Skip to main content


The TapDB SDK has a number of preset reporting events that allow us to build a very robust and complete operational module. But as you dig deeper into data analysis, there are obvious limitations to a fully templated, preconfigured analysis model. We designed custom event analytics to give you the freedom to report and query the data you need. There are some barriers to access and understanding, but the ceiling for digging into player behavior with custom event analysis is significantly higher than the pre-built report template, so we highly recommend learning and using it.

Event Analysis

Event analysis is the most basic analysis model, and its analysis object is events.

  • See all the information about the event, such as:
    • See the total number of times the "Play PVP" event was triggered, the number of users triggered, and the number of times per capita.
    • The total amount of purchase and the number of times per person for the "Buy Gift Package" event.
    • The number of "card draw" events per capita in different provinces in the last 7 days.
    • The proportion of TapTap logins and the proportion of wechat logins in daily "account logins" events.

Retention Analysis

Retention analysis is a model that analyzes the generalized retention behavior of users, and its analysis object is the device/account.

We can observe the user who triggered event A and then triggered event B. The initial event and the return event can be the same, for example:

  • Check the "Login" status of "Paid" accounts within the next 7 days.
  • Check if a "Level 10" account "pays" within the next 7 days.
  • Check the status of "rewards" within 30 days for accounts that have "purchased a monthly card".
  • Check the "login" status of accounts that triggered "login" for the first time and "login" within 90 days afterwards: this is what we usually call "account retention".

Funnel Analysis

Funnel analysis is a model for analyzing users who trigger specific events in a sequence. It is analyzed for devices / accounts.

  • To view information about those users who converted strictly according to the funnel steps, you can set the funnel window period, for example:
    • The number of accounts that "completed check-in" and the number of accounts that "engaged in PVP" within 30 minutes after "completed check-in".
    • The number of accounts that "purchased a monthly card" and the number of accounts that "purchased a monthly card" and then "purchased a newbie pack" within 7 days after "purchased a monthly card".

User Segmentation

User segmentation is an extremely powerful feature in custom event analysis. Here's how it works:

  • Find a group of users to be analyzed by looking at the data, e.g. users with very low retention rate, high ARPPU, or consistently signing in but with very low rank, and save them as a user segment.
  • Using this user segment as a condition, we observe or filter the behavioral data to try to find some behavioral characteristics, e.g., users with very low retention rates have Android versions below 7, presumably because of the large number of emulator devices included in the statistics.

User segmentation opens the way from finding features to digging deeper, and is a necessary skill for custom event analysis problems.


Dashboard is suitable for observing data. It is often used in scenarios such as "forming a daily report of core metrics" and "continuously observing a specific issue". Once you have saved a report in your analytic model, you can add it to Dashboard. Dashboard core ability is sharing: you can share your Dashboard to other users to reduce communication cost. Note that a good Dashboard should have a clear topic to avoid the distraction of combining unrelated information into one Dashboard.