When server-side events transmit bound device information, typically, the technology opts to bind the device information acquired at user registration (e.g., AFID). Subsequently, when server-side actions occur, they are transmitted based on this device information. However, microfinance products present a unique challenge where users frequently uninstall and reinstall or change devices for login. If subsequent actions are transmitted based on the device information bound during initial registration, it may falsely attribute these actions to the device from the earliest installation. From a third-party perspective, this data may appear as new installations without subsequent actions, all attributed to the user’s earliest device information and installation channel. This approach also affects retargeting ads, which cover a substantial number of reinstalled users but fail to convert, attributing actions to the channel associated with the initial device information. Considering this, adjusting the tracking logic to transmit based on the device information obtained during the last login ensures that reinstall data is attributed to the latest device information, significantly boosting conversion data for corresponding channels.
Following this logic, third-party data often reveals certain patterns: organic conversion rates are favorable, and retargeting campaigns perform well. This method of transmission allows for a more objective view of conversion data from third parties. Although many conversions stem from users reinstalling the app, any subsequent lending activity should still be considered valuable. Some clients face challenges with high costs due to tracking strategies, reminiscent of issues encountered in our past project in Pakistan.
In essence:
During third-party tracking, user information typically binds to the device information registered during initial installation rather than the device information at the user’s last login. For instance, a user with UID=111 may initially register with device information “aaa.” Over time, they may uninstall or change devices and log in again, resulting in new device information “bbb.” While backend actions are based on user behavior, conversion events reported to third parties reflect the initial device information “aaa.” However, our advertising efforts effectively recall users (even if they reinstall or switch devices), reflecting the last login device information “bbb.” Consequently, third parties may fail to report conversions under “bbb.”
Based on this logic, products with notable uninstall behaviors typically use tracking logic based on the device information registered during initial installation, resulting in high deployment costs and hindering new series launches. This phenomenon is particularly pronounced in microfinance, where users often uninstall before and after peak periods, then reinstall upon need.
In sectors like gaming and e-commerce, which exhibit higher user stickiness and fewer uninstallations, occasional errors may occur when users switch devices (e.g., changing phones) for products with extended lifecycles spanning several years.
Observable trends in advertising include:
1. Established data from older series maintains stability, while newer series struggle to launch, accumulating lower conversion rates and costs. The older the product (the larger its user base), the more pronounced this trend becomes.
2. Absence of retargeting conversion data (attributed to older channels).
3. Significantly better conversion rates and costs for older series compared to newer ones on third-party platforms.
Scientifically, updating tracking logic to transmit based on the last device information received allows for more frequent data transmission, aiding advertising learning and potentially attracting more users at lower costs (enhanced conversions benefit advertising learning). Some clients may also wish to ascertain the number of genuinely new customers, which can be achieved by adding tracking points such as “first loan/first purchase” to glean real initial conversion data and facilitate direct targeting of initial conversions.
If advertising costs appear disproportionately high, reconsidering tracking strategies from this perspective may provide insights into potential pitfalls.
