Funnel Analysis
Funnel Analysis is commonly used to observe conversion and churn at each step of a process. From a business perspective, you can configure the events that users trigger in sequence as funnel steps, and then find room for optimization by checking the conversion rates between steps.
Funnel Analysis can answer questions such as:
- New users of the day have dropped. Is there an anomaly anywhere in the process from download to registration?
- Which stage of the onboarding process has high churn and needs optimization?
- How are players on different servers progressing through the main story stages?
Definition of funnel conversion
Before you configure funnel steps, you need to understand what counts as completing a funnel conversion.
First, select an entity. Starting from the selected entity, the model calculates whether the same entity completes each funnel step in order, regardless of whether other events are triggered in between.
For example, suppose your funnel steps are A -> B -> C -> D:
- The behavior sequence A -> B -> C -> D counts as completing the entire funnel
- The behavior sequence A -> B -> E -> C -> D counts as completing the entire funnel
- The behavior sequence A -> B -> D counts as churning at step 3. Because event C isn't completed, event D is also excluded from the funnel
If a user has multiple behavior sequences that match the funnel steps within the selected period, the sequence with the most completed steps is used. If these sequences all complete the entire funnel or all churn at the same step, the earliest one in chronological order is used.
Quickly configure funnel steps
When you analyze the conversion process of new users, you can set events that are triggered in sequence, such as installation, registration, login, and onboarding, as funnel steps. If your business offers multiple options for a step and triggering any one of the events counts as completing the step, you can use Custom Events to combine two or more events into a single event.
By default, the model supports up to 30 funnel steps. To adjust the limit, contact your customer success manager.
After you configure the funnel steps, set the Conversion Window, which can be at most 180 days and at least 1 minute. Starting from the time the user triggers step 1, the model checks whether the user completes the subsequent funnel step events in order within the conversion window.
To get a more precise conversion funnel, you can also use Hold Property Constant. Users must then not only complete the step events in the specified order, but the property values of these events must also be identical. Events whose property value is null aren't used in the calculation.
For example, if you want to count the Product impression > Click > Purchase funnel, you can specify the product ID as the property to hold constant, which makes sure these actions happen on the same product.
You can also filter data or compare with group by. Note that when you configure a group-by item, you need to select the event property, user property, user cohort, or user tag of a specific step. If you add an event property of step 1, or select a user tag whose version is Match by date, the group value at step 1 in the "behavior sequence used in the calculation" is used to group the entity for the entire sequence. If the entity doesn't reach that step, the group value is null.
View overall conversion or trends
Funnel Analysis offers two views of conversion and churn. Choose one based on your analysis needs.
- Conversion Plot
If you want to see the overall conversion and churn within the time range, select the conversion plot. Each bar represents the number of entities that triggered each funnel step under the corresponding group value.
To make it easier to compare the data of different group values on the chart, the conversion plot shows up to 12 items at a time (including Totals), and shows the first 12 by default. You can select which group values to display from the Groups drop-down list in the upper-right corner of the chart based on your business scenario, or view the detailed data of all group values in the conversion table.
- Line
If you want to see the daily conversion and churn within the time range, select the line chart. Each line represents how the conversion rate of a group value changes over time.
By default, the line chart shows the conversion rate of All Steps (from step 1 to the last step, where the entity goes through the entire sequence). You can also switch to analyze how the conversion rate between any two steps changes, to pinpoint problems precisely. The table below the line chart can switch between Overall and By step to show more detailed step-by-step data.