“People are overlooked for a variety of biased reasons and perceived flaws. Age, appearance, personality. Bill James and mathematics cut straight through that.”
— Peter Brand, Moneyball
In the movie, Moneyball, Peter Brand is an analyst that revolutionizes baseball by using performance data. Baseball players tend to perform in the future as they have in the past – particularly the recent past. So past performance is the best predictor of future performance of baseball players. Now all professional baseball teams use such analytics.
This is performance analytics, and it is the prevailing approach to predictive analytics of human behavior. Performance analytics involves observing what happened through measurement of key performance indicators, and then using this data to predict what will happen. Through performance analytics, organizations have realized incredible value predicting behaviors of all sorts – consumer, employee, supplier, etc. To conduct performance analytics, organizations need to build capabilities in measurement, data management, data wrangling, and rather straightforward statistical techniques for prediction.
However, past performance is not the only useful predictor of future performance. People can change behaviors and this can impact performance rather dramatically – with changed behaviors, past performance may no longer an adequate predictor of the future. Therefore, leading organizations are now moving beyond performance analytics to what is sometimes described as “interventional” analytics that involves diagnosing and changing behaviors. Interventional analytics is significantly different from performance analytics in a variety of ways. Instead of only observing past performance, interventional analytics requires that organizations observe more fine-grained behaviors that lead to that performance. This involves richer observational techniques, stronger approaches to analysis of those observations, and other challenges that this approach entails. Organizations looking to move to interventional analytics (Moneyball 2.0) need to build capabilities that go beyond those that they developed for traditional performance analytics (Moneyball 1.0).
In this presentation, we draw on experiences with interventional analytics at the University of Notre Dame’s celebrated athletic programs to garner lessons about building these capabilities and the challenges implementing them. Through these experiences we note that there are two fundamental areas with interventional analytics that are different from performance analytics – these involve (1) observing and measuring behaviors in addition to performance, and (2) generating and analyzing interventions is different than performance analytics. Next we address each with examples.
Join the discussion moderated by Professor of Information Systems and xLab faculty director Youngjin Yoo.
Registration is required. This webinar is free and open to all.
For more information, visit the xLab website or contact xLab@case.edu