There’s no denying that data rules our world. The right data in the right hands can steer political events, help companies deduce how best to part consumers from their cash, identify areas of waste and duplicated effort, and even build predictive models to help us understand the present and anticipate the future.
Predictive analytics is one of the most important tools we have for putting humanity’s zettabytes of data to work for us. Here are four industries finding consequential ways to put this tech to good use.
1. Healthcare
If you want an idea of how much data is floating around in the healthcare market, consider that U.S. hospitals alone see more than 36 million admissions every year at an estimated cost of almost $1.1 trillion. In healthcare, predictive modeling slices through the digital “noise” generated by millions of ailing individuals to find meaningful patterns in the chaos.
These patterns, once observed, can help physicians identify and understand common risk factors across larger groups of patients and predict the likelihood of patients developing health problems. Predictive analytics that draws on these vast troves of data can predict a patient’s likely response to various treatments, too.
According to Jeff Howell of AlayaCare, his organization brought about a 73 percent reduction in ER visits among senior citizens by using daily health metrics to predict emerging health problems “among a chronically ill patient set.” No easy task — but one made far easier (and cheaper) to solve by predictive analytics.
2. Retail and Enterprise Planning
It’s not news anymore that major retail companies like Amazon.com use predictive modeling to anticipate customer needs and provide relevant suggestions. What’s more interesting is the number of smaller players getting in on this opportunity. One estimate says retailers at large have realized $1.7 trillion in “new value” by adopting data-driven enterprise tools.
Enterprise resource planning (ERP) systems are an ideal match for predictive analytics. For example, any number of outside factors, such as disasters and weather events, impact buying behavior. Drug companies are able to plot many months’ worth of demand for allergy medications far in advance of allergy season by leveraging past sales data and current patterns to make future predictions.
Combining predictive analytics with ERP systems has helped reverse past trends, including pharma companies losing up to 10 percent of their potential business during cold season due to products being out of stock.
Now, instead, companies like Bayer can use data modeling to predict up to nine months’ worth of demand, feed this information into their ERP systems, and keep an entire supply chain stable and profitable. On the flip side, a predictive ERP system can also scale-down demand based on the data so companies don’t engage in more mass production than is necessary at any given time.
3. Cybersecurity
Cybersecurity will remain a problem so long as criminals keep finding ways to defraud the public and so long as companies keep developing profitable ways to address the threat. Merely developing new weaponry to keep hackers and fraudsters at bay is like throwing down a gauntlet and begging for a worthy adversary.
The U.S. National Institute of Standards and Technology believes predictive analytics is key to realizing stronger collective cybersecurity efforts. Indeed — as proven by DARPA’s “Cyber Grand Challenge” in 2016, there are some types of cyber threats for which prediction and automation is perhaps the only effective answer. This was the first “hacking contest” to feature bots trying to out-hack each other.
One overseer of the contest called the results “astounding,” as the bots showed better-than-human abilities to find and plug security vulnerabilities. Even so, their abilities are far from perfect and some bots even stopped functioning without notice during the contest.
Even so, prediction analytics is a strong new addition to cybersecurity portfolios. Using pattern recognition and big data, automated cybersecurity solutions can warn their human counterparts about the type and severity of potential incoming threats. Too many active cyber threats go undetected until it’s too late.
Instead, analytical models can study traffic and other “signatures” in real-time and tell better than any human whether the signs indicate an impending or in-progress attack. This is the ongoing mission of several public and private institutions, including the Department of Homeland Security and telecommunications companies.
4. Human Resources
There are several tasks under the purview of HR that may benefit tremendously from the application of predictive analytical models. One is retention, and the other involves training and addressing company skill gaps.
IBM has their eye on the costly problem of retention. The company developed and deployed predictive tools that can reportedly identify (with up to 95 percent accuracy) which employees are likely to depart the company, even before they hand in their notice.
Employee turnover is costly, so having more lead time is a big deal. It means decision-makers don’t waste retention efforts on the wrong people, and that they’re not overlooking potential problems until things reach a breaking point.
The other area with lots of potential for predictive analytics involves training employees and filling skill gaps within a company. Online learning is an important tool for keeping employees knowledgeable and engaged. But Coursera and other companies want to take things further by using machine learning to predict which training courses and materials are best-suited to which employees, and help companies match aptitudes and interest with the work that needs doing.
This, too, is likely to be a boon to retention and engagement, not to mention a way for companies to weather staffing shakeups and labor shortages.
Predictive Analytics for an Unpredictable World
The widespread use of data in predictive analytics brings some new types of risks that should be on our radars, as well. The governments of the world are, rightfully, becoming more involved in the politics of privacy, for example. And studies of algorithms reveal that they inherit some of the same biases as their human creators.
Even so, predictive analytics has revealed itself as a powerful tool. In these industries and many more, expect it to continue changing the game as we know it.
