Analytics use cases
Last updated
Last updated
The goal of each game developer is to keep users in the app as long as possible and try to understand the reasons why they churn to prevent it.
That is why we try to find out which actions users do before they leave the app. For an analytical platform, it can’t be calculated because when the user stops using the app, the connection with SDK is lost and it can’t send any events to the platform.
But still, you can use several ways to learn about the actions of these users.
To see which events prevented churn:
Choose Churn in the Event dropdown menu.
Choose Flow direction – to the event.
Select basic and custom events for the report.
After you build the report, you will see the most recent event that leads to the Churn. If you want to see more previous events, just tap the event to expand it.
Another way is to process data manually (e.g via Excel). To do that, you need to export data in the Tuning -> Raw Data Export section.
Export Custom events and Sessions there.
After that, do the following:
Using the Sessions data, find the time of the last session of the user and select only users who churned (e.g. whose last session was 3 days ago).
Merge information about all custom and basic events and find the last event of the user.
Link both tables by the devtodev ID field to find out what was the last event.
You can do the same analysis in the SQL Wizard section. Here’s how:
Select in one query the Last_Seen parameter in the Users table and Date parameter in each event table.
Find the most recent event for each user and filter the list by the Last_Seen date to choose only users who churned.
The last option would be to view the actions of specific users one by one. It can be done in the Users section.
1. Filter users by the Last seen parameter which is calculated automatically. You can also add more filters to specify those users you want to explore.
2. After you apply the filter, you will see the list of users. Open their card to see the Last user activity tab. There will be the latest events that were performed by the user.
50-100 users will be enough to find patterns in their behavior.
Using these ways you can explore the last user's activity in the game. All of them have advantages and disadvantages, but please note that sometimes the last event alone is not enough to understand the reasons for churn. Especially if you use a lot of custom events in research. As a result, you can get meaningless events, such as Open the Settings tab, View the chat, etc. That is why it is better to check more events at once to see the whole picture.
It is necessary to take A/B tests during the development stage. So how is it possible to make A/B tests on the devtodev platform?
Here you can find the manual on how to run A/B tests in devtodev:
A/B testing
