Valued at roughly $694.6 million in 2013, the predictive analytics market is expected to triple by 2019, reaching more than $2.3 billion, according to market research firm Micromarket Monitor. During the next five years, the predictive analytics space will grow 25 percent annually in the United States, while the global market is expected to see a 32 percent annual growth rate, the firm forecasts.
Given the amount of data generated by consumers every day, these projections should come as no surprise, Steven Ramirez, CEO of consulting firm Beyond the Arc, says.
"Thanks to the popularity of mobile devices, data is being created at unprecedented rates," Ramirez says. "Plus, the data created through mobile devices—and other smart devices emerging out of the Internet of Things—is unlike any other data businesses have seen before. These devices are the reason predictive analytics tools are becoming as crucial and as useful as ever."
Mobile devices have opened the door to more continuous relationships with consumers and allow companies to reach their customers virtually anywhere and at any time. And because consumers constantly use their devices to browse apps, post on social media, check email, and take part in other engagements, the interaction data has not only grown in volume, but also become richer and more valuable to brands.
The addition of sensor data, transaction data, and data from Internet of Things devices has broadened the scope of customer engagements even further. "In the past, there was not enough information to go on for predictive analytics. Now, because of the mobile and Internet of Things revolutions, companies always know what customers are doing and what they're looking for, and by studying patterns in large volumes, they can predict what they'll do next," Ramirez says.
Currently, most predictive analytics solutions can help companies predict and prevent customer churn, score leads, identify the customers who will have the greatest lifetime value, and determine which customers are likely to respond to specific offers. Legacy companies such as SAP offer comprehensive solutions, but smaller vendors, such as Lattice Engines and Swiftpage, have strong offerings as well, analysts agree.
Though the technology is already "incredibly powerful," it will continue to become more sophisticated as vendors incorporate cross-channel analytics into solutions in the coming years, Ramirez expects. Increasingly, vendors will develop ways to look at customer interactions over the span of an entire customer journey, including a combination of different touch points rather than just one channel. "Unopened emails could be an indicator of customer churn, but if that same customer that isn't opening emails is actually opening the app almost every day, then there might not be a churn problem. It might just be a question of determining the preferred channel," Ramirez says, explaining that predictive analytics will become more nuanced and channel-sensitive.
Next-generation predictive analytics solutions will also make look-alike modeling a "more exact art," says Jay Famico, practice director of sales and marketing technology at Sirius Decisions, a B2B technology research firm. Look-alike modeling combines the insight collected through analytics to establish patterns of behavior and apply those patterns to noncustomers. By leveraging look-alike modeling, companies can zero in on prospects that could exhibit behaviors similar to their current customers and target them with offers and messaging.
Along the same lines, predictive analytics is paving the way for companies to identify social spheres of influence. "The next step for predictive analytics is going to be going outside the customer list and seeing who they're influenced by on Facebook or LinkedIn in order to build brand awareness at the top of the influencer chain," Famico says. The technology isn't quite there yet, but will get there in the next few years, he says.
As predictive analytics technology matures, its long-term effects will be complex, analysts expect. On one hand, many solutions that vendors are developing will be enterprise-level "workhorses" with far-reaching, scalable processing power. On the other hand, however, vendors are attempting to make these solutions simple for "any and every business user," Ramirez explains.
"These goals are almost at odds with each other, but this is where we're heading," he adds. "Data is growing quickly, so the solutions will need to scale up, but this can't come at the price of usability, because everyone in the organization needs access to predictive data. It's becoming a must-have."