For example, you can show only paths with a minimum number of users or events, or paths that occurred on selected browsers or operating systems.
Filters are applied to the analysis before the paths have been calculated.