Every action, every ad impression is a set of tags
Merge massive tags, use programmatic buying to maximize advertising value
Performance improvement = Data segmentation
App Info: APP ID, APP Name, Category, Rankings, Context...
Audience: User ID, Age, Gender, Real-time interest, previous interest...
Creative: Creative ID, Size, Ad type, Dynamic Creative...
Each combination will affect the ad tag. AdTiming will programmatically find the relationship between them, and automatically optimize it
Moderate ads frequency: people will be bored with silly & frequent ads
Explore the scenario: people are in different scenario at work or entertainment
Network Monitoring: Wifi and mobile network play inconsistent role making ads impression;
Conversion window: each conversion scenario has the best timing to match ads
Connect all data sources, internal and external network data hybrid computing, connecting to multiple big data clusters, such as Hive, HBase, MongoDB, etc. Enable hybrid computing across these clusters without ETL.
Even with stricter requirement for network security and data privacy, internal network data without external memory could be calculated together with external network data.
200+ Countries and Regions
10 Billion+ daily traffic coverage
250 Million+ daily data update
20,000+ audience tags
Using data Analysis and Ads delivery free of charge during test period