In most cases, a user has only one device and you need to achieve the following goals:
Understanding User Identification and User Association
Understanding the user association schemes supported by Sensors Data
Determine the association scheme for the products to be integrated.
Accurately identifying users and establishing user associations are fundamental to user behavior analysis. Even the slightest oversight will cast doubt on subsequent statistical or analytical results; therefore, please pay close attention to this section.
User ID and Associations
In the real world, we accurately identify a natural person using their ID card number. However, this method is no longer applicable in internet business systems. This has led to two common solutions. The first is to uniquely identify a user using their device ID. This method solves the user identification problem to some extent, but its drawbacks are also obvious. For example, if the same phone is used by multiple users, their actions will be attributed to the same "person"; and an old user switching to a new phone will also be identified as a completely new user; these problems ultimately lead to inaccurate data analysis results. The second solution is to identify users using their account or customer ID. This solution is more common in business backend systems, but actions taken by users who are not logged in cannot be identified. Therefore, this solution can only accurately record business data and mainly serves as a user identifier in the construction of business data warehouses.
Sensors Analytics has thoroughly considered the advantages and disadvantages of each solution and developed a comprehensive user identification system. For more information on how Sensors Analytics identifies users, we recommend reading "How to Accurately Identify Users - Basic Concepts" . Simply put, when a user is not logged in, Sensors Analytics selects the device ID as a unique identifier. When logged in, it selects either the login ID or the userid. A user has both a device ID (also known as an "anonymous ID") and a login ID. Next, through "user association," the device ID and login ID of the same user are linked together. This way, regardless of whether the user is anonymous or logged in, we can accurately identify that it is the same user. This is currently the most universal and accurate user identification method.
Sensors Data's User Association Solution
In most cases, a user has only one device, so a user will only have one device ID and one login ID. To address this, Sensors Data provides a one-to-one association solution between device ID and login ID. For details, see: One-to-One Association Solution .
For users with multiple devices, Sensors Analytics provides a many-to-one association solution between multiple device IDs and a login ID. See: Many-to-One Association Solution for details. Furthermore, this association solution also applies to products with multiple platforms. For example, if there are mini-programs, apps, and web platforms simultaneously, the user's device ID will be different on each platform, which is equivalent to a user having multiple devices.
One-to-one and many-to-one association schemes each have their advantages and disadvantages. Sensors Data recommends using the one-to-one association scheme by default. Of course, in some cases, products may not have a user account system, such as notepads and alarm clocks. In such cases, we can choose to use only the device ID as the user identifier. See: Using Only Device ID .
Determine the association scheme for the products to be integrated
After understanding the association solutions supported by Sensors Analytics, you need to determine the association solution for the products you are integrating. The main points to determine include the following:
Determined Items | Description of Specific Items |
|---|---|
| Association method | You need to choose between a one-to-one or many-to-one association scheme. The system defaults to a one-to-one scheme. To enable a many-to-one scheme, please contact the relevant customer service representative. |
| first_id | Sensors Data Center defaults to selecting device IDs on each platform. For reference, considering the needs of some customers, you can choose to replace it by calling the identity interface. |
| second_id | The user_id is usually chosen, but other unique IDs that can identify a user can also be used. |
| Timing of association | Typically, association is performed when a user registers, logs in, or logs in via third-party services. Additionally, after initializing the SDK, it's also necessary to call Sensors Analytics' "association interface". |
It is recommended that both the business and technical teams participate in confirming the ID association scheme. After confirmation, the business team can then proceed to the next step: Step 3: Data Collection Scheme Design . Technical personnel can also begin preparatory work, including SDK embedding and initialization, and the implementation and testing of the ID association scheme. See Step 4: Basic Data Verification for details.