Magazine article Talent Development

Data-Driven Personalization Comes to Learning: Technology Can Help Learners by Understanding Their Complex Preferences and Adapting Accordingly

Magazine article Talent Development

Data-Driven Personalization Comes to Learning: Technology Can Help Learners by Understanding Their Complex Preferences and Adapting Accordingly

Article excerpt

Like virtually every part of our lives, corporate training has changed radically in the past few decades. We may not stop to think about it too often, but it's remarkable that we only began to complement traditional training programs with computer-assisted training in the 1990s; that e-learning only gained traction in the early 2000s; and that blended, informal, and collaborative learning only came to the forefront in the past few years.

Each of these developments represents a significant step forward and benefits trainers and learners alike. At the same time, the modern workforce (and the modern learner) has changed just as quickly, and it's crucial that we keep pace. According to Bersin by Deloitte, modern learners are "overwhelmed, distracted, and impatient." They expect content to be engaging, seamless, personalized, and on demand. How can we better address that growing demand?

For McGraw-Hill Education, the answer lies in mastery-based adaptive learning that's both modular and data-driven.

Making learning modular

Modularized, bite-sized content is nothing new. When we think about how we consume content at home, the trend toward micro-content has been gathering steam for some time and is best exemplified by social media, where such bite-sized snippets as a tweet, like, pin, and meme reign supreme. When we think video, what used to be a 30-minute TV show is now likely to be a three-minute You Tube clip. Content is now optimized for swift consumption and is well-suited to short attention spans.

For learners, bite-sized content means flexibility. No longer must learning be delivered exclusively through long readings and lessons; it can now be consumed in smaller, more digestible chunks. That's particularly great news for distracted and overwhelmed learners for whom on-the-go learning is increasingly a necessity.

Importantly, micro-content isn't just about making lessons smaller; it's about making them more flexible and organizing them in a more effective manner, ensuring we have the modern learner's attention by showing what is relevant for each learner, at the right time. These bite-sized pieces of content, while valuable in their own right, become exponentially more powerful when incorporated into a fluid ecosystem--backed and synchronized by data.

The basics of data-driven personalization

Most of us are familiar with basic online user data such as click rates, time spent, bounce rates, exit and entry points, and traffic sources. These behavioral data are used by companies to optimize websites and applications and make them generally appealing, but the data also can be used to custom-tailor the experience of each user.

Take the popular personalized radio engine Pandora, for example: User preferences, in the form of a simple thumbs up or thumbs down for each song, help to build a personalized database of likes and dislikes, and ultimately determine which songs a user hears going forward. A similar dynamic is at work behind Netflix's video recommendations and Amazon's e-commerce recommendation engine--a combination of user-submitted and inferred preferences are drawn upon to tailor what is presented to the user.

In every case, the quality of the recommendations depends on two primary factors: how frequently the user interacts with the content and how finely the content is tagged, such as by genre, structure, tone, or subject.

When it comes to tagging, the options are potentially infinite, and that can be a powerful advantage. Did you enjoy the movie because of the strong female lead, or the cinematography, or the score, or the script? In reality, it was most likely due to some subtle combination of all the above--and, with the right tagging and enough user interactions, those complex preferences can be identified. Even if a particular user doesn't understand why she found something enjoyable, each interaction helps the engine refine its understanding of her preferences. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.