Mobile Information Delivery Using Projection Models for Offline Response Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for delivering information to mobile devices face inefficiencies in tracking and measuring off-line responses, as current performance metrics are often insufficient or inaccurate, particularly when off-line conversions are the primary outcome.
Innovation Solution
The system employs a geo-fencing technology that defines virtual geographical areas based on real-world boundaries and user interactions, using spatial indices and meta-data to enhance the accuracy of location-based information delivery and track off-line responses by projecting targeted response rates through frequency modeling and panel-assisted methods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If on-line activities are tracked to measure performance, then information delivery can be controlled, but measurement precision is insufficient for off-line conversions
Solution Approach 1:
The system performs preliminary actions by delivering information to mobile devices and recording on-line activities in advance. It then uses projection models to estimate off-line responses that will occur later, allowing performance measurement to extend beyond immediate on-line actions to include delayed off-line conversions.
Solution Approach 2:
The system introduces projection models as intermediary tools that bridge the gap between on-line activity tracking and off-line response measurement. These models act as mediators that translate observed on-line behaviors into estimated off-line conversion metrics, enabling accurate performance measurement across both on-line and off-line domains.
2Measurement precision
If location-based information is delivered to mobile devices, then relevance is improved, but measurement of actual user response becomes difficult
Solution Approach 1:
The system implements feedback mechanisms by continuously monitoring on-line activities and using this information to refine projection models. The measured off-line responses feed back into the system to improve future projections, creating a closed-loop measurement system that progressively enhances accuracy in detecting and measuring user responses.
Solution Approach 2:
The system replaces direct mechanical tracking of off-line responses with a computational projection model. Instead of physically tracking users offline, it substitutes a mathematical model that predicts off-line conversions based on on-line behavior patterns, making measurement feasible without direct observation.
3Measurement precision
If sponsored information is delivered based on location, then information relevance is improved, but accuracy in measuring return on investment decreases
Solution Approach 1:
The system performs preliminary actions by establishing projection models before evaluating ROI. These models are trained on historical data and are ready to accurately predict off-line responses when evaluating investment returns, ensuring reliable ROI measurement from the outset.
Solution Approach 2:
The projection models serve as intermediaries between sponsored information delivery and ROI measurement. They translate delivery metrics into meaningful performance indicators that accurately reflect return on investment, bridging the gap between information delivery actions and financial outcome measurements.
Data Source
AI summary
A system for processing information requests associated with mobile devices comprises an information server configured to build a search query for an information request based on the location features and other data therein and to search an information database for matching information documents. The matching information documents including information documents having different types of performance measure, including a first document using an impression-based performance measure, a second document using a click/call-based performance measure and a third document using an off-line site-visit-based performance measure. The information server is further configured rank the matching documents based on their respective performance measures and to select a matching document based on their respective rankings. The information server is further configured to generate a projected probability of an off-line site visit in response to the second document being selected to fulfill the each request and impressed on an associated mobile device, and to adjust a budget of the second document based on the projected probability.