Campaign prediction algorithm


Machine Learning algorithms determine the best campaign idea for a Global Marketing Research firm.

IndustryMarket Research

Year2013 – Ongoing

RoleMachine Learning, CI/ CD, Microservices

TechnologySpring boot, Grails, RabbitMQ, Azure Cloud platform, Teamcity

Predict the performance of ad campaigns

$6 bn Global Marketing Research firm was looking to predict the performance of ad campaigns to help their customers achieve better marketing ROI. A failed marketing campaign results in huge loss, sometimes even upsetting business growth plans.

Machine learning solution fosters stronger ideas

We developed a real-time campaign analytics platform, which leverages predictive analytics to quickly test an unlimited number of potential campaign message, and design options, helping one to get closer to producing ideas customers totally love. This entire process was powered by a Machine Learning algorithm that interprets how one’s customers react to different ideas. First, a targeted group is selected to evaluate creative execution. They are presented with executions and asked to choose the ones they prefer. Based on real-time consumer response, the Machine Learning algorithm breeds stronger executions while killing off weaker ones. This “survival of the fittest” process continues until ultimately the strongest and the optimal executions survive.


Smart platform improves business competitiveness

The feature to predict the success of a marketing campaign was unique, was one of its kind in the market. This significantly improved the overall product competitiveness that resulted in the increase of platform adoption. The leadership position that our customer already enjoyed was wired further.