Predictive Analytics
Predictive Analytics
The Power to Predict Who Will Click, Buy, Lie, or Die By Eric Siegel
We’re obsessed with predictions: Will it rain tomorrow? Where are interest rates going? But our usual response to the science of prediction is that it’s hard to understand or too boring to learn. But according to Eric Siegel, it’s easier to understand and far more powerful than we imagined. There’s virtually no limit to the ways predictive science can be applied. • • •
Google search trends are now being used to predict outbreaks of disease. Online dating sites correctly predict your best bet for a life mate; and, The IRS has gotten pretty good at predicting if you are cheating on your taxes.
The Era of Big Data We’re currently in a prediction revolution, powered by cheap and unfettered data. Each day, nearly every move we make - online and offline - is being recorded, and shared. Making all of this data growth affordable is the falling cost of data storage. In the late 1980s, for instance, it cost about $1 million to store a gigabyte of data. By 2010, it cost 10 cents.
The Prediction Effect There’s a central message in Siegel’s book: A little prediction goes a long way. For example, although it would be ideal to know everyone in your database who is likely to buy from you, using predictive analytics to improve your chances by even a percentage or two can make the difference between profit and loss on a marketing campaign.
Correlation Does Not Imply Causation If you discover a “predictive relationship” between variables, it doesn’t necessarily mean one causes the other. So even though increased ice cream sales correspond with increased shark attacks, one likely doesn’t cause the other (hot sunny days likely play the larger role). The good news is that for predictive models to work properly, it’s not necessary to understand causation but rather to simply predict a future occurrence with some degree of accuracy. So while ice cream sales don’t cause shark attacks, if you knew how many Drumsticks were sold that day, you could still offer a pretty reliable prediction about swimmers’ safety. “Remember,” writes Siegel, “when push comes to shove, prediction trumps explanation.”
TheBusinessSource.com
All Rights Reserved
The Power to Predict Who Will Click, Buy, Lie, or Die By Eric Siegel
We’re obsessed with predictions: Will it rain tomorrow? Where are interest rates going? But our usual response to the science of prediction is that it’s hard to understand or too boring to learn. But according to Eric Siegel, it’s easier to understand and far more powerful than we imagined. There’s virtually no limit to the ways predictive science can be applied. • • •
Google search trends are now being used to predict outbreaks of disease. Online dating sites correctly predict your best bet for a life mate; and, The IRS has gotten pretty good at predicting if you are cheating on your taxes.
The Era of Big Data We’re currently in a prediction revolution, powered by cheap and unfettered data. Each day, nearly every move we make - online and offline - is being recorded, and shared. Making all of this data growth affordable is the falling cost of data storage. In the late 1980s, for instance, it cost about $1 million to store a gigabyte of data. By 2010, it cost 10 cents.
The Prediction Effect There’s a central message in Siegel’s book: A little prediction goes a long way. For example, although it would be ideal to know everyone in your database who is likely to buy from you, using predictive analytics to improve your chances by even a percentage or two can make the difference between profit and loss on a marketing campaign.
Correlation Does Not Imply Causation If you discover a “predictive relationship” between variables, it doesn’t necessarily mean one causes the other. So even though increased ice cream sales correspond with increased shark attacks, one likely doesn’t cause the other (hot sunny days likely play the larger role). The good news is that for predictive models to work properly, it’s not necessary to understand causation but rather to simply predict a future occurrence with some degree of accuracy. So while ice cream sales don’t cause shark attacks, if you knew how many Drumsticks were sold that day, you could still offer a pretty reliable prediction about swimmers’ safety. “Remember,” writes Siegel, “when push comes to shove, prediction trumps explanation.”
TheBusinessSource.com
All Rights Reserved
