Drivers who buy the highest-octane fuel for their cars have an odd affinity for organic products. If a U.S. retail chain stocks expensive Bordeaux wine, the buyers make sure they have enough gourmet cheese on the shelves as well – wine drinkers tend to buy it when they pick up a pricey bottle at the supermarket. And a grocery store chain on the North Sea coast won't be surprised when people flood the shops on Saturday to buy steaks for the barbecue. Are these all signs of accurate gut feelings or even supernatural management abilities? Not at all – they represent a combination of talent, diligence and math – predictive analytics is a booming sub-discipline of big data analytics. Algorithms process big data and predict probabilities and trends. This boom isn't limited to the above mentioned retail examples, although the industry has been a pioneer for years. Smart companies use customer analyses, receipts, weather data, event data and past sales figures to make their purchasing more targeted and more customer-friendly.
IoT gives Advanced Analytics a boost
Specialists also expect a huge push to predictive analytics from the internet of things and Industry 4.0. Specifically, the IT can derive patterns from the data delivered by the machines and even use it to predict outages. When an engine has overheated, for example, the breakdown notification is already too late. But the root cause, such as defective bearings, can usually be detected in advance. As such, experts believe predictive maintenance will deliver revolutionary findings.
The market for predictive analytics is growing rapidly: Technavio expects global revenue to increase by 25 percent annually through 2019. And when we consider the potential savings these solutions offer, it doesn't seem far-fetched. According to figures from VDMA, a German industrial engineering association, predictive analytics can cut costs by 12 percent for unplanned repairs and nearly 30 percent for scheduled maintenance. And the experts from Roland Berger report that companies are increasingly viewing maintenance as a strategic enterprise function. They calculated that when predictive maintenance is used, it only takes up 15 percent of the total maintenance time, whereas reactive maintenance costs 40 percent of total time.
Many companies have a lot of catching-up to do
The flip side of this coin, however, is that predictive maintenance requires a lot of work. A great deal of catching up is needed, particularly on the technology side. In a recent study, "Advanced & Predictive Analytics 2016", analysts at BARC found that the business intelligence infrastructure at many companies is "insufficiently agile". "This means the technologies and processes available for the evaluations and analyses cannot be adapted to the new tasks quickly enough," say the experts. Moreover, according to BARC, more than 30 percent of the companies reported problems in their data management – such as lack of access to data sources or poor data quality – that hinder advanced analytics. Ten to 25 percent of the companies also complain about insufficient software support and inadequate system performance. BARC summarizes: "For advanced analytics initiatives to be successful, the players require technological support during the different project phases."
Predictive analytics have earned their place as a mega-trend in IT. However, companies and CIOs are well-advised to avoid rushing into things. Instead, they need to create an adequate technological foundation first. The expected massive return on such investments more than justifies the higher expenditures.
Companies nowadays need a much broader and more exact database in order to address customers successfully across all sales and communications channels. Big data is becoming a key competitive advantage.