Context Insights comes at a unique time, when the proliferation of communications technology across the developing world has made it easier to reach broad swaths of the population at a lower cost than ever before. At the same time, cloud computing has rapidly brought down the cost of storing and accessing large amounts of data while advances in machine learning technology have made pattern recognition algorithms cheap and readily accessible for a variety of applications. As a result, we are able to reach more people, store more responses cheaply, and analyze their patterns more accurately and quickly.
Our work builds on a significant history of crowdsourcing for analytic insight in high-leverage settings. Most notably, the Good Judgment Project at the University of Pennsylvania popularized the science of forecasting through its broad forecasting tournaments, which revealed remarkable data about the habits and characteristics of super-forecasters. Prediction markets have been used in a variety of settings, from political betting odds to conflict predictions for military teams operating in austere environments. Multiple studies have shown that prediction markets and crowd-sourced forecasting can systematically outperform individual forecasters, even those who are professionally trained.
Context Insights is a business model innovation, taking the same logic of super-forecasting and applying it at the local level in emerging markets. Locals in these countries are often able to access informal networks of information that are non-digitized and inaccessible by other sources. While much of the research around forecasting has centered on strategies to improve the skills of individual forecasters based in the developed world, we propose that crowdsourcing from individuals in developing countries can have a similarly positive impact on predictive power.