The term “predictive analysis” may invoke an image of giant machines crunching massive permutations of seemingly irrelevant data into nuggets of actionable information. It can just as likely invoke an image of agitated analysts running to and fro in a high-pressure atmosphere. Indisputably, some look suspiciously at this notion of predicting the future.
Predictive analytics was born to capture opportunity and mitigate risk. Capturing an opportunity means beating someone else to the finish line. Mitigating risk means eliminating threats—and possibly the competitors that take our market share.
The methods for predictive analysis are sophisticated and follow a shared approach, gathering both prior and existing data points and mapping them to a timeline. The quality and disparity of the data points boost the result’s integrity. The timeline has no end because those dates haven’t arrived yet. However, we can infer from the trend where the line may continue. If the data is solid enough, we can stretch that predictive line even further into the future.
Are the variables always objective, knowable or measurable in purely mathematical or logical terms? If only it were that easy. Notice the phrase “solid enough” in the prior paragraph. Do we trust the data? Do we trust the processes that delivered it? Do we trust the operators of those processes, or the analysts interpreting the results?
Finally, the whole engine is predicated on whether the model, or approach, is the right one for us. What if the modelers got it wrong? A model only reflects the mind of the modeler, not necessarily reality. If the modeler chose poorly, nothing useful will come of it.
How much are we willing to bet on all this? Bit of a quagmire, no?
One man used his extraordinary data mining skills over a vast network of social media to winnow-down a list of email addresses and companion phone numbers of people who were both willing and ready to buy his client’s product. This was predictive analytics gold. However, after their call-center exhausted the list, less than one percent of the targets had any interest. Why the misfire? He interviewed the call-center folks, and the feedback stunned him. It seems that while those sales targets were ready to make a purchase, they were “creeped out” that someone discovered this fact without their prior disclosure. Many of the targets asked, “How did you get this number? Who told you this information? How could you possibly know this?” The point being, people may be highly predictable in their habits and behaviors, but they don’t like the idea that predators are stalking them with it.
Stock market analysis is replete with gurus and salesmen ready to take the money and run. Bernie Madoff bilked more than $50 million dollars from his name-brand clients in a Ponzi scheme, claiming he had unbeatable predictive engines for the market. Smart traders know the truth: at any given moment, a stock has a 50/50 chance of going up or down. Nobody can predict which way.
A case in point was the fear-mongering around Y2K. Many people weren’t worried one bit about the world coming to a halt on January 1. What we didn’t predict was how people outside the computer-related industry would react. Many acted out of fear and cashed out their mutual funds and put the money in a safe place. This activity (mass-selloff of stock) caused a rapid deflation in value for those who were unafraid. The brave souls who stuck to their guns got it stuck to them as those mutual funds lost ninety percent of their value – or more. Their hard-earned nest eggs evaporated overnight because of fear, not from any technological reality.
Which brings us around to guarding against loss. One of the most important forms of predictive analysis in use today is in law enforcement, particularly with crimes-against-persons and missing persons. Big data analytics provides critical information in a timely manner. For these crimes, every second afterward is precious. Crime scene reconstruction may take place across a city or a nation as a perp is profiled, hunted and captured before harm can come to their human prey. At an international level, predictive analysis stops nefarious strategists and tacticians who practice their skills on the masses, averting the loss of life, property, and even the collateral loss of market value.
Sirius data and analytics experts help clients gain deeper insight and get business analytics right the first time. Our team can help you discover a predictive analytics technology that suits your organization. Visit www.siriuscom.com/business-analytics for more information, or contact us to learn more about our offerings.