While traditional web analytics refers to the practice of retrospectively measuring, collecting, analyzing, and reporting on web data for the purpose of understanding and optimizing the user’s online experience and website usability, the practice of predictive web analytics focuses on calculating the statistical likelihood of future online occurrences and events.
Predictive analytics encompasses a variety of statistical techniques, modeling, machine learning algorithms and data mining which analyze current and historical facts to make predictions about future or unknown events.
Predictive Analytics Software
In the world of A/B testing and personalization, predictive analytics tools are usually broken down into the following capabilities:
- Predictive Segmentation: Automatically identifies and creates meaningful visitor segments characterized by a higher probability to react in a predictable manner to certain events.
- Predictive Targeting: Predicts which experience would be most suitable for each visitor segment or individual visitor; suitability is measured by probability to meet certain objectives in a given experience, such as completing a purchase, signing up, etc.
Each of the above capabilities can arm online marketing and eCommerce managers with the power to enhance current workflows by automating manual processes and thereby reducing the amount of time and effort currently invested. However, only the presence of both complementary capabilities can yield true predictive analytics powers. Predictive Segmentation without Predictive Targeting creates valuable visitor segments but lacks insight into what the best possible experiences might be. Meanwhile, Predictive Targeting without Predictive Segmentation can match ideal experiences to each segment, but segments would have to be manually identified and defined prior to deploying experiences.
