We’re leveraging computational social science to map offline psychographic indicators with online behaviors. Our goal is to deliver the message to the right audience using more than basic demographics and to be able to target based on state of mind.
The ‘MPOWER Project is developing machine-learning/natural language processing tools that provide positive psychographic-targeted intervention to at-risk audiences across a wide spectrum of “Identity Vulnerability” susceptibility.
Distribution Model
After we have defined the psychographic and circumstantial indicators based on learnings, we will determine a weight for each based on the statistical significance for each among our sample set of investigative profiles. Data segments for each social recruitment challenge will be custom depending on the research findings for that sample set. Complete algorithms including psychographic and circumstantial indicators, social listening, predictive targeting, online user behavior and affinity for messaging will be built out from there and fed through a data management platform (DMP), similar conceptually to the way that brands use programmatic advertising.
Monitoring and Evaluation
Objective
Effectively monitor, analyze, and report on the macro VE environment as well as the specific impact of distribution and targeting of platform / campaign content – both immediately following intervention and downstream.
In regular intervals, integrate above data into distribution architecture to improve classifier balance.
Key Results
- Monitoring accuracy and effectiveness
- Data Distillation / Actionable insights
- Back-end data algorithmic oversight