Qt Insight Basics Part 4: Getting Started 3 - Funnel Analysis and Crash Analysis

This is a translation of the Japanese blog series by Mikio Hirai.

In our previous post, we discussed how to use Qt Insight to check user flow and bounce rates and compare data. In this article, we will continue with an introduction to Funnels.

  1. Funnels - Is the GUI being used as intended?

  2. Application failures - How often do abnormal terminations occur and what are the causes?

  3. Summary



Funnels can be used to obtain information on how many users are transitioning through the screen as intended by the product developer within a time limit. The name "Funnels" is derived from the fact that only a limited number of users pass through a specified path, just as a funnel gets narrower as it goes down.

Now let's open "Funnels" and open the "Example funnel" in the "All" tab. This example funnel is a pre-created demonstration funnel, but you can usually create your own customized funnel.

The following screen will then appear.

First, we will discuss "Funnel conversion" in the upper left corner.

Funnels allow you to configure the screens the user is expected to navigate through, and the user steps used for screen transitions.

Three screens are defined below: "Enters coffee-selection-screen", "Enters espresso-selected-screen" and "Enters insert-cup-screen".

In this diagram, we can see that out of the 154,319 sessions that reached the "coffee-selection-screen," 47.88% or 73,893 sessions proceeded to the "espresso-selected-screen." Furthermore, from there, only 11.94% or 18,429 sessions of the total 154,319 sessions reached the "insert-cup-screen".

On the right side of the graph, you will find "Total completion percentage," "Funnel completions," and "Percentage of sessions that started the funnel."

These metrics have the following meanings:

  • Total completion percentage: The percentage of users who reached the first step of the funnel and completed all defined steps.
  • Funnel completions: The total number of users who completed all defined steps among those who reached the first step of the funnel.
  • Percentage of sessions that started the funnel: Percentage of users who reached the first defined step in the funnel and started the funnel.

The two metrics below, "Completions per session on average" and "Sessions that started the funnel," have the following meanings.

  • Completions per session on average: The number of the user-initiated sessions (sequence of GUI application usage) that actually completed this funnel.
  • Sessions that started the funnel: Of the sessions initiated by the user, the total number of sessions that initiated this funnel.

Funnels are defined as follows. Unfortunately, new funnels cannot be created in the demo organization. Also, since you have not yet registered an application, we cannot do this together, but hopefully this information is helpful to you.

First, select "Funnel" and click "Create a new funnel".

Then, the following screen will be displayed. Set the name of the funnel, the target application, and the time limit for completion of the funnel, respectively.

Once you have completed that, it's time to define the steps.

Steps can be defined both by the screen and user actions. In this case, we have added four steps to the Funnel. Once you have finished adding them, you can create the Funnel by clicking on "Create funnel" in the top right corner.

As a result, the current data shows the composition of the Funnel's results. Unfortunately, it appears that no users have reached the end with this configuration.

Application failures

Let's dive into the details of "Application failures" which provides insight into the number and causes of abnormal application terminations. Within Application failures, there are two subcategories: "Overview" and "List of Failures".

Let's start by exploring the Overview section.

First, let's take a look at the following four items. Each one is described below.

  • Number of application failures
    The total number of sessions that experienced abnormal terminations.
  • Devices with application failures
    The total number of devices on which the application experienced abnormal terminations. For example, if the application crashed multiple times on one device, it will still be counted as one.
  • Application failure rate
    The percentage of sessions that experienced abnormal terminations.
  • Devices with application failures rate
    The percentage of devices that have experienced at least one failure.

Below that, you'll find "What leads to application failures". This section shows the user flow that leads to abnormal terminations of the application. From here, you can gather information about which actions are causing the abnormal terminations.

Next, let's take a look at the "List of failures".

When you select it from the menu on the right side, a table will be displayed as shown below.

This table shows the data for each row corresponding to the abnormal terminations, and by clicking on the arrow icon on the leftmost side of each row, you can view the detailed information for each data.

Here, we have opened the abnormal termination that occurred in the "insert-cup-screen".

By examining the expanded information, we can see that it was an abnormal termination that occurred in the Netherlands on October 30, 2023, at 18:14. The device model is "Model1", and the software version is 1.0.3. Furthermore, the resolution is 1024x768.

Below that, the user actions performed before the abnormal termination are displayed along with timestamps.

Using this information, you can gain insights into the types of abnormal terminations occurring in the market and their reproduction conditions.


So far, we have introduced the information available in Qt Insight version 1.4. Have you gained an idea of how to improve your application using these?

In the next post, we will show you what kind of analysis is possible with this data. Stay tuned!

Articles in this series

Blog Topics: