Data Model Understanding and Organizing Needs

After completing the preliminary preparation, you need to understand and analyze your project, understand the metrics you need to monitor and the events you need to analyze, and decide what data to transfer; Before we begin, we will introduce you to the Event & User model, which is critical for your understanding and use of the TA platform.

1.Introduction of Event & User Model

An Event represents a meaningful behavior of the user, such as the user adding a product to the shopping cart, browsing a video, etc. An Event mainly contains two parts of information, one part is used to describe how the behavior occurs. There are mainly the name of the behavior (What), the user who generated the behavior (Who), and when it was generated (When); the other part is the properties of the behavior, such as the name of the video browsing the video, or the payment amount in the paid event. These properties are the main object of analysis and also need careful consideration. We will give suggestions for property setting in setting user features and event properties.

User is used to describing the latest status and fixed properties of each user, such as the user's ID, registration time or accumulated payment amount, etc. The most important thing is to identify a user's ID. If you want to know how TA recognizes users, you can view the user identification rules. Through User properties, you can quickly filter out the users you want to analyze during user behavior analysis. For example, if you want to analyze the activity of paid users, you only need to analyze users whose value of the user feature 'cumulative payment amount' is greater than 0.

2.Organize Requirements

After completing the preliminary preparation, you can begin to prioritize and analyze requirements, and if you are not familiar with the process, follow the following steps to prioritize requirements:

2.1 Clear-based Analytical Indicators:

We suggest that you start with basic analysis indicators at the beginning, such as registration, login, payment, etc. If you do not have special analysis requirements for these behaviors, it is recommended to upload data using the interface provided by us. If you have special analysis requirements, we also recommend prioritizing the selection of these indicators.

2.2 Identification of Tracked Events:

After determining the basic indicators, it is recommended that you consider the importance of each system, function, and analysis point in the application, and then convert the play and analysis points that need to be analyzed into the form of an Event.

If you have doubts about how to translate the analysis requirements into the form of Events, you can refer to the following suggestions we give:

• For important behaviors, it is recommended to set as one Event, and then set the properties according to your analysis point.
• For unimportant behaviors, such as behaviors that only need to analyze the number of participants and the number of participants, you can set multiple such behaviors as an Event, and then identify specific behaviors through properties, such as setting an Event to be called a secondary event, and then through an property Event typeto explain which behavior is tracked.
• If you pay special attention to certain key behaviors, such as one-click purchases in paid purchases, you can track this behavior as an Event separately.

We strongly recommend that you organize all events in the form of a doc (such as an Excel table). The doc needs to contain the event name, description, importance, and analysis points; this doc can help you better communicate with developers who are implementing the data reporting code.

2.3 Determine the user identification fields:

TA identifies a user that requires two IDs, Account ID and Guest** ID**. We recommend that you use this Account ID directly to use the Account ID in your database. If you want to configure the Guest ID, please configure the Guest ID before uploading the event.

For different data sources, such as iOS and Android, Account ID must use the same system. If you upload data through both client SDK and background access, make sure the Account ID system is consistent.