Improve L&D Decisions with Predictive Analytics in eLearning

The business world is full of decisions that would be somuch easier with a crystal ball to help predict the future: Which applicantshould we hire or promote? Of this quarter’s new sales recruits, which ones aremost likely to succeed as team leaders? Will learners engage with thissimulation, or should we develop a serious learning game instead? Will sendingthis manager to an MBA program pay off? Using predictive analytics in eLearning—aswell as in hiring and promotion decisions—can help with these choices.

Study history to detect patterns

Data provide a record of what has happened in the past. Thisinformation is useful in a variety of ways (learn more at The eLearning Guild’sData& Analytics Summit)—but identifying historical patterns can be agreat way to predict future patterns or behaviors. For example: Collegesand universities study historical data to identify students most—orleast—likely to complete a degree. They can use that information to decidewhich students to accept. They can also use it to figure out what failingstudents have in common—and tailor assistance programs to target commonweaknesses. This approach is transferrable to corporate L&D, where using dataanalytics to identify employees’ weak areas or potential weakareas enables instructional designers to create better training, as well as to personalizeeLearning.

Some predictive analytics tools harness artificialintelligence algorithms to sift through large amounts of data—choosingapplicants to interview from a pool of thousands, for example. A caveat withautomated predictive analytics is that some historical patterns are notdesirable and could be the result of historical or implicit biasbaked into algorithms. For example, an algorithm could easily learnfrom studying successful top-level executives that, in addition to, say, havingan MBA and 10 years of management experience, the clear majority of “successfulexecutives” in US corporations are white men. Thus, use of predictive analyticsshould be balancedwith human intervention to ensure that undesirable patterns are notinadvertently imbedded in key personnel decisions. Used in conjunction withhiring committees and performance support tools that support broad corporatehiring goals, predictive analytics can play an important role in recruiting,hiring, promotion, and training strategies.

Put predictive analytics to work in eLearning

Looking at analytics data from a mindset of predicting learnerbehavior makes it possible to anticipate what employees will engage with andhow their performance could improve as a result.

On an individual level, that could mean tailoring eachlearner’s roster of training courses and performance support tools based onpast behavior. Consider a learner who has never completed a video longer thanthree minutes but who consistently engages with a chatbot that sends shortgames, quiz questions, and links to short articles. Placing that learner in avideo-based training course would be a mistake!

On the other hand, if data from course evaluations showsthat most learners engaged with the interactive video content but loathed thediscussion boards, a virtual classroom with interactive videos and otheractivities might be just the ticket.

Leveraging predictive analytics is just one way to takeadvantage of data that most companies already collect. The eLearning Guild’sresearch report PuttingData to Work explores more opportunities for using and applyingdata and using predictive analytics in eLearning.