Welcome To The Year Of 'Diverse Data Analytics'
Jon Ingalls, BlueKai's newly appointed chief marketing officer, said the next data evolution brings analytics into the equation to sort and sift through the numbers. He calls this "diverse data analysis."
Think of diverse data analysis as being an offshoot of predictive analytics, only with a bit of maturity. The electronics components industry tried to predict the amount of raw materials that should ship from suppliers to contract manufacturers in China. While most electronics original equipment manufacturers failed miserably, the execs in the online space thinks advertising can succeed. "If you think about cookies and data online, the natural 30-, 60, and 90-day renewal cookies gives you an outline to cast a predictive algorithm," Ingalls said. "The best way is to offer what I call fast hinting or directionally correct predictive analytics with certainty. Is it absolutely certain? No predictive analytics is absolutely certain."
Ingalls, who came to BlueKai through TrackSimple and Amazon, said there's a lot of benefit to predicting with 95% accuracy where conversions will occur. Statistical samples are based on data volume rather than time. At Amazon, Ingalls helped build a system that processed a third of a petabyte of data daily.
So, how will the ad industry separate the signals from the noise? Ingalls said BlueKai processes about 1.4 billion events daily. The results of the data end up in a cluster that let customers interact through a browser. They combine audiences to determine the net effect of targeting a specific demographic audience along with past purchase behavior.
The collected data gets loaded into memory, which allows users to perform fast operations. BlueKai tracks about 200 million unique cookies monthly in the U.S. and 300 million worldwide.