What News Corp/MySpace Should Have Told Reps. Markey and Barton

Yahoo isn’t alone in not being candid to Congress.  Here’s what News Corp should have also said to the Congress about its data collection, profiling and targeting system. 

It should have informed Congress how its Fox Interactive Media (FIM) data-mines its users daily.  From its data-mining company: “FIM operates some of the highest-traffic websites in the world, including MySpace, and serves over 5 billion online ads across its sites per day.  Each of these ads is optimized and targeted to specific audiences based on analysis of web traffic, user behavior and click patterns.  As part of our targeted ad serving platform, we analyze nearly 2,000 identifying variables for each of the millions of visitors to our sites every day (tracking and analyzing, for example, whether a visitor likes jazz, but also whether they respond more to car ads than to pizza ads).  While our targeting process was already one of the most advanced in the industry, we were eager to improve ad click-through rates further by fine- tuning targeting.  To get to this next level of targeting precision, we needed to analyze massive volumes of data to discover patterns and identify relevant targeting criteria across segments and demographics.  FIM implemented Greenplum for its parallel, multi-core architecture that could scale to support our massive data volumes, but also because the Greenplum data warehouse allows data analysis to be performed directly within the warehouse 

– instead of having to extract it first…our team can execute lightning-fast queries against a matrix that is 4 billion rows tall and 1 million rows wide, running tests against thousands of variables for each for the 5 billion ads FIM serves to visitors each day. What’s more, we can now complete 10,000 experiments against 20 million site visitors in just three hours. Previously, it took an entire day just to extract the data – and then another whole day to run the tests.  The result has been faster and more efficient data analysis, which has in turn enabled more precise ad targeting, delivering up to 200% higher click- through rates for the 5 billion ads served daily across the FIM network…our research analytics team uses Greenplum Database to conduct tens of thousands of real-time tests against millions of users every day, analyzing each visitor’s reaction to ads against over with an absolute deluge of data.”