Dynamic Geo-Fence Design Using Performance-Driven Geo-Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing location-based targeting systems face challenges in accurately translating location signals into meaningful signals for serving relevant information to mobile device users, as static geo-fences often fail to capture user intentions and demographics effectively, leading to unpredictable user volumes and limited performance optimization.
Innovation Solution
The system partitions urban areas using geographical information of transportation routes and natural boundaries to create geo-blocks that are relevant to specific human activities, and uses mobile device signals to fine-tune and enrich these geo-blocks for dynamic geo-fence design, optimizing performance metrics while ensuring a sufficient user volume.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If static geo-fences are used for location-based targeting, then the system structure is simple and easy to implement, but the ability to capture user intentions and demographics effectively deteriorates
Solution Approach 1:
The patent divides the geographical space into discrete blocks (e.g., 100m x 100m tiles or zip code-based blocks) rather than using continuous static geo-fences. Each block is independently analyzed and scored based on multiple signals, enabling precise user intention capture while maintaining systematic simplicity through modular block-based processing.
Solution Approach 2:
The system dynamically changes the parameters used to define and evaluate geographical blocks. Instead of fixed geo-fence boundaries, the system adjusts block scoring based on multiple signals including location history, device attributes, and contextual data, allowing the effective targeting parameters to adapt while the underlying block structure remains simple.
2Reliability
If large safety margin of reachable user volume is ensured in geo-fence design, then the reliability of user volume prediction is improved, but the space for performance optimization deteriorates
Solution Approach 1:
The patent applies different scoring weights and evaluation criteria to different geographical blocks based on their local characteristics. High-performing blocks receive higher scores and are prioritized for targeting, while maintaining sufficient overall user volume through the aggregation of many blocks. This allows performance optimization at the local block level while ensuring reliability at the aggregate level.
Solution Approach 2:
The system dynamically adjusts the composition of the geo-fence by selecting and reweighting individual blocks based on real-time performance data and changing user behavior patterns. This dynamic block selection enables continuous performance optimization while maintaining reliable user volume through diversified block participation.
3Loss of information
If zip code level partitioning is used for location-based targeting, then demographic data availability is improved, but the precision of location-to-signal mapping deteriorates
Solution Approach 1:
The patent further segments zip code areas into smaller functional blocks (e.g., 100m x 100m tiles or activity-based zones) while retaining zip code-level demographic data for each block. This hierarchical segmentation allows precise location mapping at the block level while preserving the demographic information availability from the zip code level through data aggregation and inheritance.
4Measurement precision
If 1 billion 100 meter by 100 meter virtual tiles are used to cover the world, then the granularity of location partitioning is improved, but the complexity of processing and the challenge of determining location intent deteriorates
Solution Approach 1:
The patent applies enhanced processing and multiple signal analysis only to blocks within active geo-fences rather than uniformly processing all 1 billion global tiles. Each block within the geo-fence is evaluated with locally-relevant signals and criteria, reducing overall processing complexity while maintaining high granularity precision where it matters most for active campaigns.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
System and methods for running a location-based information campaign (campaign) select one or more first geo-blocks to form a first geo-fence from a plurality of geo-blocks each corresponding to a geographical region having at least one border defined by a real- world object and overlapping substantially with a targeted region associated with the location-based information campaign, and process information requets with respect to the first geo-fence. Each of the one or more first geo-blocks is associated with a respective performance score above a first threshold. The system and methods further monitor a pacing status associated with the campaign, and in response to a pacing goal associated with the campaign not being met, define a second geo-fence for the campaign, the second geo-fence including the one or more first geo-blocks and one or more second geo-blocks each associated with a respective performance score above a second threshold that is below the first threshold.