Software and SaaS News, Tips, and Thought Leadership | SaaS Tips, Software Engineering Blog, and Software News

Granularity is Key; Machine Learning for Personalization

Written by Pete Prestipino | May 5, 2015 5:00:00 AM

Consumers react well to personlization in their digital lives, but it's increasingly difficult for today's marketers to keep pace. Fortunately, thanks to some innovative approaches (and powerful technologies, of course), that is beginning to change.

Data analytics company Globys, for example, has launched a self-optimizing personalization product for mobile marketing campaigns that is powered by machine learning, identifying the "optimal experience" for each customer and context resulting in increased targeting granularity and higher revenue per experience across all campaigns. The result of its machine automated approach, according to Globys, is that Amplero is able to optimize marketing 200-500 percent better than manually coded targeting rules.

As you are likely aware, most personalization platforms today demand that enterprises manually design/develop rules for targeting users, depend on traditional testing and optimization tools, and, of course, spend a great deal of time fine tuning their efforts (using A/B testing to run experiments for example). With the immense amount of customer data available, however, those traditional approaches (and tools) just aren't keeping pace and as a result, marketers are beginning to see diminishing returns.

Globys' Amplero solution, however, leverages a continuous cycle of self-designed experimentation and machine learning to help marketers discover and execute that optimal experience for each customer throughout the entire digital journey.  The Amplero solution essentially uses "science-based" target and control groups to test combinations of customer, experience and execution contexts and then apply a range of analytics across numerous variables to find targeting optimizations with statistically valid lift (Amplero creates a revenue lift score for every combination of experience, customer and execution context). By using machine learning, marketers can then essentially eliminate the processes of designing, developing, testing, and optimizing targeting rules by automatically discovering the conditions and context of optimal targeting. 

"Granularity is the key to breaking the 80/20 rule where twenty percent of the customers deliver eighty percent of the campaign result," said Dr. Olly Downs, chief scientist at Globys.  "Amplero scores experiences in thousands of contexts to ensure the other 80 percent of the customers get the experience that is relevant for them."