I just read Martin Couzin’s article on curation and this is a fascinating and important topic. It was interesting to find that in his survey 45% of respondents find curation important while most of the rest think its a personal project or something that sounds “interesting” but not important quite yet.
Let me start this article by stating that yes, of course, curation is incredibly important. Nobody wants to read news, articles, or view instructional content that is irrelevant, incorrect, or possibly misleading to our goals. This problem now plagues Facebook, Twitter, and really every social system we use, especially those we use to learn.
In the case of learning, I think there are quite a few challenging dimensions to curation, which make it even more difficult. And in the interest of opening up the discussion, I’d like to present this for people to consider and give us feedback.
- What is my job? If we as L&D professionals know what a person’s job entails, we can hopefully curate content that is relevant, proven to be useful, and structurally needed to do work well. A sales person who puts bad data into Salesforce, for example, needs curated content to make him more effective. A production worker who is breaking things needs content to improve quality and safety. And these are topics which should be carefully “curated” based on job needs, job activity, and observed experience.
- What is my career goal? We all want to do our job well now, but we also want to be promoted, move ahead, or change roles in the future. How do I curate learning that is “most appropriate” for the job I want to do next? Can we assume everyone wants to move into management? Of course not. This is a second “dimension” to curation, and one we have to seriously develop using assessment, data about successful career paths, and our own organizational needs. Just because I’m a data scientist, do I want to learn more about algorithms? Or visualization? Or maybe how to get data out of our SAP system? These are tricky questions to answer.
- What is my credentialing goal? What credential, degree, or other form of attribution does the individual want to achieve? If I want to become a “certified specialist” and get promoted, the curation should clearly show me how to do this. We have both individual and organizational credentials to deal with – a simple “click here to view the senior specialist path” is a good curation option to provide, for example. Is that in your roadmap?
- What will AI and data tell us? As I discuss in this article about the LXP market, we have lots of good ideas how to use machine learning to recommend content. But is it working? Does it reflect these topics above? Can it find me the “right expert” or the “right author” or the “right content” I am most likely to enjoy and use? Is it simply a matter of recommending the most “liked content?” It’s much harder than it sounds.
As we built out the Bersin research library, for example, we found that certain articles and research reports were 10-20X more popular than others. We were excited to think that “aha if we look at these articles we can better curate our library.” But there wasn’t much of a pattern. Why?
It turned out the “most frequently read” material was those we promoted in our newsletters! They were “top of the list” and easiest to click, so they tended to get read a lot more. Over time people gravitated to the authors and topics they wanted the most, but it was dauntingly hard to find algorithms that do this well. In the end we settled on some fairly simple rules, and “curate” based on traffic, related topics, and click paths others have taken.
All I”m trying to say is that this problem is trickier than it looks, and we need lots of good ideas, discussion, and models to make curation the powerful tool we all know it can be.
I’m excited to be on the Board of LPI and look forward to comments, ideas, and feedback.