Behavior analysis displays the data about user behavior. It mainly analyzes the matters about when and where the app users do what actions, through which channels, and how long they take to do the actions. The analysis can help you understand the users’ operation rules, access paths, and the characteristics of behavior.
To check the behavior analysis report, complete the following steps:
- Log in to the console, select Products and Services > Mobile PaaS, and then select an application.
- On the left navigation bar, click Mobile Analysis Service > Basic analysis.
- Click the Behavior analysis tab.
Through the behavior analysis report, you can view the corresponding behavior indicators and charts in the following sections. Different behavior indicator analysis results are illustrated through charts:
- Hover your mouse cursor on the question mark of an indicator (), and you can see the description and calculation rule of the indicator.
- The statistical data is based on the historical daily aggregation. The daily aggregate result of all statistical items is stored in the database, and the data of the current day is not taken into aggregation.
- For more information about the indicator calculation rule, see Calculation rules for basic indicators.
A user’s active hours is subject to the time the user starts to use the app, which is calculated based on the time the log is reported minus the backend duration. Active hours show the time periods (with every two hours as a segment, for example, 2:00 - 4:00 am) when users use the app in a day starting from 0:00.
Click the Page name in this area to go to the detailed analysis of this page.
- Duration/day: Calculate the average daily length of time the user uses the app by dividing the total usage time of the app by the number of device IDs.
- Times/day·user: Calculate the average number of times the app is opened by a user per day by dividing the non-distinct data (PV) on the background by the distinct data (device ID) on the background.
Based on the active device/user tracking, calculate the top three channels for app downloads. When calculating, deduplicate multiple downloads of the same user in the same channel.
For example, user A downloads the app 1 time through channel C1, and downloads it 2 times through channel C2. When the data is classified, the downloads through channel C1 and C2 are only counted once respectively.
Startup speed counts the average startup time for the first startup and non-first startup, in seconds.
The first startup refers to starting an app for the first time after the app is installed; non-first startup refers to restarting an app after the first-time app startup and exit.
Page traffic indicates where the visitors of a specific page come from (source) and go to (destination). Based on the pid and the refer fields, the system calculates the sources of the current page and the proportion of each source to the total sources, and the destinations of the current page and the total proportion of each destination.
You can switch pages through the drop-down box on the upper right to view the corresponding page data. The pages displayed in the drop-down list are all pages added on the Page configuration tab page.
The active user tracking log sorts the users by region, and shows the top five regions with the highest proportion of distinct users. At the same time, the proportions of users in the first five regions to the total number of users are displayed.
- For world map, the proportion of regional distinct users is calculated by country.
- For map of China, the proportion of regional distinct users is calculated by province.