Prediction Model for Mobile App Installation Targeting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Content delivery systems often drop requests from end-users with unknown identities, resulting in a significant percentage of missed opportunities to present relevant content, as they cannot make informed targeting decisions without complete user profile information.
Innovation Solution
A prediction model is trained to compute the likelihood of application installation or activation based on features such as preloaded applications, categories, and network connections, allowing for content item selection and targeting of end-users without requiring known identities, and enabling more relevant content delivery to a broader audience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If content delivery systems require complete user profile information for content selection, then content targeting accuracy is improved, but the number of droppped requests increases
Solution Approach 1:
The system performs preliminary actions by collecting and storing user profile information, device characteristics, and content interaction data in advance through user agents and tracking systems. This pre-collected data enables the content selection system to make informed decisions even when complete profile information is not immediately available, reducing dropped requests while maintaining targeting accuracy
Solution Approach 2:
The patent introduces intermediary components including user agents that collect user information, content delivery exchanges that mediate between content providers and publishers, and prediction models that bridge the gap between incomplete data and content selection decisions. These intermediaries enable the system to handle requests with partial information while still delivering relevant content
2Reliability
If downstream exchange systems filter requests from users with unknown identities, then content provider quality control is improved, but content delivery opportunities are lost
Solution Approach 1:
The system changes the parameters used for request validation by shifting from requiring complete user identity information to accepting multiple alternative parameters including device characteristics, content request context, prediction model scores, and partial profile data. This parameter flexibility allows the system to process requests that would otherwise be filtered out while maintaining quality control through multi-factor evaluation
Solution Approach 2:
The system performs preliminary quality assessment through prediction models that evaluate request validity, user intent, and content relevance before the main content delivery decision. This pre-evaluation enables downstream exchange systems to accept requests from users with unknown identities while maintaining quality control through automated scoring and filtering mechanisms
Data Source
AI summary
Techniques for optimizing content item delivery for installations or activations of a mobile application are provided. In one technique, a machine-learned model is trained based on multiple training instances that individually indicate whether an entity performed a particular action relative to a mobile application. In response to receiving a content item request from a third-party content delivery exchange, it is determined whether a client device that initiated the content item request has activated a particular application. In response to determining that the client device has not activated the particular application, multiple feature values of the content item request are identified. Based on inputting the feature values into the model, a score is generated that indicates a likelihood that an entity of the client device will perform the particular action relative to the particular application. Based on the score, a content item is transmitted over a network to the client device.