Funnel Analysis

Overview

A funnel represents a sequence of user actions that lead to a goal in your product. For example, a purchase flow may include these steps: View Product → Add to Shopping Cart → Submit Order → Complete Purchase. This sequence forms a funnel. Funnel Analysis helps you understand where users convert or drop out across these steps. You can then improve your product or run operational campaigns to boost conversion and meet business goals. After running a funnel query, you can save the results as a report and add it to a custom dashboard for visualization and reporting.

The time range you select in the interface refers to when the first step of the funnel occurs.

Scenarios

Funnel Analysis measures conversion rates across key user journeys. For example, you can measure the overall conversion rate from viewing a product to completing a purchase and identify which step has the highest drop-off rate. If registration completion rates are low between entering details and finishing registration, you can investigate possible causes, such as failed SMS CAPTCHA delivery.

• Select an analysis entity: Device ID

• Select and configure funnel steps: Registration Page → Registration Success

• Set the conversion period: 1 day

• Select a time range

• Click Start Analysis22-10-23

How to Use

Page Layout

The Funnel Analysis feature consists of the following components:

1. Action toolbar: Save reports, view saved reports, and export query results.2. Configuration panel: Set events, metrics, properties, groups, and time ranges.3. Results area: View visual charts and detailed data after analysis.

How to Use

Select an Analysis Entity

In Funnel Analysis, you can link user behavior by Device ID, Account ID, or Entity ID. The dropdown menu displays Device ID, Account ID, and Entity ID. Device ID is selected by default.

• Device ID: A unique ID assigned by Quick Tracking to each device.

• Account ID: A user account ID that you pass through an API. It uniquely identifies your users.

• Entity ID: A user identifier generated by Quick Tracking. It uses ID-Mapping to link a Device ID and an Account ID, which lets you connect user behavior before and after they log in.

Add Funnel Steps

1. Click the dropdown menu to select an event for the funnel step. You can also assign an alias to the step.2. Add filters:

• Filter by event properties. For more information about categories and examples, see Appendix 2: Add Filters.

• Different property types support different filter operators. For a complete list and their definitions, see Appendix 2: Add Filters.

3. Set the conversion period:

• Define the maximum time allowed for a user to progress from the first step to the last. Only funnels completed within this period are counted as successful conversions.

• You can set the conversion period in minutes, hours, or days.

4. Add associated properties:

• You can use associated properties when multiple steps must share the same property value. For example, a View → Order flow requires the same Product ID for both steps.

• The associated property can be different across steps, but their values must match for the conversion to be counted.

Select Property GroupsGroup

You can group metrics by properties to compare performance. For example, you can compare conversion rates across different traffic channels using two properties.

• For a funnel to be included in a group, all of its steps must meet that group's property conditions.

• A user who meets the conditions for multiple groups is included in each of those groups.

• Group properties support dictionary-based grouping, such as City → Region.

Add Global Filters

When you select two or more events for your steps, you can use global filters to apply common filtering criteria across all of them. The filtering capabilities are the same as those for single-event property filtering.

Add User SegmentsSegment_20210422173239

You can use user segments to analyze specific audiences. For more information about creating a segment, see Audience Insights.

Select Time Range

You can choose a time range and granularity for your query. You can select either a relative or a fixed time range. By default, a relative time range is set to the past 7 days, displayed by day.

• A relative time range moves forward with the current date. Options include Past X Days/Weeks/Months and Since Launch. You can also define custom ranges, such as Past X Days/Weeks/Months. You can choose whether to include the current day. The rules are as follows:

A. Past n days: calculated by counting back exactly n full days from the current time.

B. Past n weeks: Count back n complete weeks from the current time. If the current time falls on the last day of a week, the period of past n weeks includes the week that contains the current time. For example: If the current time is 7.20 (Tuesday), the past 1 week is 7.12–7.18 (Monday to Sunday). If the current time is 7.18 (Sunday), the past 1 week is 7.12–7.18.

C. Past n months: This period is defined as the span of n full months before the current date. If the current date is the last day of the month, the period includes the month containing the current date. For example, if the current date is 7.20, then Past 1 month spans 6.1–6.30; if the current date is 6.30, then Past 1 month spans 6.01–6.30.

• For fixed time ranges, you can select start and end dates from a calendar or enter a number of days to define the range. Click Apply to run the analysis.

View ChartsRetail

After you configure your query and click Start Analysis, the results are displayed.

• You can click Export Data in the top-right corner to download the results as an Excel file.

• To save common metrics for reuse, click Save in the top-right corner.

* Overall conversion rate: The percentage of users who triggered Step 1 and completed all subsequent steps in order. This calculation is constrained by the selected time range and conversion period and considers all applied filters and grouping dimensions.

* Step-to-step conversion rate: The percentage of users who completed a specific step and then proceeded to the next one. This calculation is also constrained by the time range and conversion period.* Devices or logged-in users at Step n:

A. If n = 1: The number of unique devices or logged-in users who triggered Step 1 within the specified time range and met all filter conditions (both step-level and global), grouped by the selected dimension.

B. If n > 1: The number of unique devices or logged-in users who completed Step n-1 and then triggered Step n. The trigger for Step n must occur within the specified time range and conversion period, and the event must meet all filter conditions (both step-level and global), grouped by the selected dimension.

View Detailed DataAudience

The detailed data table shows the results for each step. You can click any row in the results to save that group of users as an audience.

Save Results as a ReportUser Count

You can click the Save button in the upper-right corner. In the Save Report dialog box that appears, specify or select the following information:

1. Enter a report name.
2. Select a time period for the report. The selected time period determines the data that is included in the report.

• Relative: The report uses the time range selected in the dashboard.

• Fixed: The report always uses the specified fixed time range, regardless of the selection in the dashboard.

• If no time scope is selected, the report follows the time selection of the dashboard by default.

3. Click OK to save the report to the report list.Add Report to DashboardFor saved reports, you can click Add to Dashboard. In the dialog box that appears, enter the following information:

1. Report name2. Target dashboard3. Chart type4. Chart layout in the custom dashboard5. Click OK

Funnel Calculation Logic

1. Define funnel steps

A funnel consists of two or more steps that users must complete in a specific order.

Each step is defined by one event and can include zero or more filters.

2. Set time range and conversion period

A funnel requires a defined time limit for completion. Only funnels completed within this period are considered successful. You must also define a time range for the analysis, such as the past 7 or 30 days.

• Time range: The window for the analysis, such as the past 7 days.

• Conversion cycle: The conversion cycle can be set in minutes, hours, or days.

Example: Assume the funnel is Home Page → Select Payment Method → Payment Successful. The analysis window is Day 1 to Day 3. A user's behavior is as follows:

• With a 24-hour conversion period, if a user starts on Day 2 at 23:00, they must complete the final step by Day 3 at 23:00 for it to count as a conversion.

3. Apply funnel filters

You can apply event-specific or global filters to funnels. The logic is as follows:

A. You can filter events using event and global filters.

B. Global filter conditions apply to all steps. C. If you configure both event-specific and global filter conditions, the system combines them using AND logic: Event filter conditions & Global filter conditions. For example, assume that a funnel has the following steps: Visit home page, Select payment method (filter condition: Payment method = WeChat), and Payment successful. The following examples show the behaviors of different users:

User 1: Access the home page → Select Alipay as the payment method → Select WeChat as the payment method → Payment successful

User 2: Access the home page → Select a payment method (Alipay) → Access the home page → Access the home page → Payment successful

User 1 is counted in the final conversion funnel, but User 2 is not.

4. Group and compare funnels

The grouping feature in Funnel Analysis lets you add group filters to query conditions, facilitating the comparison of funnels across different groups.

A. Supports grouping prebuilt and custom properties that are common to all (step)s.

B. A group funnel is complete only if the properties of all steps satisfy the group’s properties.

C. If you meet the criteria for multiple grouping funnels within a time range, you are assigned to multiple groups.

Example:

Group the funnel by the custom property Product Brand. The funnel steps are: View Product → Add to Shopping Cart → Pay. The user behavior is as follows:

Example 1: View Product (Apple) → View Product (Nokia) → View Product (Samsung) → Add to Shopping Cart (Samsung) → Pay (Samsung)

Example 2: View Product (Apple) → View Product (Samsung) → View Product (Apple) → Add to Shopping Cart (Apple) → Pay (Apple)

Example 3: View Product (Apple) → View Product (Samsung) → Add to Shopping Cart (Samsung) → View Product (Apple) → Add to Shopping Cart (Apple) → Pay (Apple) → View Product (Nokia) → Add to Shopping Cart (Nokia)

Example 4: View Product (Apple) → Add to Shopping Cart (Samsung) → Pay (Samsung)

In the preceding example, the number of conversions is:

5. Interpret funnel metrics

Calculate the number of users at each stage, and then calculate the conversion rate.

A. Number of devices in step n / Number of logged-in users:

If n = 1, an event from step 1 that meets both the event-specific and global filtering conditions is triggered within the time range, and the deduplicated device count or login user count is calculated based on the grouping dimension.

If n is greater than 1, the metric counts the number of unique devices or logged-on users—grouped by a specified dimension—that triggered step n−1 and then triggered the event for step n. The trigger for step n must occur within a specified time range and within a defined time period after step 1 was triggered. In addition, the event must satisfy both event-specific and global filter conditions.

B. Overall conversion rate = Unique devices or logged-in users at final step ÷ Unique devices or logged-in users at first step

C. Step-to-step conversion rate = Unique devices or logged-in users at Step n ÷ Unique devices or logged-in users at Step n−1